Task Conflict Control Mechanism – Characteristics and ...

Task Conflict Control Mechanism –

Characteristics and Implications

Thesis submitted in partial fulfillment of the requirements for the degree

of "DOCTOR OF PHILOSOPHY"

By

Eyal Kalanthroff

Submitted to the Senate of Ben-Gurion University of the

Negev

16 July, 2014

Beer-Sheva

ii

Task Conflict Control Mechanism –

Characteristics and Implications

Thesis submitted in partial fulfillment of the requirements for the degree

of "DOCTOR OF PHILOSOPHY"

By

Eyal Kalanthroff

Submitted to the Senate of Ben-Gurion University of the

Negev

Approved by advisor

Approved by the Dean of the Kreitman School of Advanced Graduate Studies

16 July, 2014

Beer-Sheva

iii

This work was carried out under the supervision of Prof. Avishai Henik

In the Department of Psychology

Faculty of Humanities and Social Sciences

iv

Research-Student's Affidavit when Submitting the Doctoral Thesis for Judgment

I Eyal Kalanthroff, whose signature appears below, hereby declare that:

I have written this Thesis by myself, except for the help and guidance offered by

my Thesis Advisors.

The scientific materials included in this Thesis are products of my own research,

culled from the period during which I was a research student.

This Thesis incorporates research materials produced in cooperation with others,

excluding the technical help commonly received during experimental work.

Therefore, I am attaching another affidavit stating the contributions made by myself

and the other participants in this research, which has been approved by them and

submitted with their approval.

Date: _________________ Student's name: Eyal Kalanthroff

Signature:______________

v

Acknowledgments

My gratitude to Prof. Avishai Henik, my supervisor and mentor, who ignited in me

the enthusiasm and enjoyment gained from research and who believed in me and my

ideas from day one. Thanks for useful comments on any subjects – it was always great

to know that I could count on his honest care on any subject.

To Dr. Gideon Anholt, who introduced me to the field of OCD in both clinic aspects

and the lab, for his personal support, and for always being willing to help during both

successful and disparaging days.

To Desiree Meloul for her useful comments, and especially for a friendly smile and

helping hand on every possible issue.

To Prof. Marius Usher for his endless patience when discussing and explaining

scientific ideas, theories and methods and for believing in me as a researchers at every

step of the way.

To Dr. Liat Goldfarb who assisted me greatly in conceptualizing the basic ideas on

which this research is based.

During my PhD I also conducted an internship in clinical psychology. I would like to

thank Zvi Fajerman, Ornit Rosenblat, Limor Swisa, and Daphna Snir who guided and

tutored me in my very first steps of becoming a clinical psychologist. Your

acceptance and support allowed me to truly combine clinical practice and research.

I would also like to thank my parents, Nili and Efraim Kalanthroff, my sister Galit

Kalanthorff-Klain, and my brothers Rafi and Ziv Kalanthroff for their unconditional

love and support, for being there to actively encourage my studies from before I

learned how to read and write and for much after I finished my PhD. You gave me

vi

(and still give me) the basic tools to study and investigate, and the confidence to dare

and try new things. Most importantly you planted curiosity in me at very early stages

of my life.

Last, I would like to thank my beloved wife Shani for being there with me and for me

at every minute of every hour of every day. You inspire everything I do and give me

the confidence to do it. All of this (and much more) would not have been possible

without you! Words cannot express my gratitude and appreciation. Thanks! I dedicate

my PhD dissertation to our little personal miracle, Gili.

vii

Contents

CHAPTER I: Introduction 1

Cognitive Control 1

Task Conflict Control 3

The Stroop Task 5

Obsessive-Compulsive Disorder (OCD) 6

The Current Work 7

CHAPTER II: Experimental Work 10

Experiment Group 1: Stop Interfering: Stroop Task Conflict Independence

from Informational Conflict and Interference 10

Experiment 2: Individual but not Fragile: Individual Differences in Task

Control Predict Stroop Facilitation 23

Experiment Group 3: Evidence for Interaction between the Stop Signal and

the Stroop Task Conflict 31

Experiment Group 4: Preparation Time Modulates Pro-Active Control and

Enhances Task Conflict in Task Switching 46

Theoretical model: What Should I (Not) Do? Control Over Irrelevant Tasks

in Obsessive-Compulsive Disorder Patients 60

CHAPTER III: General Discussion 65

References 71

viii

Abstract

Using behavioral tasks and imaging techniques, it has been previously shown that task

sets can be activated by the perception of a stimulus attribute that is strongly

associated with a particular task set—when one sees an object in the environment one

immediately perceives not only the external features but also directly perceives what

tools it affords in terms of meaningful actions. Task control is a mechanism

responsible for promoting and maintaining goal-directed actions and suppressing

automatic but irrelevant ones. Not much attention has been given to the task control

idea and to its important implications on cognitive psychology, clinical psychology

and neuropsychology. The main goal of the present study was to identify and

characterize the task control mechanism and to specify its important implications. The

main task that I used in order to investigate task control was the color-word Stroop

task, in which performance reflects two conflicts—informational (between the

incongruent word and ink color) and task (between relevant color naming and

irrelevant word reading). Thus, the Stroop task allows a behavioral indication for task

conflict—a reverse facilitation effect. In the first experiment I found that task conflict

is independent from informational conflict, and thus it will occur even if the automatic

task always leads to the same response as the relevant task (hence informational

conflict is not possible). In the second experiment, I conducted an individual

differences study and found that the efficiency of task control is correlated with the

efficiency of inhibitory control. The third experiment continued that idea and found

indications for task conflict in momentary inhibitory control failure trials. In the

fourth experiment I used a version of the task-switching paradigm and found that task

control was reduced if the task was not constant (i.e., frequently changing) and that

preparation time positively affected the ability to manage task conflict. In the fifth

section, I discuss some possible implications for task control research for people

suffering from obsessive-compulsive disorder (OCD). In the latter section I base our

theoretical model on findings showing indications for task conflict in OCD patients,

which also correlate with symptoms severity. Finally, in the Discussion section I

propose a computational neural-networks model for task conflict that can account for

the results in all the experiments of the current work.

Keywords: OCD, executive functions, inhibitory control, task conflict, Stroop

1

CHAPTER I: Introduction

Cognitive studies on executive control of both healthy and clinical populations

have focused mainly on the abilities to manage multiple pieces of information.

Mainly, the ability to differ between relevant and irrelevant (though distracting)

information and the ability to effectively process more than one source of information

(sometimes containing contradicting information) were studied. Many times

information in our environment is processed automatically. The concept of

automaticity commonly (though not always) refers to the cognitive system's tendency

to processes specific kinds of stimuli in an obligatory fashion, even when conscious

choice to process this information is absent. Researchers usually focus on the features

of these stimuli, which tend to be automatically processed (and hence tend to

influence our behavior), or focus on the characteristic of the cognitive system that

aims to allows us to suppress this automatic processing and maintain a goal-directed

behavior. This line of research has proved very interesting and useful to our

understanding of the cognitive system and its functioning. It has taught us greatly

about various clinical populations and led to some groundbreaking discoveries in

various fields such as the influence of emotional stimuli and the processing of

numbers and quantities. Nevertheless, focusing cognitive research merely on

information processing neglects a very important aspect of executive control—task

selection and automaticity of specific tasks. Though one may think that tasks are

evoked only by deliberate intentions, in this work we will suggest that tasks can also

be automatically evoked simply by the perception of a stimulus in the environment.

We will try to show that these automatically evoked tasks can cause a conflict if they

contradict other deliberate tasks and that a specific cognitive mechanism is needed in

order to manage this conflict. Finally, we will argue that a deficit in this cognitive

mechanism has far reaching consequences, and we will demonstrate this on a specific

clinical population—those with obsessive-compulsive disorder (OCD).

Cognitive Control

Our environment is composed of an endless amount of stimuli that we can

processes at any given moment. These stimuli contain pieces of information that are

2

often irrelevant, not connected, or even contradict our current stream of thoughts or

actions. In many cases these stimuli “invite” us to use them (like a television or a

phone), though often doing this can impair our current goals (like writing your PhD

thesis). The reason we cannot process or attend to all this information and tasks

simultaneously is that our cognitive resources are limited. Thus, on many occasions

we face the need to focus on one object, dimension, situation, or task, and to ignore,

inhibit, or defer other aspects for later processing. Moreover, we often are forced to

solve a conflict in information processing or a conflict between two simultaneous

tasks. Conflicts usually include the need to suppress or inhibit a non-appropriate or an

irrelevant response, thought or emotion. Solving these conflicts is considered to be a

central function of cognitive control (Allport, 1980; Norman & Shallice, 1986).

Cognitive control is a key human capacity that enables us to flexibly respond to the

environment and produce goal-directed behavior, freeing us from the constraints of

automaticity or stimulus bounds (Miller & Cohen, 2001). It is generally believed that

control is not unitary and that different control processes (e.g., conflict monitoring,

inhibition, and task selection) are required in different tasks (Miyake et al., 2000).

It had been suggested that inhibition is a hallmark of executive function (van

Veen & Carter, 2006; Verbruggen & Logan, 2008). Inhibition is defined as the ability

to stop the current course of thoughts or actions. An act of inhibition requires

distinguishing early on between a relevant and irrelevant stimulus in relation to the

specific response. The concept of inhibition has been largely investigated, and many

researchers urged discerning between different types of inhibition and different

inhibitory mechanisms (Harnishfeger, 1995; Nigg, 2000; Rafal & Henik, 1994).

Another aspect of cognitive control is task demand control. In order to

effectively perform relevant tasks, the ability to keep in mind the task set and

instructions regarding the task, while clearing the noise of irrelevant thoughts, is

extremely important. vanVeen and Carter (2006) suggested that control is often

conceptualized as the ability to represent and maintain the task requirements.

Recently, a dual-mechanism of control framework was proposed by Braver (2012; see

also De Pisapia & Braver, 2006). According to this framework, much of the

variability in control processes is due to the balance between two types of control

processes: a pro-active process that is deployed in advance of the stimulus (e.g., pay

3

attention to color and ignore the word) and a reactive process that is deployed after

the stimulus is processed and generates some type of conflict (I see the word “red” in

blue ink color, but I need to respond to the color, so it’s blue; see Braver, 2012; Figure

1). As Braver discusses in his review, this account explains aspects of control

variability in a series of behavioral and imaging studies. To summarize, cognitive

control is needed in order to manage a huge amount of information with our limited

cognitive system. Inhibition is one of its most important features and keeping the task

set activated is one of its major goals.

Many researchers have tried to detect the brain structures in the frontal lobe

that are responsible for cognitive control. There is evidence that the anterior cingulate

cortex (ACC) and the dorso-lateral prefrontal cortex (DLPFC; MacDonald, Cohen,

Stenger, & Carter, 2000) are related to conflict solving. It has been suggested that the

ACC is responsible for detecting the conflict, and the DLPFC is responsible for

dealing with it (Botvinick, Nystrom, Fissell, Carter, & Cohen, 1999; Botvinick,

Braver, Barch, Carter, & Cohen, 2001; Carter, Botvinick, & Cohen, 1999; Carter et

al., 1998; MacDonald et al., 2000). When considering inhibition, two brain regions

had been found to play a crucial role. The first is the right inferior frontal gyrus

(rIFG). Neuroimaging studies have suggested that the rIFG is involved in stopping a

response (e.g., Aron, Behrens, Smith, Frank, & Poldrack, 2007; Rubia, Smith,

Brammer, & Taylor, 2003). The rIFG shows increased activation when stopping is

successful, and the magnitude of the activation correlates negatively with the SSRT

(stop-signal reaction time; e.g., Aron et al., 2007; Aron & Poldrack, 2006; Rubia,

Smith, Taylor, & Brammer, 2007). The second region that is associated with

inhibition is the pre-supplementary motor area (pre-SMA; Aron et al., 2007). Some

researchers suggested that the pre-SMA is mainly involved in monitoring or resolving

the conflict between the go task and the stop task (which requires inhibition; e.g.,

Nachev, Wydell, O’Neill, Husain, & Kennard, 2007).

Task Conflict Control

The underlying idea of task conflict is that a stimulus may trigger performance

of a task that has acquired a strong association with it (Allport & Wylie, 2000; Rogers

& Monsell, 1995; Waszak, Hommel, & Allport, 2003). For example, a pen can trigger

4

a writing task, a telephone can trigger a dialing task, and words can trigger a reading

task (MacLeod & MacDonald, 2000). Thus, in many everyday situations, tasks that

are usually irrelevant are automatically triggered. If one wants to perform a specific

task, these automatically triggered irrelevant tasks can cause a cognitive conflict—a

task conflict between the relevant task and the automatic irrelevant task triggered by a

stimulus in the environment (Goldfarb & Henik, 2007; Haggard, 2008; Steinhauser &

Hubner, 2009). Monsell (2003) proposed that task sets can be activated either by

deliberate intentions that are governed by goals ("endogenous") or by the perception

of a stimulus attribute that is strongly associated with a particular task set

("exogenous"). Gibson (1979) suggested that people directly perceive what tools

afford in terms of meaningful actions—a feature of the object that clearly identifies

how the object could be used (e.g., grasp-ability of a stick, lift-ability of an object, or

click-ability of a light switch). Interestingly, Lhermitte, Pillon, and Serdaru (1986)

suggested that action-plans are evoked by stimuli and that certain frontal lobe patients

show a utilization behavior—an inability to suppress these action-plans and as a

result, perform tasks triggered by irrelevant stimuli (e.g., one patient that saw a

syringe tried to use it to inject the physician). Similarly, Cisek (2006) suggested that

multiple motor plans are generated automatically across visuo-motor regions of the

cortex in response to attended stimuli. Makris, Hadar, and Yarrow (2011) showed that

physical properties of objects automatically activate specific motor codes in the brain.

Furthermore, these researchers found that motor codes that are triggered by objects

are rapid and relatively short-lived. This fits with the suggestion that task control is

usually very efficient (Goldfarb & Henik, 2007; La Heij, Boelens, & Kuipers, 2010).

Aarts, Roelofs, and van Turennouta (2009) found a localized neural region in the

lateral prefrontal cortex that is associated specifically with task conflict.

Task conflict is commonly managed by task control, which is an efficient

control mechanism aimed at ensuring acting on the relevant task while inhibiting the

irrelevant automatic task (Goldfarb & Henik, 2007; La Heij et al., 2010). Braver's

(2012) dual mechanism of control framework suggests that the pro-active control unit

maintains constant amplification of the relevant task units and constant inhibition of

the irrelevant tasks (see also MacDonald et al., 2000). In contrast, in the absence of

effective pro-active control, the relevant task will not be strong enough (to be

5

processed without interference) and task conflict will be high, that is, interference

from conflicting tasks will be high.

The Stroop Task

A review of the literature reveals that over the last few decades many tasks

have been related to executive control. One very widespread laboratory task used to

study executive control is the Stroop task (Stroop, 1935). This task requires

participants to identify the color in which a color-word is printed, while ignoring the

word meaning. Since word reading is automatic a color-word can cause interference

when presented in incongruent color (e.g., GREEN written in red). In such cases, one

has to ignore the irrelevant word (GREEN) and respond instead to the color (red). The

Stroop effect (i.e., longer reaction time (RT) for incongruent than for congruent (e.g.,

RED written in red) stimuli), the interference effect (i.e., slower RT for incongruent

than for neutral (e.g., XXXX written in red) stimuli), and the facilitation effect (i.e.,

faster RT for congruent than for neutral stimuli) demonstrate that participants have

difficulty in ignoring the irrelevant word altogether (MacLeod, 1991). The low error

rate in normal participants and the increased error rate in participants with executive

and frontal deficits (Cohen & Servan-Schreiber, 1992) demonstrate the role of

executive control and of the prefrontal cortex in guiding task-relevant behavior and

suppressing automatic responses (Cohen, Dunbar, & McClelland, 1990).

Until recent years, most studies referred to the incongruent Stroop condition as

the only condition in the Stroop task that causes a conflict. Yet, a number of

neuroimaging studies showed that the ACC (as mentioned earlier, a brain area thought

to monitor conflict) is more active, not only when contrasting incongruent Stroop

trials to neutral trials, but also when contrasting congruent trials to neutral trials (e.g.,

Bench et al., 1993; Carter, Mintun, & Cohen, 1995; Roelofs, van Turennout & Coles,

2006; for older participants see Milham et al., 2002,). Thus there seems to be a

contradiction between the neuroimaging findings indicating conflict in congruent

trials and behavioral findings indicating that congruent trials are easier (and faster)

than neutral trials. Goldfarb and Henik (2007) suggested that one way to solve this

contradiction is to distinguish between two types of conflict in the Stroop task: the

information (or response) conflict, which is between the contradictory information

6

that arises from the word meaning and the information that arises from the word

color; and the task conflict, which is a conflict between the relevant task of color

identification and the irrelevant but automatic reading task. The information conflict

involves the content of the stimulus and the response needed, and differs for

congruent and incongruent Stroop stimuli, whereas the task conflict involves the task

associated with the stimulus and differs for congruent and neutral non-word Stroop

stimuli. Specifically, an intriguing Stroop reverse facilitation (slower RT to Stroop-

congruent compared to Stroop-neutral trials) can be observed. This supports the

presence of an additional conflict in the Stroop task―the task conflict. The idea is

that in both Stroop congruent and Stroop incongruent trials, but not in Stroop neutral

trials (when a meaningless letter string is used as a neutral), the participant faces a

task conflict (i.e., should they name the font color or read the word?). In the current

work I use the Stroop task as the main task used to investigate task conflict control.

Obsessive-Compulsive Disorder

OCD is a highly debilitating anxiety disorder with a lifetime prevalence rate of

2%-3% (Huppert, Simpson, Nissenson, Liebowitz, & Foa, 2009; Weisman et al.,

1994). OCD is characterized by recurrent intrusive thoughts or impulses (obsessions),

and repetitive, irresistible behaviors (compulsions) aimed at feared consequences and

to reduce anxiety (American Psychiatric Association, 2000). Compulsive behaviors

inflict paradoxical effects of increasing rather than decreasing the anxiety caused by

obsessions, effectively perpetuating compulsions (Salkovskis, 1999; van den Hout,

Engelhard, de Boer, du Bois, & Dek, 2008; van den Hout & Kindt, 2003).

Understanding factors affecting individual proneness to developing OCD is

paramount to improving OCD treatment, particularly since knowledge of etiological

factors underlying OCD is lacking (Gava et al., 2007; Grabill et al., 2008).

Both neurological and behavioral studies have found a deficit in executive

control in OCD patients (e.g., Abramovitch, Dar, Schweiger, & Hermesh, 2011;

Lucey et al., 1997; Meiran, Diamond, Toder, & Nemets, 2011; Penades, Catalan,

Andres, Salamero, & Gasto, 2005; Stein, 2002; for reviews see also Greisberg &

McKay, 2003; Kuelz, Hohagen, & Voderholzer, 2004). Some researchers suggested

executive control impairments are a core symptom of OCD (e.g., Anholt, Linkovski,

7

Kalanthroff, & Henik, 2012; de Wit et al., 2012; Muller & Roberts, 2005). The most

robust and stable differences between OCD patients and healthy controls were found

on tasks that required response inhibition (Bannon, Gonsalvez, Croft, & Boyce, 2002;

Penades et al., 2007) and some researchers proposed response inhibition deficit to be

an endophenotype of OCD (Chamberlain, Blackwell, Fineberg, Robbins, & Sahakian,

2005; Menzies et al., 2007).

As can be seen, the executive and inhibitory control approach for OCD seems

quiet promising. Nevertheless, there is an ongoing debate between approaches that

stress the inhibitory deficit and approaches that stress other cognitive symptoms of

OCD (for a review on the cognitive approach see Calkins, Berman, & Wilhelm, 2013;

and for a reply see Anholt & Kalanthroff, 2013). Specifically, the debate concentrates

on whether the executive deficits are indeed core deficits of OCD or are they

epiphenomena (Abramovitch, Dar, Schweiger, & Hermesh, 2012). One of the most

important downsides of the inhibitory deficit approach is that in general researchers

do not find correlations between executive deficits and symptom severity

(Abramovitch, Abramowitz, & Mittelman, 2013). Hence, finding such correlations are

of the essence in the behavioral approach.

In light of the executive and inhibitory deficit that was found in OCD patients

and in light of the debate about the importance of these deficits to understanding (and

treating) OCD, we aim to investigate task conflict control in OCD patients in order to

check whether or not this control mechanism can have a unique contribution to

understanding OCD. In this work I propose that a pro-active/task control deficit is a

potential link between poor inhibitory control and OCD. In other words, I propose

that OCD patients experience increased difficulties to inhibit normal, associatively

evoked tasks, which in turn, starts a vicious cycle of compulsive behaviors and

intrusive thoughts (Anholt et al., 2012). This suggestion will be elaborated in the final

part of this work.

The Current Work

In the current work we will examine the task conflict and its control

mechanism, alongside with suggesting some implications of our results, mainly

8

applying to OCD patients. In most studies that examine cognitive control using the

Stroop task, researchers focus on incongruent trials alone (commonly they do not

even use neutral or congruent trials). This work will mainly focus on Stroop

congruent trials as an indication for task conflict. This work is based on the findings

and suggestions made by Goldfarb and Henik (2007). Similar to Tzelgov, Henik, and

Berger (1992), Goldfarb and Henik used a version of the Stroop task in which 75% of

the trials were neutrals. This manipulation was found to “relax” the control. In order

to increase control relaxation, Goldfarb and Henik also included prime cues, in 50%

of the trials, indicating whether the upcoming trial was going to be a word trial

(congruent or incongruent) or a non-word trial (neutral). A reverse facilitation, and

indication for task conflict, was obtained when control was relaxed and not awakened

(the non-cued trials).

In Goldfarb and Henik's (2007) study there is no way to determine if the task

conflict arises as a consequence of participants' worrying about responding to the

wrong dimension (i.e., the written word), hence making mistakes. In Experiment

Group 1 we aimed to examine whether task conflict is independent from

informational conflict. To that end we partially replicated Goldfarb and Henik’s

experiment only we used no incongruent trials at all. The next question that the

current work addresses is what is the mechanism on which task control is based on.

Previous evidence in the literature indicated that task control may be based on

response inhibition control. After reviewing this evidence, in Experiment 2 we

investigated the correlation between task control, as expressed in Stroop reverse

facilitation, and response inhibition, as expressed in the stop-signal task (Verbruggen

& Logan, 2008). Next, in Experiment Group 3 we combined the Stroop and stop-

signal tasks and investigated the Stroop task conflict in momentarily inhibitory

control failure trials, as expressed in erroneous response in stop-signal trials.

As mentioned earlier, Braver (2012) suggested the term “pro-active control” to

describe the control mechanism that amplifies specific task demands for as long as

one engages the task at hand. As can be seen in De Pisapia and Braver’s (2006)

model, pro-active control is relies highly on inhibition. The current work, therefore,

suggests that pro-active control is needed in order to properly manage task conflict. In

Experiment Group 4 we aimed to investigate this assumption. In order to do that we

9

combined the Stroop and task switching paradigm so that the task would not be

constant and pro-active control would be low. This would allow investigating task

conflict under a low pro-active control condition. To summarize the results of these

experiments and to elaborate on our perception of the task control mechanism, I will

present a neural network computational model in the Discussion. Finally, based on the

evidence reviewed earlier about the connection between OCD and an inhibitory

control deficit, we will suggest a novel theory for the role of task control deficit in the

development and maintenance of OCD.

10

Chapter II

Experiment Group 1

Stop Interfering: Stroop Task Conflict Independence

from Informational Conflict and Interference

Eyal Kalanthroff, Liat Goldfarb, Marius Usher, & Avishai Henik

Quarterly Journal of Experimental Psychology

2013

Stop interfering: Stroop task conflict independence frominformational conflict and interference

Eyal Kalanthroff1, Liat Goldfarb2, Marius Usher3, and Avishai Henik1

1Department of Psychology and Zlotowski Center for Neuroscience, Ben-Gurion University of the Negev, BeerSheva, Israel2E. J. Safra Brain Research Center for Learning Disabilities, University of Haifa, Haifa, Israel3School of Psychology and School of Neuroscience, Tel-Aviv University, Tel-Aviv, Israel

Performance of the Stroop task reflects two conflicts—informational (between the incongruent word andink colour) and task (between relevant colour naming and irrelevant word reading). This is supported byfindings showing that the anterior cingulate cortex ismore activated by congruent and incongruent stimulithan by nonword neutral stimuli. Previously, researchers demonstrated behavioural evidence for task con-flict—a reverse facilitation effect under a reduced task conflict control condition.The boundary conditionsof this Stroop reverse facilitation effect are not yet clear. The current study aimed to investigate whethertask conflict arises, and task control is needed, whenever there are two possible tasks, even if the irrelevanttask cannotmislead one to give erroneous responses (i.e., stimuli do not contain an informational conflict).To this end, in both experiments no incongruent stimuli were presented. In Experiment 1, participantsconducted a Stroop task with a high proportion of nonword neutrals and with a neutral/congruent cuein 50% of the trials. In Experiment 2, the nonword neutral was replaced by a real non-colour-word.We found the reverse facilitation effect in the noncued trials of Experiment 1. Moreover, as expected,this effect was eliminated when a noncolour neutral word that induced task conflict was used(Experiment 2). We conclude that task conflict control is reactively activated whenever there are atleast two possible tasks, even in the absence of any possibility of informational conflict.

Keywords: Task conflict; Stroop; Reverse facilitation; Executive functions; Anterior cingulate cortex.

Executive control is a key human function thatmediates the ability to guide behaviour in accord-ance with internal goals (Banich, 2009; Miller &Cohen, 2001; Miyake et al., 2000; Shallice &Norman, 1986). An important part of thisprocess, and perhaps a hallmark of executive func-tion, is the suppression of irrelevant information(Verbruggen & Logan, 2008). This is needed in

daily situations that go against routine. To studythis process in the laboratory, consider the Strooptask (Stroop, 1935), which requires participants toidentify the colour in which a colour-word isprinted, while ignoring the word meaning. Sinceword-reading is automatic, this presents partici-pants with a challenge when presented with incon-gruent stimuli (e.g., GREEN written in red). In

Correspondence should be addressed to Eyal Kalanthroff, Department of Psychology, Ben-Gurion University of the Negev,

P.O.B. 653, Beer Sheva, Israel 84105. E-mail: [email protected]

We thank Ms. Desiree Meloul for helpful comments and useful input on this article.

1356 # 2013 The Experimental Psychology Society

THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013

Vol. 66, No. 7, 1356–1367, http://dx.doi.org/10.1080/17470218.2012.741606

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

8:31

03

July

201

3

11

such cases, one has to ignore the irrelevant word(GREEN) and respond instead to the colour(red). The Stroop effect (i.e., longer reactiontime, RT, for incongruent than for congruentstimuli; e.g., RED written in red) and interferenceeffects (i.e., slower RT for incongruent than forneutral stimuli; e.g., XXXX written in red) demon-strate that participants have difficulty in ignoringthe irrelevant word altogether. The low error ratein normal participants and the increased error ratein participants with executive and frontal deficits(Cohen & Servan-Schreiber, 1992) demonstratethe role of executive control and of the prefrontalcortex in guiding task-relevant behaviour and sup-pressing automatic responses (Cohen, Dunbar, &McClelland, 1990).

A number of investigations have suggested thatexecutive control is allocated in an adaptive andstrategic way in response to conflict (Botvinick,Braver, Barch, Carter, & Cohen, 2001; DePisapia & Braver, 2006). This means that partici-pants do not allocate executive control equally inall Stroop trials; rather, control relaxes after easy(neutral) stimuli, but is engaged by difficult, con-flict-inducing (incongruent) stimuli. This sugges-tion is supported by data showing that themagnitude of the Stroop effect increases in blocksthat have a majority of neutrals (e.g., 12.5% incon-gruent, 12.5% congruent, and 75% neutrals), com-pared to blocks that have a majority of colour-wordStroop stimuli (e.g., 37.5% incongruent, 37.5%congruent, and 25% neutral; Tzelgov, Henik, &Berger, 1992). As suggested by Tzelgov et al.(1992), and consistent with a role of conflict inthe allocation of executive control in the Strooptask, an increase in the frequency of neutrals mayresult in putting executive control to sleep, resultingin enhanced Stroop interference. This interpret-ation is supported by neuroimaging data thatprovide evidence for the role of conflict in top-down control regulation by showing that theanterior cingulate cortex (ACC)—a brain areathought to monitor conflict (Botvinick et al.,2001; Botvinick, Nystrom, Fissell, Carter, &Cohen, 1999; Carter, Botvinick, & Cohen, 1999;Carter et al., 1998)—is more active in incongruentStroop trials than in neutral trials (e.g., Bench et al.,

1993; Carter, Mintun, & Cohen, 1995; Milhamet al., 2002, for older participants). Interestingly,however, these studies also show that not only dothe incongruent Stroop stimuli trigger higherACC activations than the neutrals do, but the con-gruent Stroop trials do so too. This may be surpris-ing since on the face of it, responding to suchstimuli should be easy, as indicated by facilitationeffects (faster response to congruent than toneutral Stroop stimuli; MacLeod, 1991). Onemay note, though, that Stroop facilitation effectsare generally smaller and less robust than Stroopinterference effects (e.g., Dalrymple-Alford &Budayr, 1966).

The nature of the contradiction between theneuroimaging and the behavioural data on conflictand putative facilitation in Stroop congruent trialshas been subject to a number of recent investi-gations. One way to resolve this apparent contra-diction is to distinguish between two types ofconflict: information (or response; i.e., a conflictbetween the contradictory information that arisesfrom the word meaning and the information thatarises from the word colour) versus task conflict(i.e., a conflict between the relevant identificationof the colour task and the irrelevant but automaticreading task). While information conflict involvesthe content of the stimulus and the responseneeded, differing between congruent and incongru-ent Stroop stimuli, task conflict involves the taskassociated with the stimulus and differs betweencongruent and neutral nonword Stroop stimuli(Kalanthroff, Goldfarb, & Henik, in press). Thisis consistent with the suggestion that a stimulushas the ability to evoke performance of a task thathas a strong association with it (Allport & Wylie,2000; Rogers & Monsell, 1995; Waszak,Hommel, & Allport, 2003) and, specifically, thatwords trigger an automatic tendency to readwritten words (MacLeod & MacDonald, 2000).Such task conflict may arise in particular in situ-ations in which the proactive top-down controlmechanism (Braver, 2012) is diminished. In thesesituations, the stimulus may trigger a reactiveresponse of the associated task demands.

Recent research in our lab provided furthersupport for the presence of task conflict in

THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (7) 1357

TASK CONFLICT INDEPENDENCY

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

8:31

03

July

201

3

12

congruent Stroop stimuli by demonstrating a novelreverse facilitation effect (i.e., faster response tononword neutrals, e.g., XXXX written in red,than to congruent Stroop stimuli, e.g., REDwritten in red). To show this, Goldfarb andHenik (2007) used a high-neutral-nonword-fre-quency, low-control condition, as a manipulationaimed to put the task conflict guard to sleep. Tofurther reduce proactive task control, they used acueing procedure in which in half of the trials avalid cue indicated whether the upcoming trialwas going to be a neutral or a colour-word trial.This allowed participants to relax the proactivecontrol progressively further as they got used torelying on cues for activating it. The results led toa slow-down in RTs for congruent compared tononword neutral trials—that is, a significantreverse facilitation in the noncued trials. Becausethe large proportion of neutral trials and cueingwere the cause for the changes in RTs in congruenttrials, these results provided behavioural evidencefor task conflict between responding to the wordreading task and responding to the colour namingtask. The reverse facilitation revealed that the taskconflict is present in congruent Stroop trials inthe absence of proactive control (noncued con-dition), in line with the neuroimaging evidence(e.g., Bench et al., 1993; Carter et al., 1995).

One important question arises, however, regard-ing the boundary conditions of the Stroop reversefacilitation, which has implications for the natureof the task conflict slow-down. Is this slow-downan adaptive control process, meant to protect theparticipant from the risk of making a mistake inincongruent Stroop trials (half of the colour-wordtrials were incongruent in all the studies above)?Or, is the task conflict slow-down a mandatoryand automatic process that is triggered by the pres-ence of any word-like stimuli, even if no danger formisleading information exists, as long as a compet-ing task exists (e.g., a word in the Stroop task)? Ifthe latter is the case, we should expect that theStroop reverse facilitation be maintained, even ifthere are no incongruent Stroop trials at all. It hasbeen suggested that the Stroop effect is sensitiveto context (Melara & Algom, 2003; Sabri,Melara, & Algom, 2001). Will there be an

indication for task conflict when incongruenttrials are nonexistent? Accordingly, the aim ofthis study is to test the boundary condition of theStroop reverse facilitation. Will participants haveslowed-down responses in congruent Stroop trialseven though the irrelevant pathway will activatethe same response, and no potential danger willbe present? Such a nonadaptive outcome is consist-ent with studies indicating that the Stroop taskconflict is independent of informational conflictand that the reading task is automatically activatedwhenever there is a real word presented. Forexample, in a neuroimaging Stroop study, Benchet al. (1993) presented stimuli in pure blocks ofcongruent or neutral (i.e., letter string) trials.Even though they used pure blocks, they stillfound larger ACC activation for a pure congruentblock than for a pure neutral block, indicatingthat task control is not adaptive and that task con-flict arises even if there is no danger for incongruentinformation. Bench et al.’s findings are restricted tothe neural level since at the behavioural level theyfound a common (small) facilitation. In thecurrent study, we aim to test this at the behaviourallevel.

EXPERIMENT 1

Our study is based on the experimental design usedby Goldfarb and Henik (2007). In their study, 75%of the trials were neutral, 12.5% congruent, and12.5% incongruent. In half of the trials, subjectswere presented a cue telling them whether theupcoming trial was going to be neutral or not.The automatic response to the word (reading)could potentially cause an erroneous response;hence, it was worthwhile inhibiting this irrelevanttask.

In this case, performing the irrelevant task canlead to some “bad consequences”. Efforts investedin inhibiting the automatic task (or managing thetask conflict) slowed the RTs for congruent trialsand caused a reverse facilitation effect. But, whatwould happen had the participants learned thatthe competing task could never mislead them? Inthis experiment, we replicated the procedure from

1358 THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2013, 66 (7)

KALANTHROFF ET AL.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

8:31

03

July

201

3

13

our previous study (Goldfarb & Henik, 2007), onlywith no incongruent trials. This would allow us toanswer the question of whether task conflict isautomatically triggered whenever a written wordappears or whether control is adaptive and stopsmanaging that conflict if it learned that it is notworthwhile.

Since we believe that task conflict exists when-ever there are (at least) two possible tasks and thatthis conflict will appear whenever proactivecontrol is low, we expect to replicate Goldfarband Henik’s (2007) finding. That is, in a highneutral nonword condition, proactive task conflictcontrol is expected to be relaxed, and therefore innoncued trials, task conflict should be present, asindicated by a reverse facilitation. In the samehigh neutral nonword condition, in cued trialscontrol can be easily and efficiently recruited, and,thus, task conflict is less likely to appear; hence,we do not expect a reverse facilitation.

Unlike in Goldfarb and Henik’s (2007) study,the cue in our experiment revealed whether theupcoming trial was going to be neutral or congru-ent. We address the potential problems that thiscould cause as well as our suggested solutions inthe Discussion section of the current experimentand in Experiment 2.

Method

ParticipantsNineteen first-year psychology students (15 femalesand 4 males) of Ben-Gurion University of theNegev (Israel) participated for partial fulfilment ofcourse requirements and credit. All participantshad normal or corrected-to-normal vision, wereright-handed, had no history of attention deficitor dyslexia, and were native speakers of Hebrew,and all were naive as to the purpose of the exper-iment. One participant (male) was excluded fromfurther analysis due to extremely slow RT(average RTs were more than 3 standard deviationsfrom the mean in all conditions).

StimuliParticipants were presented with a four-colour,manual Stroop task. Each stimulus consisted of

one of four colour-words (blue, red, green, andyellow), or a four-letter string in Hebrew—

(meaningless repetition of a Hebrewletter, parallel to XXXX in the English version).All four colour-words have four letters inHebrew. The ink colour was red, blue, green, oryellow. There were four different congruent combi-nations of words and ink colours, and four differentneutral stimuli. An “X” cue meant that a congruenttrial was going to appear, an “O” cue meant that aneutral letter string was going to appear, and a “?”cue meant that any stimulus could appear. Cueswere 100% valid. Seventy-five percent of the trialswere neutral (37.5% cued and 37.5% noncued),and 25% were congruent (12.5% cued and 12.5%noncued). The words were presented at the centreof a screen on a black background and were 0.98inches high and 2.36 inches wide. Cues were pre-sented at the centre of the screen.

ProcedureFor most accounts the procedure was similar to thatin Goldfarb and Henik’s (2007) study. Data collec-tion and stimuli presentation were controlled by aDELL OptiPlex 760 vPro computer with anIntel core 2 duo processor E8400 3 GHz. Stimuliwere presented on a DELL E198PF 19′′ LCDmonitor. A keyboard was placed on a tablebetween the participant and the monitor.Participants were tested individually. They satapproximately 23.5′′ from the computer screen.Coloured stickers were taped on four regular key-board keys according to the colours they rep-resented. The “v” key represented red, the “b” keyrepresented blue, the “n” key represented green,and the “m” key represented yellow. Participantswere asked to use only their right hand.

The experiment started with 12 key-matchingpractice trials. In these trials, a coloured asteriskappeared at the centre of the screen, three timesin each of the four colours, in a random order.The asterisk disappeared on a key-press or after3,500 ms, and then the next trial began after a1,000-ms interval. Feedback was given only forincorrect trials.

Following this, participants were presented withan explanation about the task and the meaning of



Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

8:31

03

July

201

3

14

the cues. Instructions emphasized that after an “O”

cue a letter string would appear, after an “X” cue acolour-word would appear, and after a “?” cue eitherone of the two could appear. Participants wereasked to refer only to the ink colour and to pressthe correct key as fast as possible without makingmistakes. Instructions did not reveal that therewould be no incongruent trials or the proportionof the congruency conditions. Nevertheless, partici-pants were told that the practice block would beidentical to the experimental block (only that theexperimental block would be longer and wouldnot include feedback); so in fact, participantsknew there would be no incongruent trials afterthe practice block. Subjects underwent 32 practicetrials (which were not analysed further) and 192experimental trials. In both the experiment blockand the practice block, 25% of the trials were con-gruent, and 75% were neutral. Half of the trials ineach congruency condition were cued (i.e., “X” or“O” preceded the target), and half were not (“?” pre-ceded the target). Within each condition, thestimuli were presented equally. The order of thetrials was random.

Each trial started with a 1,000-ms fixation (awhite+ sign at the centre of a black screen), fol-lowed by a 500-ms interval of an empty black

screen. Then the cue appeared for 1,000 ms, fol-lowed by another 500-ms interval of an emptyblack screen. After this the Stroop target appearedand stayed in view for 2,500 ms or until a key-press. RT was calculated from the appearance ofthe target Stroop stimulus to the reaction. Eachtrial ended with a 1,000-ms intertrial interval.

Results and discussion

Mean RTs of correct responses were calculated foreach participant in each condition. A two-wayANOVA (analysis of variance) with repeatedmeasures was applied to RT data with congruency(congruent and neutral) and cue (cued ornoncued) as within-subject factors. A significantinteraction was found, F(1, 17)= 7.616, MSE=1,041.829, p, .01, η2p= .309. In addition, wefound a main effect for cue, F(1, 17)= 6.420,MSE= 876.74, p, .05, η2p= .274, but no effectfor congruency, F(1, 11), 1. We found that RTfor noncued congruent stimuli (633 ms) was 39ms longer than that for cued congruent stimuli(594 ms), t(17)= 3.161, p, .01. On the otherhand, there was no significant difference in RTbetween the cued (622 ms) and the noncued (619ms) neutral trials t(17), 1 (Figure 1).

To further investigate the interaction, a plannedpost hoc analysis was carried out. We analysed thesimple effect of congruency in the two cueing con-ditions. For the cued condition, we found a mar-ginally significant facilitation, F(1, 17)= 3.631,MSE= 1,856.185, p= .074. For the noncued con-dition, we found a significant reverse facilitation of15 ms, F(1, 17)= 5.46, MSE= 352.39, p, .05.Note that two different patterns of Stroop effectswere found within the same experiment. Areverse facilitation (RTs for neutral trials shorterthan RTs for congruent trials) was found in thenoncued condition, and the regular facilitationeffect (RTs for neutral trials longer than RTs forcongruent trials) was found when a cue was given(see Figure 1). Reverse facilitation in the noncuedcondition, which occurred without any incongruenttrials, indicates that task conflict arises automati-cally whenever a word is presented, regardless ofthe information it contains. Moreover, the reverse

Figure 1. Mean RT (reaction time) and error rates (in parentheses)

in the congruency conditions for Stroop trials with and without

cueing in Experiment 1. Error bars represent one standard error

from the mean (using Cousineau’s, 2005, method to compute the

error bars in within-subjects designs).


KALANTHROFF ET AL.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

8:31

03

July

201

3

15

facilitation effect also indicates that task conflictcontrol is needed, and it slows down congruentRT responses, even when there are no incongruenttrials, or when the competing task cannot misleadyou. The lack of a reverse facilitation in the cued(higher control) condition also supports theinterpretation that the task conflict effect is contin-gent upon low proactive control.

Using an ex-Gaussian analysis, Steinhauser andHubner (2009) showed that a task conflict effect isobtained in the exponential (slow) portion of theRT distribution, while response conflict is obtainedin the Gaussian (fast) portion of the RT distri-bution. In light of this, one can argue that task con-flict disappeared in cued trials only due to speedinginduced by the cue, and not due to increased proac-tive control. Though fast RTs and high controlexplanations do not necessarily contradict eachother, we conducted an additional analysis inorder to investigate this alternative explanation.Mean RTs (of correct responses) were calculatedfor congruent and neutral cued conditions onlyfor the slowest quarter of the trials for each partici-pant. Similar to the general RT analysis, we found amarginally significant common facilitation effect(62 ms), F(1, 17)= 4.245, MSE= 8,051.369,p= .055, indicating that the facilitation effect isnot an artefact of generally fast RTs in the cuedcondition.

The most important finding of this experimentis that our results in the noncued conditionreplicate those of Goldfarb and Henik (2007)in a Stroop task without incongruent trials.In the cued condition, however, we found a mar-ginally significant facilitation, whereas theyfound a marginally significant reverse facilitation.Furthermore, the magnitude of the reverse-facili-tation effect in the noncued condition (about 15ms) was smaller than the one found in Goldfarband Henik’s study (more than 100 ms). Althoughit is difficult to draw firm conclusions from themagnitude of the facilitation in different studies,one possibility is that the effect of the task con-flict—the amount of inhibition it triggers on theresponse—depends to some degree on the corre-lation between task conflict and errors. While inGoldfarb and Henik’s study task-conflict trials

were those in which there was a danger to makean erroneous response (half were incongruent), inthe present experiment, this was not the case (alltask-conflict trials were congruent). Thus, it ispossible that while task conflict is automatic, theamount of response inhibition it triggers dependson task contingencies (Melara & Algom, 2003;Sabri et al., 2001). We defer this question tofuture studies that vary the frequency of the incon-gruent stimuli in a within-subject design. Theimportant result of our present study is thatStroop reverse facilitation emerges even under astringent condition in which the congruent wordsdo not pose any danger towards naming errors.

A possible alternative explanation for the resultin the cued condition is that in the present study,participants switched to an explicit reading strategyin congruent trials. This could explain the fasterRTs in the cued congruent trials. In order toreject this alternative explanation, we conductedan additional analysis. Allport, Styles, and Hsieh(1994) showed that there is a large switching cost(i.e., longer RTs to switch trials than to repeattrials) when, in the Stroop task, participantsswitched from a colour naming task to a wordreading task. Because we wanted to rule out thepossibility that participants responded to the wordin the cued congruent trials, we looked for a switch-ing (to word response) cost. If participantsresponded to the word, it is most likely that theydid it in cued congruent trials, whereas in theneutral trials they surely responded to colours;thus, responding to a word on trial n would belonger if the preceding trial n – 1 was a neutralthan when it was a word. Mean RTs (of correctresponses) in the cued congruent condition werecalculated separately for trials that were precededby a cued congruent condition (repeat) and fortrials that were preceded by neutrals (supposedlyswitch trials). No significant difference was foundbetween “repeat” trials (589 ms) and “switch”trials (595 ms), F, 1 (the observed statisticalpower for this null effect was .06), indicating noswitching cost (hence, no switching). Note thatprevious studies argued that when switching fromreading to colour naming, there is no (or muchless) switching cost (Allport et al., 1994); thus,



Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

8:31

03

July

201

3

16

we calculated “switching cost” caused by (possibly)switching to a word response.

We further discuss the possibility that partici-pants responded to the word on cued congruenttrials in the Discussion of Experiment 2.

EXPERIMENT 2

Experiment 2 was conducted in order to verifyfurther that nothing but the level of task conflictcontrol produced the effects observed inExperiment 1. In order to do this, we used amanipulation used in our previous studies(Goldfarb & Henik, 2007, Experiment 2;Kalanthroff et al., in press, Experiments 2 and 3).We replaced the neutral homogenous letterstring stimulus in Experiment 1 with a non-colour-word. This would still allow for no informa-tional conflict in the neutral condition. Using theStroop task, Bugg, McDaniel, Scullin, and Braver(2011) found a decreased interference effect(increased proactive control) in a high-neutral-word-proportion block. This suggests that neutralwords cause a conflict and require executivecontrol. According to MacLeod and MacDonald(2000), words raise an automatic tendency toread; thus a non-colour-word should activate anautomatic word reading response (the irrelevanttask), whereas a string of letters should not.Goldfarb and Henik (2007), and Kalanthroffet al. (in press) found that the reverse facilitationeffect indeed disappeared when a non-colour-word was used as a neutral. Accordingly, underthis condition, task conflict should occur in alltrials, and its effect should not be seen in any trial—task conflict control will be high in both cuedand noncued conditions, and as in Goldfarb andHenik’s (Experiment 2) and Kalanthroff et al.’sstudies (Experiments 2 and 3), no task conflicteffects will be revealed.

As we noted in Experiment 1, the cue effectcould alternatively be attributed to the participants’strategy in which they switch to a reading task inthe congruent cued trials. Experiment 2’s designstill allows participants to carry out this strategy.Theoretically, they can switch to a reading task in

the congruent cued trials regardless of the kind ofneutral trials used in the design. If participantsused such a strategy, and the effect found in thecued condition of Experiment 1 had nothing todo with the task conflict, then the same patternof results as those found in Experiment1 wouldbe found in Experiment 2.

Because we believe that task conflict will nowappear in both congruent and neutral trials, wepredict that a common facilitation will appear inthe noncued condition. Moreover, because webelieve that participants follow the instructionsand do not read (as also evidenced from the lackof switching cost in Experiment 1), we predictthat in the cued condition the facilitation effectwill not be larger than that in the noncuedcondition.

Method

ParticipantsSixteen first-year Psychology students (12 femalesand 4 males) of Ben-Gurion University of theNegev (Israel), who did not take part inExperiment 1, participated for partial fulfilmentof course requirements and credit. All participantshad normal or corrected-to-normal vision, wereright-handed, had no history of attention deficitor dyslexia, and were native speakers of Hebrew,and all were naive as to the purpose of theexperiment.

StimuliIn this experiment, the meaningless homogenousletter string was replaced by the Hebrewnon-colour-word (building), which matchedthe colour-words by length and word frequencyand did not begin with the same letter as any ofthe colour-words (see also Goldfarb & Henik,2007). Except for this, all the stimuli and the pro-cedure in Experiment 2 were identical to those inExperiment 1.

Results and discussion

Mean RTs of correct responses were calculated foreach participant in each condition. A two-way


KALANTHROFF ET AL.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

8:31

03

July

201

3

17

ANOVA with repeated measures was applied toRT data with congruency (congruent and neutral)and cue (cued or noncued) as within-subjectfactors. As expected, a main effect for congruencywas found, F(1, 15)= 6.948, MSE= 1,580.698,p, .05, η2p= .317, and no main effect for cueing,F, 1, or interaction, F, 1, was found. Theobserved statistical powers for those null effectswere .053 and .063 for the main effect for cueingand for the interaction, respectively. As can beseen in Figure 2, when using a neutral word,response to the cued and noncued conditions didnot differ. It seems that under a high-control con-dition, caused by the continuous encounters withthe need to manage the task conflict, task controlwas very efficient in quickly solving the taskconflict.

Planned comparisons showed that in both thecued and the noncued conditions, RT for congru-ent words was significantly shorter (611 ms and615 ms for the cued and noncued condition,respectively) than RT for neutral non-colour-words (640 ms and 639 ms for the cued andnoncued conditions, respectively)—the facilitationeffect, F(1, 15)= 4.933, MSE= 1,334.38, p ,.05, and F(1, 15)= 4.537, MSE= 991.377, p ,.05, respectively (Figure 2). Unlike in Experiment

1, task conflict control was high, and thus acommon facilitation was demonstrated. It seemsthat word neutrals increase task conflict controlbecause task control is needed in order to preventresponding to words.

Importantly, Experiment 2 strengthens theassumption that results obtained in Experiment 1were not caused by reading, or switching toreading, in the congruent trials. There are two indi-cations that participants did not respond to theword in the congruent cued condition: (a) Ifreading had taken place in the congruent cued con-dition, participants would have used the same strat-egy in the parallel condition in Experiment2. Hence, the facilitation effect should have beenhigher (or the RTs in congruent trials would havebeen shorter) in the cued condition than in thenoncued condition in this experiment. Moreover,if switching from identifying the colour toreading was the cause for the reverse facilitationeffect obtained in Experiment 1, we would haveexpected the same pattern of results inExperiment 2. The fact that the facilitation effectwas positive and similar in both conditions (30ms and 24 ms for cued and noncued condition,respectively) and that the interaction between con-gruency and cueing was not significant, does notsupport this option. (b) If participants switched toreading in the congruent cued condition, wewould have expected a switching cost (Allportet al., 1994) in Experiment 1; no such switchingcost was found.

Another important finding of Experiment 2 wasthat there was no main effect for cueing, meaningthat the cue did not help to speed up responding.We suggested earlier that conflict arose whenevera word was presented. Accordingly, participantsknew that a conflict was expected in each andevery trial. Thus, the cue was not informativeregarding the upcoming conflict. Additionally,since a conflict was presented in all trials (unlikeExperiment 1 in which conflicts were rare), ourresults indicate that proactive control was highthroughout the block. In Experiment 1, proactivecontrol was relaxed, and the existence of cues wea-kened it even more. In Experiment 2, control wasalready high, and a cue did not make a difference.

Figure 2. Mean RT (reaction time) and error rates (in parentheses)

in the congruency conditions for Stroop trials with and without

cueing in Experiment 2. Error bars represent one standard error

from the mean (using Cousineau’s, 2005, method to compute the

error bars in within-subjects designs).



Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

8:31

03

July

201

3

18

Moreover, the idea that a high proportion of wordneutrals increases proactive control is strengthenedby Bugg et al.’s (2011) study that was mentionedearlier.

GENERAL DISCUSSION

The main result of this investigation is the replica-tion of the Stroop reverse facilitation (faster RT toneutral stimuli than to congruent Stroop colour-words; Goldfarb & Henik, 2007), in a study thatdoes not include any incongruent Stroop stimuli.This supports the idea that congruent Stroopstimuli trigger task conflict—between the relevanttask and the irrelevant task that is triggered by astrong association with a stimulus—automatically.

These results are consistent with the idea that atleast part of the interference that is commonlyobserved in Stroop tasks is due to competitionbetween the relevant task and the irrelevant auto-matic task (Monsell, Taylor, & Murphy, 2001);they are also consistent with the results of morerecent studies on task conflict in the Stroop task.For example, David et al. (2011) found evidencefrom an event-related potential (ERP) study forthe existence of task conflict in the Stroop match-ing task; Steinhauser and Hubner (2009) showedan empirical dissociation between task conflictand response conflict in the Stroop task byshowing that the task conflict effect is obtained inthe exponential portion of the RT distribution,while response conflict is obtained in theGaussian portion of the RT distribution; andfinally, Bugg et al. (2011) showed decreasedStroop interference (evidence for increased proac-tive control) in a high-neutral-word-proportioncondition, similar to that found in a high-incongru-ent-word-proportion condition—another indi-cation that a written word causes controlrecruitment. This is consistent with the results ofExperiment 2 of the current study in which thecue did not help participants to improve perform-ance. In line with Bugg and colleagues, wesuggest that a high proportion of word-neutraltrials increases proactive control throughout theblock. Evidence for task conflict was found in a

few studies on task switching, a paradigm thatincreases load on proactive task maintenancecontrol (e.g., Aarts, Roelofs, & van Turennouta,2009; Braverman & Meiran, 2010). Thus, theexistence of task conflict is supported by evidencefrom a number of converging paradigms.

In the Stroop task, the task conflict, unlike theinformational conflict, exists in both congruentand incongruent trials. Goldfarb and Henik(2007) argued that a very efficient task controlmechanism prevents the behavioural expression ofthe task conflict, which becomes visible underlow-control conditions. In a recent study,Kalanthroff et al. (in press) combined the Strooptask with the stop-signal method (Logan, 1994;Logan & Cowan, 1984). Participants were pre-sented with Stroop stimuli, but in some of thetrials they were presented with a stop signal, requir-ing the inhibition of a response. If inhibition of thestop signal and inhibition of irrelevant (but auto-matic) pathways in the Stroop task involve acommon process (Miyake et al., 2000), we shouldexpect the two types of inhibitory control to be cor-related. That is, when participants respond in spiteof the stop signal (low inhibitory control), stoppingfails, and the proactive Stroop control is diminishedalso. Indeed, we found that erroneous responses tostop-signal trials showed slower RTs to congruentthan to neutral nonword Stroop trials—a reversefacilitation. Consistent with this, in an additionalindividual differences study (Kalanthroff &Henik, 2012), we found significant correlationsbetween the Stroop facilitation effect and stop-signal reaction time (SSRT). These results(Stroop reverse facilitation in the absence of proac-tive task control) are further supported by develop-mental studies with Stroop-like tasks, whichreported behavioural evidence for task conflict inchildren (La Heij & Boelens, 2011; La Heij,Boelens, & Kuipers, 2010; see Discussion section).

Crucially, all of these studies used incompatibletrials so that acting on the competing task couldpotentially lead one to an incorrect response.Hence, it was impossible to determine whethertask control depended on the existence of incon-gruent trials in the design—namely, whether itemerged automatically, even in the case where no


KALANTHROFF ET AL.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

8:31

03

July

201

3

19

danger for incongruent information existed. Incontrast, the current work shows that task conflictappears automatically, even when the competingtask (i.e., reading) could not lead to errors.

The current findings strengthen the claim thatstimuli have the ability to evoke the performanceof a task that has a strong association with them(Allport & Wylie, 2000; Rogers & Monsell,1995; Waszak et al., 2003). In our previous work(Kalanthroff et al., in press), which was mentionedearlier, we found a reverse facilitation effect in erro-neous response to stop-signal trials. Interestingly,RTs for incongruent trials were not longer thanRTs for congruent trials in these erroneousresponses. We concluded that at least some dis-sociation between two separate control mechanismsexists; the first control mechanism is responsible formanaging the task conflict and is recruited when atask conflict appears; the second is responsible forsolving the informational conflict and is recruitedwhen facing stimuli that produce incongruentinformation. Our previous findings showed thatfailure in the task control mechanism is not necess-arily followed by a failure in the informationalcontrol mechanism. Results in the current studystrengthen this by showing that task conflictoccurs even if there is no potential for informationalconflict.

Even though we basically replicated the mainfindings of Goldfarb and Henik (2007) in whicha significant reverse facilitation was observed inthe noncued condition, there were still some differ-ences between the current study results (in whichincongruent trials were omitted) and the previousones (in which incongruent trials were present)regarding the cued condition. In the currentstudy, in both experiments a common facilitationeffect was obtained under the cued condition(though it was only marginally significant inExperiment 1), whereas in Goldfarb and Henik’sstudy, under the cued condition there was no facili-tation effect but rather a nonsignificant reversefacilitation. As the magnitude of the reverse facili-tation effect was smaller than that obtained instudies that mixed incongruent stimuli, there is apossibility that the magnitude of the response inhi-bition triggered by the task conflict depends on the

relation between task conflict and error-prone(incongruent) responses. This might imply thattask control is adaptive to some extent; whenthere are no incongruent trials, there is less taskconflict, slowing down the RTs for congruenttrials less than when incongruent trials exist. Anadaptation might occur after subjects learn thatno incongruent trials exist, and, thus, the readingtask interferes less. In the cued condition, taskcontrol is not relaxed enough, and thus, its (low)levels are sufficient to prevent the reversefacilitation.

Though the current study implies that task con-flict emerges in congruent trials, it seems that thesystem can detect that the competing task is lessdangerous, and, hence, less control is needed. Ittakes a more “relaxing effort” of proactive controlin order to reveal the task conflict (or reverse facili-tation). Importantly, our design does not allow usto draw unequivocal conclusions about the exist-ence and characteristics of a subtle adaptationmechanism (Melara & Algom, 2003). Futureresearch that varies the proportion of incongruenttrials in a within-subject design is needed tofurther examine this issue. Over all, we believethat the facilitation effect in the cued conditionsupports our suggestion that task control is contin-gent upon low (noncued condition) proactivecontrol.

Braver and colleagues (Braver, 2012; De Pisapia& Braver, 2006) suggested two modes of cognitivecontrol: proactive control, which “reflects the sus-tained and anticipatory maintenance of goal-relevant information”, and reactive control, which“reflects transient stimulus-driven goal reactivation. . . based on interference demands or episodicassociations” (Braver, 2012, p. 106). When the pro-portion of nonword neutrals is high, proactivecontrol is likely to be less active due to long-termadaptation. It is reasonable to assume that low acti-vation of proactive control is the cause for theappearance of behavioural evidence for the taskconflict. Our finding seems to imply that taskcontrol is reactive (when the frequency ofnonword neutrals is high, and cues are used towarn of upcoming Stroop stimuli) and is recruitedwhenever task conflict surprises subjects in a



Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

8:31

03

July

201

3

20

specific trial. However, when proactive control ishigh, the task conflict is managed very fast;hence, there is almost no behavioural evidence forit. Nevertheless, we also found indications forlong-term adaptations; hence, we suspect thatproactive control influences the activation of reac-tive control. The model proposed by De Pisapiaand Braver (2006) accounts for this specific notion.

To conclude, we found that task conflict occurs,and that task control is needed, whenever there areat least two possible tasks. This is the case even ifthere is no possible informational conflict andeven if the competing tasks cannot mislead one toan erroneous response. Though task conflictoccurs even if there are no incongruent trials, webelieve there is lower task conflict under this con-dition than where there are incongruent trials.This could possibly be explained by the commonactivation of the ACC by the task conflict andinformational conflict; activation of this conflictdetector mechanism is lower when there is noinformational conflict. These results indicate thattask conflict, together with the informational con-flict, mediates Stroop performance.

Original manuscript received 3 June 2012

Accepted revision received 22 August 2012

First published online 20 November 2012

REFERENCES

Aarts, E., Roelofs, A., & van Turennouta, M. (2009).Attentional control of task and response in lateraland medial frontal cortex: Brain activity and reactiontime distributions. Neuropsychologia, 47, 2089–2099.

Allport, D. A., Styles, E. A., & Hsieh, S. (1994).Shifting intentional set: Exploring the dynamiccontrol of tasks. In C. Umiltà & M. Moscovitch(Eds.), Attention and performance XV (pp. 421–452).Hillsdale, NJ: Lawrence Erlbaum Associates.

Allport, A., &Wylie, G. (2000). Task-switching, stimu-lus–response bindings and negative priming. InS. Monsell & J. Driver (Eds.), Control of cognitive pro-cesses: Attention and performance XVIII (pp. 35–70).Cambridge, MA: MIT Press.

Banich, M. T. (2009). Executive function: The search foran integrated account. Current Directions in

Psychological Science, 18, 89–94.Bench, C. J., Frith, C. D., Grasby, P. M., Friston, K. J.,

Paulesu, E., Frackowiak, R. S., et al. (1993).Investigations of the functional anatomy ofattention using the Stroop test. Neuropsychologia, 32,907–922.

Botvinick, M. M., Braver, T. S., Barch, D. M., Carter,C. S., & Cohen, J. D. (2001). Conflict monitoringand cognitive control. Psychological Review, 108,624–652.

Botvinick, M. M., Nystrom, L. E., Fissell, K., Carter, C.S., & Cohen, J. D. (1999). Conflict monitoring versusselection-for-action in anterior cingulate cortex.Nature, 402, 179–181.

Braver, T. S. (2012). The variable nature of cognitivecontrol: A dual mechanisms framework. Trends in

Cognitive Sciences, 16, 106–113.Braverman, A., &Meiran, N. (2010). Task conflict effect

in task switching. Psychological Research, 74, 568–578.Bugg, J. M., McDaniel, M. A., Scullin, M. K., & Braver,

T. S. (2011). Revealing list-level control in the Strooptask by uncovering its benefits and a cost. Journal ofExperimental Psychology: Human Perception and

Performance, 37, 1595–1606.Carter, C. S., Botvinick, M. M., & Cohen, J. D. (1999).

The contribution of the anterior cingulated cortex toexecutive processes in cognition. Reviews in

Neuroscience, 10, 49–57.Carter, C. S., Braver, T. S., Barch, D. M., Botvinick, M.

M., Noll, D., & Cohen, J. D. (1998). Anterior cingu-late cortex, error detection, and the online monitoringof performance. Science, 280, 747–749.

Carter, C. S., Mintun, M., & Cohen, J. D. (1995).Interference and facilitation effects during selectiveattention: An H215O PET study of Stroop task per-formance. NeuroImage, 2, 264–272.

Cohen, J. D., Dunbar, K., & McClelland, J. L. (1990).On the control of automatic processes: A parallel dis-tributed processing account of the Stroop effect.Psychological Review, 97, 332–361.

Cohen, J. D., & Servan-Schreiber, D. (1992). Context,cortex and dopamine: A connectionist approach tobehavior and biology in schizophrenia. PsychologicalReview, 99, 45–77.

Cousineau, D. (2005). Confidence intervals in within-subjects designs: A simpler solution to Loftus andMasson’s method. Tutorials in Quantitative Methods

for Psychology, 1, 42–45.


KALANTHROFF ET AL.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

8:31

03

July

201

3

21

Dalrymple-Alford, E. C., & Budayr, B. (1966).Examination of some aspects of the Stroop color-word test. Perceptual and Motor Skills, 23, 1211–1214.

David, I. A., Volchan, E., Vila, J., Keil, A., deOliveira, L.,Faria-Júnior, A. J. P., et al. (2011). Stroop matchingtask: Role of feature selection and temporal modu-lation. Experimental Brain Research, 208, 595–605.

De Pisapia, N., & Braver, T. S. (2006). A model of dualcontrol mechanisms through anterior cingulate andprefrontal cortex interactions. Neurocomputing, 69,1322–1326.

Goldfarb, L., & Henik, A. (2007). Evidence for taskconflict in the Stroop effect. Journal of ExperimentalPsychology: Human Perception and Performance, 33,1170–1176.

Kalanthroff, E., Goldfarb, L., & Henik, A. (in press).Evidence for interaction between the stop-signaland the Stroop task conflict. Journal of ExperimentalPsychology: Human Perception and Performance. doi:10.1037/a0027429.

Kalanthroff, E., & Henik, A. (2012). Individual but notfragile: Individual differences in task control predict

Stroop facilitation. Manuscript submitted forpublication.

La Heij, W., & Boelens, H. (2011). Color–object inter-ference: Further tests of an executive control account.Journal of Experimental Child Psychology, 108,156–169.

La Heij, W., Boelens, H., & Kuipers, J. R. (2010).Object interference in children’s colour and positionnaming: Lexical interference or task-set competition?Language and Cognitive Processes, 25, 568–588.

Logan, G. D. (1994). On the ability to inhibit thoughtand action: A user’s guide to the stop signal paradigm.In D. Dagenbach & T. H. Carr (Eds.), Inhibitory pro-cesses in attention, memory and language (pp. 189–239).San Diego, CA: Academic Press.

Logan, G. D., & Cowan, W. B. (1984). On the ability toinhibit thought and action: A theory of an act ofcontrol. Psychological Review, 91, 295–327.

MacLeod, C. M. (1991). Half a century of research onthe Stroop effect: An integrative review.Psychological Bulletin, 109, 163–203.

MacLeod, C. M., & MacDonald, P. A. (2000).Interdimensional interference in the Stroop effect:Uncovering the cognitive and neural anatomy ofattention. Trends in Cognitive Sciences, 10, 383–391.

Melara, R. D., & Algom, D. (2003). Driven by infor-mation: A tectonic theory of Stroop effects.Psychological Review, 110, 422–471.

Milham, M. P., Erickson, K. I., Banich, M. T., Kramer,A. F., Webb, A., Wszalek, T., et al. (2002).Attentional control in the aging brain: Insights froman fMRI study of the Stroop task. Brain &

Cognition, 49, 277–296.Miller, E. K., & Cohen, J. D. (2001). An integrative

theory of prefrontal cortex function. Annual Reviewof Neuroscience, 24, 167–202.

Miyake, A., Friedman, N. P., Emerson, M. J., Witzki,A. H., Howerter, A., & Wager, T. D. (2000). Theunity and diversity of executive functions and theircontributions to complex “frontal lobe” tasks: Alatent variable analysis. Cognitive Psychology, 41,49–100.

Monsell, S., Taylor, T. J., & Murphy, K. (2001).Naming the color of a word: Is it responses or tasksets that compete? Memory and Cognition, 29,137–151.

Rogers, R. D., & Monsell, S. (1995). Costs of a predict-able switch between simple cognitive tasks. Journal ofExperimental Psychology: General, 124, 207–231.

Sabri, M., Melara, R. D., & Algom, D. (2001). A con-fluence of contexts: Asymmetric versus global failuresof selective attention to Stroop dimensions. Journal ofExperimental Psychology: Human Perception and

Performance, 27, 515–537.Shallice, T., & Norman, D. (1986). Attention to action:

Willed and automatic control of behavior. InR. Davidson, G. Schwartz, & D. Shapiro (Eds.),Consciousness and self-regulation: Advances in research

and theory (Vol. 4, pp. 1–18). New York, NY:Plenum Press.

Steinhauser, M., & Hubner, R. (2009). Distinguishingresponse conflict and task conflict in the Strooptask: Evidence from ex-Gaussian distribution analy-sis. Journal of Experimental Psychology: Human

Perception and Performance, 35, 1398–1412.Stroop, J. R. (1935). Studies of interference in serial

verbal reactions. Journal of Experimental Psychology,18, 643–662.

Tzelgov, J., Henik, A., & Berger, J. (1992). ControllingStroop effects by manipulating expectations for colorwords. Memory & Cognition, 20, 727–735.

Verbruggen, F., & Logan, G. (2008). Response inhi-bition in the stop-signal paradigm. Trends in

Cognitive Sciences, 12, 418–424.Waszak, F., Hommel, B., & Allport, D. A. (2003). Task

switching and long-term priming: Role of episodicstimulus–task bindings in task-shift costs. CognitivePsychology, 46, 361–413.



Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 0

8:31

03

July

201

3

22

23

Chapter II

Experiment 2

Individual but not Fragile: Individual Differences in Task

Control Predict Stroop Facilitation

Eyal Kalanthroff & Avishai Henik

Consciousness and Cognition

2013

Author's personal copy

Short Communication

Individual but not fragile: Individual differences in taskcontrol predict Stroop facilitation

E. Kalanthroff *, A. HenikDepartment of Psychology and Zlotowski Center for Neuroscience, Ben-Gurion University of the Negev, Beer Sheva, Israel

a r t i c l e i n f o

Article history:Received 19 September 2012Available online 15 February 2013

Keywords:StroopTask conflictAnterior cingulate cortexExecutive functionStop signal

a b s t r a c t

The Stroop effect is composed of interference and facilitation effects. The facilitation is lessstable and thus many times is referred to as a ‘‘fragile effect”. Here we suggest the facilita-tion effect is highly vulnerable to individual differences in control over the task conflict(between relevant color naming and irrelevant word reading in the Stroop task). We rep-licated previous findings of a significant correlation between stop-signal reaction time(SSRT) and Stroop interference, and also found a significant correlation between SSRTand the Stroop facilitation effect—participants with low inhibitory control (i.e., long SSRT)had no facilitation effect or even a reversed one. These results shed new light on the ‘‘frag-ile” facilitation effect and highlight the necessity of awareness of task conflict, especially inthe Stroop task.

� 2013 Elsevier Inc. All rights reserved.

1. Introduction

Executive control is a key human function that mediates the ability to maintain goal-directed behavior (Banich, 2009;Miller & Cohen, 2001; Miyake et al., 2000; Shallice & Norman, 1986). A popular task to investigate this process in the labis the Stroop task (Stroop, 1935). In the common task, individuals are asked to identify the ink color in which a word isprinted (i.e., word color) while ignoring the word meaning. The word color and word meaning can be either congruent (e.g., RED written in red), incongruent (e.g., GREEN written in red) or neutral (e.g., XXXX written in red). It is commonly as-sumed that incongruent trials create a conflict, and a control mechanism is recruited to settle this conflict. Hence, the reac-tion time (RT) for incongruent trials is commonly longer than RT for neutral trials. This is known as the interference effect,which is a very strong and stable effect (MacLeod, 1991). In congruent trials, both word color and word meaning activate thesame response; hence, RT is usually faster than in neutral trials. This is known as the facilitation effect, which is a much smal-ler and less stable effect than the interference effect (e.g., Dalrymple-Alford & Budayr, 1966). The facilitation effect is some-times absent (Kalanthroff, Goldfarb, & Henik, 2012; Mathis, Schunck, Erb, Namer, & Luthringer, 2009), and in many studiesreferred to as a ‘‘fragile effect” (e.g., Fuentes & Ortells, 1993; Goldfarb & Henik, 2007; Logan & Zbrodoff, 1998; MacLeod &MacDonald, 2000). The current study aims to address the fragility of the facilitation effect and examines whether individualdifferences in the ability to execute control produce differences in the size of the facilitation effect.

Several neuroimaging studies demonstrated that the anterior cingulated cortex (ACC)—a brain area thought to monitorconflict (Botvinick, Braver, Barch, Carter, & Cohen, 2001; Botvinick, Nystrom, Fissell, Carter, & Cohen, 1999; Carter, Botvinick,& Cohen, 1999; Carter et al., 1998)—is more activated in the incongruent condition than in the neutral condition (e.g., Benchet al., 1993; Carter, Mintun, & Cohen, 1995; Milham et al., 2002, for older participants). Interestingly, however, these studies

1053-8100/$ - see front matter � 2013 Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/j.concog.2013.01.010

* Corresponding author. Address: Department of Psychology, Ben-Gurion University of the Negev, P.O.B. 653, Beer Sheva 84105, Israel. Fax: +972 86472072.

E-mail address: [email protected] (E. Kalanthroff).

Consciousness and Cognition 22 (2013) 413–419

Contents lists available at SciVerse ScienceDirect

Consciousness and Cognition

journal homepage: www.elsevier .com/locate /concog

24


also showed that not only the incongruent Stroop stimuli triggered higher ACC activations than the neutrals did, but the con-gruent Stroop trials did so too. Hence, the findings of ACC activation mentioned above imply that conflict occurs not only inthe incongruent condition but also in the congruent condition.

The contradiction between the behavioral (RT to congruent trials faster than to neutral trials) and neurological (higherACC activation to congruent than neutral trials) findings has been subject to a number of recent investigations. Goldfarband Henik (2007) suggested that the contradiction can be settled by distinguishing between two conflicts reflected in theStroop task—informational (between the incongruent word and ink color) and task (between relevant color naming and irrel-evant word reading). While the information conflict involves the content of the stimulus and the response needed, differingbetween congruent and incongruent Stroop stimuli, task conflict involves the task associated with the stimulus, and differsbetween congruent and non-word neutral Stroop stimuli (Kalanthroff, Goldfarb, & Henik, 2012). This is consistent with thesuggestion that a stimulus has the ability to evoke performance of a task that has a strong association with it (Allport & Wy-lie, 2000; Rogers & Monsell, 1995; Waszak, Hommel, & Allport, 2003) and specifically, that words trigger an automatic ten-dency to read written words (MacLeod &MacDonald, 2000). Monsell (2003) proposed that a cognitive task, and the efficiencywith which we perform it, results from an interplay between 2 mechanisms: (1) endogenous control, which emphasizes therole of deliberate intentions governed by goals, and (2) exogenous influences, which are caused by alternative possible tasksavailable by the stimulus. In the Stroop task, endogenous control will lead to focusing attention on the color naming task,whereas word reading will be the strongest exogenous influence. Monsell’s proposal seems to fit with the task conflict rea-soning. Recently, David et al. (2011) found evidence, in an event-related potential (ERP) study, for the existence of task con-flict in the Stroop matching task; Steinhauser and Hubner (2009) found evidence for the dissociation of task conflict from theresponse conflict in the ex-Gaussian distribution analysis.

Wilkinson and Halligan (2004) suggested that one of the cases in which behavior expected from brain activation will notappear is when another brain structure contributes to the reduction of this behavior. Accordingly, the contradiction betweenbehavior and brain activation in the Stroop task indicates the existence of an active control mechanism aimed at eliminatingthe task conflict. In line with this, Goldfarb and Henik (2007) claimed that the task conflict is usually not seen due to a veryefficient task control mechanism. Thus, task conflict does delay responding to the congruent condition but (due to the effi-cient control mechanism) not enough to make it slower than responding to the neutral condition. Task control is a cognitivemechanism that is responsible for solving the task conflict or the ‘task set competition’; namely, the need to perform a targettask and play down an irrelevant task (Banich, 2009; Goldfarb & Henik, 2007; Monsell, 2003; Monsell, Taylor, & Murphy,2001). Recently, Braver and colleagues (Braver, 2012; De Pisapia & Braver, 2006) suggested the dual mechanism theory ofpro-active (reflecting the strategic block-wise control) and reactive (reflecting the stimulus driven trial-by-trial control)mechanisms. Though this framework does not account for task conflict, it seems reasonable to assume that task conflict willusually become visible when pro-active control (or the ‘task demand’ component introduced in earlier models, e.g., Botvinicket al., 2001) is low. Goldfarb and Henik (2007) tried to ‘put the task conflict (pro-active) guard to sleep’ by increasing the pro-portion of nonword neutral trials and by providing cues for whether the upcoming trial was going to be a neutral or a conflictone (i.e., congruent or incongruent). This reduced the task conflict control and resulted in a reversed facilitation effect inwhich the congruent RT was longer than the neutral RT—a behavioral evidence for the task conflict. In a more recent study(Kalanthroff, Goldfarb, Usher, & Henik, 2012), we found that the reverse facilitation effect appears in a similar procedureeven if no incongruent trials exist. This suggests that task conflict appears whenever there is an alternative possible taskregardless of the response it initiates. Thus, it seems reasonable to assume that RTs to Stroop congruent trials are contingentupon the efficiency of the task control mechanism.

Task conflict and the idea that the direction of the facilitation effect depends on the activation of an irrelevant task hasbeen discussed in the task-switching literature (e.g., Aarts, Roelofs, & van Turennouta, 2009; Braverman & Meiran, 2010;Rogers & Monsell, 1995). Considering that task switching reduces pro-active control, it is not surprising that a reverse facil-itation is often seen in task-switching studies (see also Aron, Monsell, Sahakian, & Robbins, 2004).

The current study aims to investigate whether individual differences in task control can explain the variance in the Stroopfacilitation effect. As mentioned earlier, the Stroop facilitation effect is small, unstable and often referred to as ‘‘fragile”—some individuals and some subject group samples do not show the effect (e.g., Fuentes & Ortells, 1993; Kalanthroff, Goldfarb,& Henik, 2012; Logan & Zbrodoff, 1998; MacLeod & MacDonald, 2000). We predict that individual differences in the effi-ciency of the task control mechanism can explain why some individuals have a facilitation effect and some do not.

In the Stroop task subjects have to inhibit their irrelevant but automatic, prepotent reading responses (Miyake et al.,2000). Such inhibition is needed to improve performance in both incongruent and congruent trials. La Heij and Boelens(2011) recently suggested that children’s low ability to manage task conflict can result from their immature inhibition ofa prepotent response mechanism. Accordingly, we believe that the ability to inhibit a prepotent response can predict themagnitude of both the Stroop interference and facilitation effects.

A task that is commonly used to measure inhibition of a prepotent response is the stop-signal task (Logan, 1994; Logan &Cowan, 1984), which examines the ability to suppress an already initiated action that is no longer appropriate. A great deal ofresearch has been carried out in the past few decades with the stop-signal task (for review see Verbruggen & Logan, 2008). Inthe classic task, participants are asked to address a visual stimulus (go signal) with a motor response as fast as possible. Inabout 25% of the trials, an auditory stimulus (stop signal) comes right after the visual go signal. Participants are instructed toinhibit their motor response when they hear the stop signal. The duration between the go signal and the stop signal (SSD;stop-signal delay) is subjected to a tracking procedure—in the subsequent stopping trial, the SSD will become a bit longer

414 E. Kalanthroff, A. Henik / Consciousness and Cognition 22 (2013) 413–419

25


(more difficult) following successful inhibition, or a bit shorter (easier) following erroneous response to the stop signal.Eventually, it becomes possible to estimate the stop-signal reaction time (SSRT), which is the time needed for successful inhi-bition. Logan and Cowan (1984), and Logan, Cowan, and Davis (1984) compared the performance in the stop-signal task to a‘horse race’ between the go process (triggered by the go signal) and the stop process (triggered by the stop signal). While RTis the time needed for the go process to finish, SSRT is time needed for the stop process to finish. Logan et al. (1984) arguedthat ‘‘response inhibition phenomena are consistent with a hierarchical theory of attention in which a high level processdetermines the significance of incoming stimuli and decides whether to abort the current stream of thought and action. . .”(p. 290). Indeed, the stop process represents high-level control and SSRT has proven to be an important measure of cognitivecontrol (Verbruggen & Logan, 2008), and has proven useful in developmental and pathology studies (for reviews see Boucher,Palmeri, Logan, & Schall, 2007; Logan, 1994).

In a previous study (Kalanthroff, Goldfarb, & Henik, 2012), we analyzed stop-signal control failure trials (i.e., erroneousresponse to stop-signal trials). In that study, we found longer RTs in the Stroop congruent condition compared with theStroop neutral condition. Thus, a momentary failure in stopping results in both an erroneous response to a stop signaland in slow management of the task conflict. These results strengthened the notion that there could be some overlap be-tween the inhibition of a prepotent response and task control or the control needed to manage the task conflict in the Stroopcongruent condition. This fits in with Friedman and Miyake’s (2004) suggestion that stopping and Stroop belong to the samelatent variable.1

Here we suggest that variance in the Stroop facilitation effect is not random and that it can be explained in terms of indi-vidual differences in task control. The current study aims to disclose a factor that contributes to the variance in the Stroopfacilitation effect—the task conflict that relies on inhibitory control. According to the notion that task control relies on inhi-bition of a pre-potent response, we predicted that individual differences in SSRT would correlate with individual differencesin Stroop facilitation. More specifically, we expected that a long SSRT (which indicates low inhibitory control) would predictno (or even reversed) Stroop facilitation, and that a short SSRT (high inhibitory control) would predict large facilitation.

2. Method

2.1. Participants

One hundred and thirty-six undergraduate students (87 females and 49 males) of Ben-Gurion University of the Negev(Israel) participated for partial fulfillment of course requirements and received course credit for participating. All partici-pants had normal or corrected-to-normal vision, had no history of attention deficit or dyslexia, were native speakers of He-brew, and all were naive as to the purpose of the experiment. Participants who had more than a 15% error rate or an averageRT of more than 1200 ms in the Stroop task were excluded from the analysis. In addition, since SSRT is an estimation of thetime needed for a participant to stop on 50% of the trials, if there were significantly more or less than 50% successful inhi-bitions in the stop-signal task, the SSRT would not be valid and hence, these participants were excluded from the analysis(estimation method by Verbruggen & Logan, 2009; see also Verbruggen, Logan, & Stevens, 2008). Four participants (twomales and two females) were excluded due to these criteria (two due to high error rate, one due to invalid SSRT, and onedue to both). From the remaining 132 participants, the youngest was 20 years old and the eldest was 30 years old(mean = 24.1 years, SD = 1.6).

2.2. Stimuli

Participants were presented with stimuli that were appropriate for either the Stroop or the stop-signal tasks.

2.2.1. Stroop stimuliEach stimulus consisted of one of four Hebrew color words— לוחכ (blue), םודא (red), קורי (green), and בוהצ (yellow)—or

a four-letter string in Hebrew— שששש (meaningless repetition of a Hebrew letter, parallel to XXXX in the English version).The meaningless letter string would not trigger automatic reading, and as such, would cause no task conflict. The ink colorsof the stimuli were red, blue, green or yellow. There were 20 combinations of words and ink colors: 4 congruent, 4 neutraland 12 incongruent. The words were presented at the center of a screen on a black background and were 0.98 inches highand 2.36 inches wide.

2.2.2. Stop-signal stimuliIn the stop-signal task, we used the ‘‘Stop-it” program (Verbruggen et al., 2008), where the go signal was a white square or

circle on a black background. The stop signal was an auditory tone (750 Hz, 75 ms).

1 It is not possible to examine the correlation between the stop signal and task conflict in the Friedman and Miyake (2004) study. This is because: (a) theStroop effect was calculated using word-neutral stimuli, which potentially also cause task conflict (thus, task conflict exists in both incongruent and neutraltrials and the difference between them does not reflect task conflict); and (b) no congruent trials were used, thus it is impossible to examine whether thevariance in the Stroop facilitation effect is contingent upon individual differences in inhibition of a prepotent response.

E. Kalanthroff, A. Henik / Consciousness and Cognition 22 (2013) 413–419 415

26


2.3. Procedure

Data collection and stimuli presentation were controlled by a DELL OptiPlex 760 vPro computer with an Intel core 2 duoprocessor E8400 3 GHz. Stimuli were presented on a DELL E198PF 1900 LCD monitor. A keyboard was placed on a table be-tween the participant and the monitor. Participants were tested individually. They sat approximately 23.62 in. from the com-puter screen. In the stop-signal task participants were asked to wear headphones. Participants were asked to complete amanual Stroop task followed by a stop-signal task.

2.3.1. Stroop taskThe experiment started with 12 key-matching practice trials. In these trials a colored asterisk appeared at the center of

the screen, three times in each of the four colors, in a random order. Participants were instructed to hit the ‘‘m” key on thekeyboard for blue, the ‘‘n” key for yellow, the ‘‘b” key for red, and the ‘‘v” key for green (keys were marked with colored stick-ers). The asterisk disappeared on a key-press or after 3500 ms, and then the next trial began after a 1000 ms interval. Feed-back was given only for incorrect trials.

Following this, the Stroop task started. The task included 40 practice trials (which were not analyzed further) and 288experimental trials. Proportions for each congruency condition were equal. During practice, participants received feedbackfor accuracy and RT. Each trial started with the presentation of a fixation (a white plus sign in the center of a black screen) for500 ms, followed by a visual stimulus. Participants were instructed to hit the appropriate color-stickered key. They wereasked to respond with their dominant hand, which was placed on the keyboard with four fingers (all but the thumb) onthe four different response keys. They were asked to ignore the meaning of the word and to press the correct key as quicklyand accurately as possible. The visual stimulus stayed in view for 2000 ms or until a key-press. RT was calculated from theappearance of the visual stimulus to the reaction. Each trial ended with a 1000 ms inter-trial interval.

2.3.2. Stop-signalThe experiment included one practice block of 32 trials and three experimental blocks of 64 trials each. Each trial started

with a 250 ms fixation (a white plus sign in the center of a black screen), followed by a visual go stimulus. Response keyswere ‘‘z” for the appearance of a square and "/” for a circle. Participants were asked to respond with the index fingers of bothhands. The instruction stated to press the correct key as fast and accurately as possible, and emphasized not to wait for apotential stop signal. The visual stimulus stayed in view for 1250 ms regardless of the latency of the response. RT was cal-culated from the appearance of the visual stimulus to the response. Each trial ended with a 2000 ms inter-trial interval. On arandom selection of the trials, an auditory stop signal was sounded. The stop signal was presented after a variable SSD thatwas initially set at 250 ms and adjusted by a staircase tracking procedure: after each successful stopping the SSD was ex-tended by 50 ms and after each unsuccessful stopping the SSD was shortened by 50 ms. Between blocks, participants hadto wait for 10 s before they could start the next block. During this interval, they received information about their perfor-mance in the previous block. For further details on the stimuli and procedure see Verbruggen et al. (2008).

3. Results

First we will analyze each task separately and then we will address the correlations between tasks. For the stop-signaltask we used the ‘‘Analyze-it” program (Verbruggen et al., 2008) for all results. All analyses in this study are based on datafrom all trials.

3.1. Stroop analysis

Mean RTs for correct responses were calculated for each participant for each congruency condition. A one-way analysis ofvariance (ANOVA) with repeated measures was applied to RT data with congruency (congruent, neutral, and incongruent) asa within subject factor.

A significant main effect was found, F(2,262) = 27.368, MSE = 4271.613, p < .001, partial eta square (PES) = .173. RTs (anderror rates) for the congruency conditions were 617 ms (5.3%), 629 ms (6.3%), and 673 ms (6.5%) for congruent, neutral andincongruent trials, respectively. These results indicate a strong congruency effect (i.e., incongruent minus congruent trial RT),F(1,131) = 33.719, MSE = 6219.692, p < .001, PES = .205; a large interference effect (i.e., incongruent minus neutral trial RT), F(1,131) = 120.239, MSE = 1102.504, p < .001, PES = .479; and a non-significant small facilitation effect (neutral minus congru-ent trial RT), F(1,131) = 1.604, MSE = 5492.643, p = .208. As mentioned earlier, these results correspond with the commonStroop effect (Macleod, 1991).

3.2. Stop-signal analysis

The mean RT for no-signal trials was 496 ms (SD = 119 ms). SSRT was calculated by subtracting the mean SSD from theno-signal mean RT using the Analyze-it program (Verbruggen & Logan, 2009). Mean SSRT was 222 ms (SD = 70 ms).


27


3.3. Correlations

As expected, a significant correlation was found between the SSRT and the Stroop interference effect, r = 0.209, p < .02.This corresponded with Friedman and Miyake’s (2004) findings. The correlation between the Stroop facilitation effect andSSRT was also significant, r = �0.226, p < .01 (both correlations were parametric). In order to make sure that there wereno artifacts of general RTs (even though SSRT is not supposed to be affected by general RTs), we conducted a block regressionmodel to predict RTs of incongruent or congruent conditions using SSRT after entering the neutral condition (a measurementfor general RTs) as a predictor. The neutral condition contributed to the prediction of both incongruent (standardizedb = .933, p < .001) and congruent conditions (standardized b = .549, p < .001). SSRT also contributed significantly to the pre-diction of incongruent RTs (standardized b = .083, p < .01), and congruent RTs (standardized b = .191, p < .01). Entering SSRTin addition to neutrals significantly increased the R2 for incongruent ðR2

change ¼ :007; Fchangeð1;129Þ ¼ 6:664;p < :01Þ and for congruent conditions ðR2

change ¼ :036; Fchangeð1;129Þ ¼ 6:934; p < :01Þ.In order to examine our a priori assumption, we divided all subjects into 6 equal groups (N = 22 in each group) according

to SSRT (from short to long). A one-way ANOVA was applied to facilitation with SSRT-group as a between subject factor. Asignificant main effect was found, F(5,126) = 2.552, MSE = 10,371.002, p = .03, PES = .092. A trend analysis revealed a signif-icant linear component, t(1) = �3.291, p < .001. The quadratic, cubic, and quartic components were not significant. As can beseen in Fig. 1, these results indicate that the magnitude of the facilitation decreases with the increase in SSRT and reverses inthe long SSRT group.

4. Discussion

In recent years many studies have demonstrated task conflict, or ‘task set competition’ in various tasks (e.g., Aarts et al.,2009; Banich, 2009; Braverman & Meiran, 2010; David et al., 2011; Monsell, 2003; Steinhauser & Hubner, 2009). Goldfarband Henik (2007) demonstrated the existence of this conflict in the Stroop congruent condition and indicated that the taskconflict is usually not behaviorally seen due to a very efficient task control mechanism. Similarly, La Heij, Boelens, and Kui-pers (2010) argued that task conflict in the object interference task only shows in children due to their yet undeveloped con-trol. The goal of the current study was to investigate the influence of a specific control mechanism on the Stroop facilitationeffect. Accordingly, we argue that the commonly found fragile Stroop facilitation (e.g., Fuentes & Ortells, 1993; Logan &Zbrodoff, 1998) can be better understood in terms of individual differences in task control.

In order to identify individual differences in inhibition of a prepotent response, we used the stop-signal task (Logan, 1994;Logan & Cowan, 1984). We found a significant correlation between SSRT and the Stroop interference effect, which replicatesFriedman and Miyake’s (2004) findings. Most importantly, there was a significant correlation between SSRT and the Stroopfacilitation effect that was not an artifact of general RT. More specifically, we found that fast SSRT (indicating very efficientinhibitory control) predicts large facilitation while slow SSRT (less efficient inhibitory control) predicts no facilitation or evena reverse facilitation. Though this effect had a significant linear component, it was more robust for the extreme participantresults, and the difference for those in the middle (e.g., SSRT-groups 3–4) was not significant. It seems that for a big part ofthe participant pool, who fall in the middle of the distribution, there is no significant difference in task control ability, andthese insignificant differences hardly affect the facilitation effect. For the 22 participants with the longest SSRTs in our studywe found a reverse facilitation effect. It seems that individuals who have low inhibitory control need longer time to solve atask conflict than individuals characterized by high inhibitory control.

The current findings strengthen the assumption that task control relies on inhibition of a prepotent response and fit inwith previous findings of Kalanthroff, Goldfarb, and Henik (2012) that demonstrated this connection. Specifically, in thatstudy we found that on no-stop trials there was a regular Stroop effect composed of interference and facilitation, whereason erroneous responses to stop trials there was interference and a reverse facilitation. We suggested that task conflictwas more visible (i.e., reverse facilitation) when inhibitory control was low. Furthermore, Friedman and Miyake (2004)

-80

-60

-40

-20

0

20

40

60

80

169 207 225 238 260 302

Fac

ilita

tion

Eff

ect

(RT

in m

s)

SSRT Group

Fig. 1. Facilitation effect in the different SSRT groups (mean SSRT in milliseconds for each group appears on the X axis). Error bars represent one standarderror from the mean.


28


showed that stopping and Stroop performance belong to the same individual differences factor. Using structural equationmodeling, these researchers reported that prepotent response inhibition and resistance to distractor interference (as mea-sured in the Stroop incongruent condition) were closely related. Our results fit in with these findings and reveal some over-lap between the Stroop interference effect and stop-signal inhibition. Importantly, dividing the Stroop effect intointerference and facilitation effects revealed that a correlation between stop-signal inhibition and facilitation also exists.

In previous studies (Kalanthroff, Goldfarb, & Henik, 2012; Kalanthroff et al., 2012) we found that task conflict in theStroop task occurs whenever a word is presented, regardless of the existence of informational conflict. Tzelgov, Henik,and Berger (1992) found larger facilitation and smaller interference in a Stroop task with a word neutral (e.g., lion in red)compared to a Stroop task with a non-word neutral (e.g., xxxx in red). In other words, response to the word neutral conditionwas a bit slower compared to a non-word neutral. Though they used different participants for each task, these researchers’findings imply that the appearance of any word causes a task conflict. Bugg, McDaniel, Scullin, and Braver (2011) found adecreased interference effect in a block with a high proportion of word-neutral. This strengthens the notion that neutralwords cause a conflict and require executive control. The idea that word neutrals might themselves be a source of interfer-ence and slow the RT is not new. About 30 years ago Kahneman and Henik (1981), and Kahneman and Chajczyk (1983)showed that attention may be required, even for autonomous processing, like word reading in the Stroop task. The idea thatany word will cause a conflict corresponds with our findings indicating task conflict in the Stroop congruent condition.

Monsell et al. (2001) suggested that at least part of the interference in Stroop tasks was due to competition between therelevant task and the irrelevant automatic task. The irrelevant automatic task concept can be better understood according toMacLeod and MacDonald (2000), who argued that words are associated with an automatic tendency to read. Hence, in anyStroop trial there is a conflict between two tasks—the relevant color naming task and the irrelevant word reading task, whichis triggered by the word stimulus. We argue that this task conflict affects RTs in the congruent condition and thus the effi-ciency of dealing with this conflict (i.e., task control) modulates the facilitation effect. Therefore, individual differences intask control efficiency account for individual differences in the facilitation effect. In light of this, we can broaden Monsellet al.’s suggestion by adding that the part of the Stroop interference effect due to task competition is highly vulnerable toindividual differences in task control. Importantly, the Stroop informational conflict also affects RTs in both the incongruentcondition, where it slows responding, and the congruent condition, where it speeds up (facilitates) responding.

Our results indicate that inhibitory control has a unique contribution to the prediction of task conflict managing time, inaddition to the contribution of general cognitive processing and motor speed. However, we cannot rule out the possibilitythat executive function in general (and not specifically inhibitory control) can predict congruent RTs. This is an importantlimitation that should be considered. Although, we do not believe this to be the case, the idea that executive control effi-ciency can predict the Stroop facilitation effect is novel.

It is also important to mention that although task control was shown to contribute to the prediction of the magnitude ofStroop facilitation, our results indicated that there is still much variance of the Stroop facilitation effect that is not explainedby task conflict control. Further research is needed in order to reveal those factors.

5. Conclusions

Our findings suggest that the Stroop facilitation effect is highly affected by the efficiency of task control, which relies oninhibitory control – a novel suggestion considering that facilitation was not attributed to executive functions before. Thus,individual differences in inhibitory control can account for individual differences in the Stroop facilitation effect. These re-sults support the suggestion that fast reaction to Stroop congruent trials requires fast resolution of the task conflict.

Acknowledgment

We thank Desiree Meloul for helpful comments and useful input on this article.

References

Aarts, E., Roelofs, A., & van Turennouta, M. (2009). Attentional control of task and response in lateral and medial frontal cortex: Brain activity and reactiontime distributions. Neuropsychologia, 47, 2089–2099.

Allport, A., & Wylie, G. (2000). Task-switching, stimulus-response bindings and negative priming. In S. Monsell & J. Driver (Eds.), Control of cognitiveprocesses: Attention and performance XVIII (pp. 35–70). Cambridge, MA: MIT Press.

Aron, A. R., Monsell, S., Sahakian, B. J., & Robbins, T. W. (2004). A componential analysis of task-switching deficits associated with lesions of left and rightfrontal cortex. Brain, 127, 1561–1573.

Banich, M. T. (2009). Executive function: The search for an integrated account. Current Directions in Psychological Science, 18, 89–94.Bench, C. J., Frith, C. D., Grasby, P. M., Friston, K. J., Paulesu, E., Frackowiak, R. S., et al (1993). Investigations of the functional anatomy of attention using the

Stroop test. Neuropsychologia, 32, 907–922.Botvinick, M. M., Braver, T. S., Barch, D. M., Carter, C. S., & Cohen, J. D. (2001). Conflict monitoring and cognitive control. Psychological Review, 108, 624–652.Botvinick, M. M., Nystrom, L. E., Fissell, K., Carter, C. S., & Cohen, J. D. (1999). Conflict monitoring versus selection-for-action in anterior cingulate cortex.

Nature, 402, 179–181.Boucher, L., Palmeri, T. J., Logan, G. D., & Schall, J. D. (2007). Inhibitory control in mind and brain: An interactive race model of countermanding saccades.

Psychological Review, 114, 376–397.Braver, T. S. (2012). The variable nature of cognitive control: A dual mechanisms framework. Trends in Cognitive Sciences, 16, 106–113.Braverman, A., & Meiran, N. (2010). Task conflict effect in task switching. Psychological Research Psychologische Forschung, 74, 568–578.


29


Bugg, J. M., McDaniel, M. A., Scullin, M. K., & Braver, T. S. (2011). Revealing list-level control in the Stroop task by uncovering its benefits and a cost. Journal ofExperimental Psychology: Human Perception and Performance, 37, 1595–1606.

Carter, C. S., Botvinick, M. M., & Cohen, J. D. (1999). The contribution of the anterior cingulated cortex to executive processes in cognition. Reviews inNeuroscience, 10, 49–57.

Carter, C. S., Braver, T. S., Barch, D. M., Botvinick, M. M., Noll, D., & Cohen, J. D. (1998). Anterior cingulate cortex, error detection, and the online monitoring ofperformance. Science, 280, 747–749.

Carter, C. S., Mintun, M., & Cohen, J. D. (1995). Interference and facilitation effects during selective attention: An H215O PET study of Stroop taskperformance. NeuroImage, 2, 264–272.

Dalrymple-Alford, E. C., & Budayr, B. (1966). Examination of some aspects of the Stroop color-word test. Perceptual and Motor Skills, 23, 1211–1214.David, I. A., Volchan, E., Vila, J., Keil, A., de Oliveira, L., Faria-Júnior, A. J. P., et al (2011). Stroop matching task: Role of feature selection and temporal

modulation. Experimental Brain Research, 208, 595–605.De Pisapia, N., & Braver, T. S. (2006). A model of dual control mechanisms through anterior cingulate and prefrontal cortex interactions. Neurocomputing, 69,

1322–1326.Friedman, N. P., & Miyake, A. (2004). The relations among inhibition and interference control functions: A latent-variable analysis. Journal of Experimental

Psychology: General, 13(1), 101–135.Fuentes, L. J., & Ortells, J. J. (1993). Facilitation and interference effects in a Stroop-like task: Evidence in favor of semantic processing of parafoveally-

presented stimuli. Acta Psychologica, 84, 213–229.Goldfarb, L., & Henik, A. (2007). Evidence for task conflict in the Stroop effect. Journal of Experimental Psychology: Human Perception and Performance, 33,

1170–1176.Kahneman, D., & Chajczyk, D. (1983). Tests of automaticity of reading: Dilution of Stroop by color-irrelevant stimuli. Journal of Experimental Psychology:

Human Perception and Performance, 9, 497–509.Kahneman, D., & Henik, A. (1981). Perceptual organization and attention. In M. Kubovy & J. R. Pomerantz (Eds.), Perceptual organization (pp. 181–211).

Hillsdale, NJ: Erlbaum.Kalanthroff, E., Goldfarb, L., & Henik, A. (2012). Evidence for interaction between the stop-signal and the Stroop task conflict. Journal of Experimental

Psychology: Human Perception and Performance. E-pub ahead of print.Kalanthroff, E., Goldfarb, L., Usher, M., & Henik, A. (2012). Stop interfering: Stroop task conflict independence from informational conflict and interference.

Quarterly Journal of Experimental Psychology, 1, 1–2. E-pub ahead of print.La Heij, W., & Boelens, H. (2011). Color–object interference: Further tests of an executive control account. Journal of Experimental Child Psychology, 108,

156–169.La Heij, W., Boelens, H., & Kuipers, J. R. (2010). Object interference in children’s colour and position naming: Lexical interference or task-set competition?

Language and Cognitive Processes, 25, 568–588.Logan, G. D. (1994). On the ability to inhibit thought and action: A user’s guide to the stop signal paradigm. In D. Dagenbach & T. H. Carr (Eds.), Inhibitory

processes in attention, memory and language (pp. 189–239). San Diego, CA: Academic Press.Logan, G. D., & Cowan, W. B. (1984). On the ability to inhibit thought and action: A theory of an act of control. Psychological Review, 91, 295–327.Logan, G. D., Cowan, W. B., & Davis, W. B. (1984). On the ability to inhibit simple and choice reaction time responses: A model and a method. Journal of

Experimental Psychology: Human Perception and Performance, 10, 276–291.Logan, G. D., & Zbrodoff, N. J. (1998). Stroop-type interference. Congruity effects in color naming with typewritten responses. Journal of Experimental

Psychology: Human Perception and Performance, 24, 978–992.MacLeod, C. M. (1991). Half a century of research on the Stroop effect: An integrative review. Psychological Bulletin, 109, 163–203.MacLeod, C. M., & MacDonald, P. A. (2000). Interdimensional interference in the Stroop effect: Uncovering the cognitive and neural anatomy of attention.

Trends in Cognitive Sciences, 10, 383–391.Mathis, A., Schunck, T., Erb, G., Namer, I. J., & Luthringer, R. (2009). The effect of aging on the inhibitory function in middle-aged subjects: A functional MRI

study coupled with a color-matched Stroop task. International Journal of Geriatric Psychiatry, 24, 1062–1071.Milham, M. P., Erickson, K. I., Banich, M. T., Kramer, A. F., Webb, A., Wszalek, T., et al (2002). Attentional control in the aging brain: Insights from an fMRI

study of the Stroop task. Brain and Cognition, 49, 277–296.Miller, E. K., & Cohen, J. D. (2001). An integrative theory of prefrontal cortex function. Annual Review of Neuroscience, 24, 167–202.Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., & Wager, T. D. (2000). The unity and diversity of executive functions and their

contributions to complex ‘‘frontal lobe” tasks: A latent variable analysis. Cognitive Psychology, 41, 49–100.Monsell, S. (2003). Task switching. Trends in Cognitive Sciences, 7, 134–140.Monsell, S., Taylor, T. J., & Murphy, K. (2001). Naming the color of a word: Is it responses or task sets that compete? Memory and Cognition, 29, 137–151.Rogers, R. D., & Monsell, S. (1995). Costs of a predictable switch between simple cognitive tasks. Journal of Experimental Psychology: General, 124, 207–231.Shallice, T., & Norman, D. (1986). Attention to action: Willed and automatic control of behavior. In R. Davidson, G. Schwartz, & D. Shapiro (Eds.).

Consciousness and self-regulation: Advances in research and theory (Vol. 4, pp. 1–18). New York: Plenum Press.Steinhauser, M., & Hubner, R. (2009). Distinguishing response conflict and task conflict in the Stroop task: Evidence from ex-Gaussian distribution analysis.

Journal of Experimental Psychology: Human Perception and Performance, 35, 1398–1412.Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of Experimental Psychology, 18, 643–662.Tzelgov, J., Henik, A., & Berger, J. (1992). Controlling Stroop effects by manipulating expectations for color words. Memory and Cognition, 20, 727–735.Verbruggen, F., & Logan, G. D. (2008). Response inhibition in the stop-signal paradigm. Trends in Cognitive Sciences, 12, 418–424.Verbruggen, F., & Logan, G. D. (2009). Models of response inhibition in the stop-signal and stop-change paradigms. Neuroscience and Biobehavioral Reviews,

33, 647–661.Verbruggen, F., Logan, G. D., & Stevens, M. A. (2008). STOP-IT: Windows executable software for the stop-signal paradigm. Behavior Research Methods, 40,

479–483.Waszak, F., Hommel, B., & Allport, D. A. (2003). Task switching and long-term priming: Role of episodic stimulus-task bindings in task-shift costs. Cognitive

Psychology, 46, 361–413.Wilkinson, D. T., & Halligan, P. W. (2004). The relevance of behavioural measures for functional imaging studies of cognition. Nature Reviews Neuroscience, 5,

67–73.


30

31

Chapter II

Experiment Group 3

Evidence for Interaction between the Stop Signal and the

Stroop Task Conflict

Eyal Kalanthroff, Liat Goldfarb, & Avishai Henik

Journal of Experimental Psychology: Human Perception and Performance

2013

Evidence for Interaction Between the Stop Signal and the StroopTask Conflict

Eyal KalanthroffBen-Gurion University of the Negev

Liat GoldfarbPrinceton University

Avishai HenikBen-Gurion University of Negev

Performance of the Stroop task reflects two conflicts—informational (between the incongruent word andink color) and task (between relevant color naming and irrelevant word reading). The task conflict isusually not visible, and is only seen when task control is damaged. Using the stop-signal paradigm, a fewstudies demonstrated longer stop-signal reaction times for incongruent trials than for congruent trials.This indicates interaction between stopping and the informational conflict. Here we suggest that“zooming in” on task-control failure trials will reveal another interaction—between stopping and taskconflict. To examine this suggestion, we combined stop-signal and Stroop tasks in the same experiment.When participants’ control failed and erroneous responses to a stop signal occurred, a reverse facilitationemerged in the Stroop task (Experiment 1) and this was eliminated using methods that manipulated theemergence of the reverse facilitation (Experiment 2). Results from both experiments were replicatedwhen all stimuli were used in the same task (Experiment 3). In erroneous response trials, only the taskconflict increased, not the informational conflict. These results indicate that task conflict and stop-signalinhibition share a common control mechanism that is dissociable from the control mechanism activatedby the informational conflict.

Keywords: task conflict, inhibition, executive function, Stroop task, stop signal

Our cognitive resources are limited. This often forces us to focuson one object, dimension, situation, or task, and to ignore, inhibit,or defer other aspects for later processing. To this end, we exertcognitive control (or executive functioning), a major vehicle ofwhich is inhibition. Van Veen and Carter (2006) suggested thatcontrol should be conceptualized as the ability to suppress irrele-vant information, and Verbruggen and Logan (2008) argued thatresponse inhibition is a hallmark of executive function. It is gen-erally believed that control is not unitary, and that different controlprocesses are required in different tasks (Miyake et al., 2000).

The current study used a combination of two inhibitory tasks:stop-signal (Logan, 1994; Logan & Cowan, 1984) and Stroop(Stroop, 1935). Both tasks demand the suppression of one processin order to successfully execute the other. Our objective was toinvestigate the underlying mechanisms activated in a stoppingsituation. In particular, we examined the task conflict controlmechanism (Goldfarb & Henik, 2007). This was carried out by“zooming in” on trials in which the control mechanism failed in

one task (stop-signal) in order to examine what happens in theother task (Stroop).

We first briefly review each task and then discuss the commonproperties of them as well as possible interactions between them.

The Stroop Task

In the classic color�word Stroop task (Stroop, 1935), individ-uals are asked to identify the ink color in which a word is printed(i.e., word color) while ignoring the word meaning. Word colorand word meaning can be either congruent (e.g., RED written inred), incongruent (e.g., BLUE written in red), or neutral (e.g.,XXXX written in red). Commonly, reaction time (RT) for incon-gruent trials is the longest while RT for congruent trials is theshortest (see MacLeod, 1991, for a review of Stroop task research).The Stroop task is particularly useful for investigating executiveaspects of attentional control. Due to the automatic tendency toread a word (MacLeod & MacDonald, 2000), the Stroop taskrequires not only the ability to attend to one attribute over another,but also requires executive aspects of attentional control to inhibitautomatic word reading while prioritizing the less automatic wordcolor recognition. Shallice (1982) argued that inhibition of routineprocesses is one of the most important aspects of executive atten-tional control.

It is commonly assumed that when word meaning and wordcolor are incongruent, a conflict arises and a cognitive controlmechanism is recruited to settle this conflict. Hence, the RT forincongruent trials is commonly longer than RT for neutral trials.This is known as the interference effect. In congruent trials, both

This article was published Online First March 5, 2012.Eyal Kalanthroff and Avishai Henik, Department of Psychology and

Zlotowski Center for Neuroscience, Ben-Gurion University of the Negev,Beer Sheva, Israel; Liat Goldfarb, Department of Psychology and Centerfor the Study of Brain, Mind, and Behavior, Princeton University.

Correspondence concerning this article should be addressed to EyalKalanthroff, Department of Psychology, Ben-Gurion University of theNegev, P.O.B. 653, Beer Sheva 84105, Israel. E-mail: [email protected]

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

Journal of Experimental Psychology: © 2012 American Psychological AssociationHuman Perception and Performance2013, Vol. 39, No. 2, 579–592

0096-1523/13/$12.00 DOI: 10.1037/a0027429

579

32

word color and word meaning activate the same response; hence,RT is usually even faster than in neutral trials. This is known as thefacilitation effect (MacLeod, 1991). The facilitation effect is muchsmaller and more fragile than the interference effect. Commonly,the interference effect is referred to as evidence only for the firsttype of conflict—the informational conflict—which is the conflictbetween the contradictory information that arises from the irrele-vant word meaning and the relevant word color. Kornblum (1994)and Kornblum and Lee (1995) suggested a full taxonomy ofconflict types. According to this taxonomy, the informationalconflict can be conceptualized as Type 8 ensembles that are“characterized by having the response set overlap with both therelevant and the irrelevant stimulus dimensions which themselvesoverlap” (Kornblum, 1994, p. 131). In terms of this theory, theinformational conflict consists of both “response conflict” and“coding conflict.”

There is also a potential for a second conflict. A few researchershave suggested that stimuli acquire associations with the tasks inwhich they occur. Hence, stimuli have the ability to evoke theperformance of a task that has a strong association with it (Allport& Wylie, 2000; Rogers & Monsell, 1995; Waszak, Hommel, &Allport, 2003). As mentioned earlier, MacLeod and MacDonald(2000) argued that words are strongly associated with the task ofreading, and they trigger an automatic tendency to read a writtenword. Additionally, both interference and facilitation effects indi-cate that word reading has taken place, despite the fact that testsubjects are instructed to ignore the word meaning. This under-scores the potential for a second conflict: the task conflict, namely,a conflict between two tasks—the relevant color naming task andthe irrelevant word reading task, which was triggered by thestimulus (i.e., the word). Because the irrelevant word reading taskarises whenever a real word is presented, the task conflict willpresumably appear equally in congruent and incongruent trials. Inmost reports, the conflict referred to is informational. ConsideringKornblum’s taxonomy of conflicts type (Kornblum, 1994; Korn-blum & Lee, 1995): Task conflict can be mistakenly conceptual-ized as a conflict at the response level (conflict between twooptional responses). Steinhauser and Hubner (2009) showed anempirical dissociation between task conflict and response conflictin the Stroop task. They modeled RTs with the ex-Gaussiandistribution and demonstrated that task conflict was mainly shownin the exponential component, while the response conflict wasmainly shown in the Mu parameter of the Gaussian component.This serves as evidence that task conflict is not a conflict at theresponse level, but rather a conflict at the processing level. Wesuggest that the task conflict fits Kornblum’s Type 4 ensemble bestbecause the conflict occurs at the coding level and the irrelevanttask does not overlap with the response. Presentation of the stim-ulus automatically activates two task codes because both arepotential candidates to be the relevant dimension.

Neuroimaging studies also support the task conflict assumption.Some studies have indicated greater anterior cingulate cortex(ACC) activation in congruent and incongruent trials comparedwith neutral trials (e.g., Bench et al., 1993; Carter, Mintun, &Cohen, 1995; Milham et al., 2002, for older participants). Severalresearchers have proposed that the ACC monitors conflict situa-tions and signals this information to different structures in thefrontal cortex. This allows for making strategic adjustments incognitive control, meant to maintain goal-directed behavior (Bot-

vinick, Braver, Barch, Carter, & Cohen, 2001; Botvinick, Nys-trom, Fissell, Carter, & Cohen, 1999; Carter, Botvinick, & Cohen,1999; Carter et al., 1998). According to this notion, the findingsthat indicate greater ACC activation in nonneutral trials imply thatcongruent trials, as well as incongruent trials, cause a conflict.More specifically, these findings also indicate a task conflict inboth congruent and incongruent trials in the Stroop task.

However, in contrast to what would be expected from theneuroimaging results, behavioral findings imply that the congruentcondition causes the same or less conflict than the neutral condi-tion; thus the RT for the congruent condition is commonly faster orsimilar to the RT for the neutral condition. Goldfarb and Henik(2007) argued that the reason for this contradiction is the existenceof an active control mechanism aimed at eliminating the taskconflict. Thus, there is no behavioral expression for the taskconflict because solving the task conflict delays responding for acongruent trial for only a short time, keeping the RT still shorterthan in the neutral condition. Accordingly, Goldfarb and Heniktried to put the task conflict guard to sleep. This was done byincreasing the proportion of nonword neutral trials to 75% and byproviding a valid cue for whether the coming trial would be neutralor have a conflict (i.e., congruent or incongruent) in 50% of thetrials. These manipulations were conducted in order to reduce thetask conflict control. Indeed, increasing the nonword neutral pro-portions resulted in a reversed facilitation effect—in which thecongruent RT was longer than the neutral RT. Moreover, thereversed facilitation effect was larger, due to longer RTs inthe congruent trials, in the noncued condition, indicating that themuch needed control was low in the these trials. Results from thisstudy provide behavioral evidence for the existence of task con-flict. because the proportions and cueing were the cause for thechanges in RTs in congruent trials, these results provide evidencefor a task conflict between the word reading task and the colornaming task.

Another example of task conflict was reported recently by LaHeij, Boelens, and Kuipers (2010). They asked children to namethe color of namable objects (e.g., house, coat) and abstract forms.Like Prevor and Diamond (2005) before them, La Heij et al. founda difference in RT between the namable forms and the abstractforms. This effect was not found in adults. La Heij et al. showedthat this effect is preserved when using difficult-to-name objectsand thus suggested that the object-interference effect is not due tocompetition between lexical representations, but to competitionbetween two task sets: color naming and object naming. Theabsence of the effect in adult participants was attributed to theirmore efficient inhibitory control.

In a recent study, Kalanthroff and Henik (2011) manipulatedcue-target intervals (CTI) in a task-switching Stroop task. In thistask, participants were given a cue about the task in the comingtrial—whether the next target would require word reading or colornaming. By frequently changing task identity, task switching wasexpected to increase task conflict. Furthermore, although the pro-cesses that occur during the CTI are still debated, there is muchagreement that the CTI is used to recruit control, at least to someextent (Meiran, 1996; Monsell, 2003). Indeed, with short CTIs areverse facilitation was found, indicating task conflict. Bravermanand Meiran (2010) also reported findings that emphasized theimportance of task conflict control in task switching. These recent

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

580 KALANTHROFF, GOLDFARB, AND HENIK

33

task conflict studies provide additional support for the suggestionthat task conflict effects are caused by the need to recruit control.

Thus, there is much evidence for the existence of a task conflict.It should be noted that task conflict and the existence of habitualresponses require certain control processes that may differ fromthose required for solving information conflict. The current studywill attempt to supply new information about the differentiationbetween these control mechanisms.

The Stop-Signal Task

The stop-signal task (Logan, 1994; Logan & Cowan, 1984)examines the ability to suppress an already initiated action that isno longer appropriate. A great deal of research has been done inthe past few decades with the stop-signal task (see Verbruggen &Logan, 2008, for review). In the classic task, participants are askedto address a visual stimulus (go signal) with a motor response asfast as possible. In about one third of the trials, an auditorystimulus (stop signal) comes right after the visual go signal.Participants are instructed to inhibit their motor response whenthey hear the stop signal. In the common version of the stop-signaltask, the duration between the visual stimulus and the stop-signaldelay (SSD) changes from one trial to the next and is based on theparticipant’s success in inhibiting his or her response. If the par-ticipant succeeds in response inhibition, the next stopping trial willbe more difficult (the SSD will be longer), and if the participantfails to inhibit his response, the next stopping trial will be easier(the SSD will become shorter). Eventually, it is possible to esti-mate the stop-signal RT (SSRT), which is the time needed forsuccessful inhibition. SSRT has proven to be an important measureof cognitive control (Verbruggen & Logan, 2008).

Logan and Cowan (1984) and Logan, Cowan, and Davis (1984)compared the performance in the stop-signal task to a horse racebetween the go process, triggered by the presentation of the gosignal (e.g., usually a visual stimulus), and the stop process,triggered by the stop signal (e.g., usually an auditory stimulus).According to this metaphor, the SSRT is the time needed for thestop process to finish, while the RT is the time required for the goprocess to finish. If the stop process, which starts the race laterthan the go process, finishes before the go process, there will be asuccessful stop. Logan et al. (1984) argued that “response inhibi-tion phenomena are consistent with a hierarchical theory of atten-tion in which a high level process determines the significance ofincoming stimuli and decides whether to abort the current streamof thought and action or to queue the new stimuli along with theold ones, to be processed as resources become available” (p. 290).In these terms, the stop process represents high-level control,whereas the go process is a more automatic process that might bestopped sometimes. This notion leads us to the conclusion that anerroneous response to a stop signal can occur because (a) theprimary go process was extremely fast or (b) the inhibitory controlwas momentarily less efficient and produced slower stopping.Because control is an effortful process, one can reasonably assumethat, in many trials, an erroneous response to a stop signal willresult from less efficient performance of the control mechanism.

Most neuroimaging studies have indicated that stopping ismainly associated with activation of the inferior frontal gyrus(IFG; ventrolateral prefrontal cortex), and the presupplementarymotor area (pre-SMA) (Aron, Behrens, Smith, Frank, & Poldrack,

2007; Aron, Fletcher, Bullmore, Sahakian, & Robbins, 2003;Chambers et al., 2007; Chevrier, Noseworthy, & Schachar, 2007;Rubia, Smith, Brammer, & Taylor, 2003). Though the IFG and thepre-SMA seem to be most important for triggering the stop pro-cess, the ACC has also been found to be a crucial component ofstopping. Using electrophysiological measurement, van Boxtel,van der Molen, and Jennings (2005) found different ACC involve-ment in erroneous responses to stop-signal trials compared tosuccessful stopping trials. Brown and Braver (2005) found thatACC activity could predict the likelihood of an erroneous responseoccurring in a stop-signal task. Considering the notion that theACC, via the magnitude of its activity, signals the conflict situa-tion, these findings indicate that the stopping creates a conflict thatis being monitored by the ACC (see also Eimer, 1993; Jodo &Kayama, 1992; Nieuwenhuis, Yeung, Wildenberg, & Ridderink-hof, 2003, for electrophysiological evidence of ACC involvementin monitoring the conflict caused by the need to inhibit respond-ing).

ACC activation in stop-signal trials, as in the nonneutral Strooptrials, signals a conflict situation and the need to inhibit a response.These neuroimaging findings indicate that some overlap existsbetween the Stroop task and the stop-signal task, mainly throughthe activation of the ACC in stop-signal conditions and in non-neutral Stroop conditions.

Combined Stop-Signal and Stroop Tasks

Several studies have investigated control mechanisms by com-bining stop-signal and various control tasks. Though some founddissociation (Logan, 1981; Scheres et al., 2003), most found in-teraction between stop-signal inhibition and interference control(e.g., Chambers et al., 2007, found behavioral interaction andanatomical dissociation; Kramer, Humphrey, Larish, Logan, &Strayer, 1994; Ridderinkhof, Band, & Logan, 1999; Verbruggen,Liefooghe, & Vandierendonck, 2004, 2006). To our knowledge,only Verbruggen et al. (2004) have combined the Stroop task andthe stop-signal task. They found longer SSRT for incongruenttrials than for neutral trials. Using a combination of a stop-signaland flanker task, they also found longer SSRT for incongruenttrials than for neutral and congruent trials. These findings indicateinteraction between the informational conflict and stopping. Morespecifically, it seems to be more difficult to stop in trials thatcontain informational conflict. The lack of a Stroop congruentcondition in Verbruggen et al.’s (2004) study does not make itpossible to reach a valid conclusion about the interaction betweenstop-signal inhibition and task conflict. Considering the efficiencyof the task control mechanism, we can assume that SSRT incongruent trials will not be longer than SSRT in neutral trials. Inother words, no behavioral effect of task conflict on SSRT will bevisible. The fact that task conflict is not visible does not mean it isnot there. As mentioned before, task conflict is commonly notvisible, and a complicated manipulation or control damage isneeded to reveal it (e.g., Goldfarb & Henik, 2007; Kalanthroff &Henik, 2011). We believe that a more complicated manipulationand analysis will be necessary to reveal the interaction betweenstopping and task conflict.

There are theoretical bases to the assumption that there is acommon mechanism for task conflict and stopping, and that theinteraction does exist. As mentioned earlier, in the Stroop task the

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

581INTERACTION BETWEEN STOP SIGNAL AND TASK CONFLICT

34

word reading is automatic. Hence, one needs to inhibit or overridethe tendency to carry out the more dominant or automatic wordreading. For that reason, Miyake et al. (2000) considered that theStroop task required inhibition of prepotent responses. The taskconflict reasoning seems consistent with this idea. The stop-signaltask is also a form of inhibition of prepotent responses. Thus, it isreasonable to assume that both the Stroop task conflict and theconflict created by the stop-signal task are monitored by a similarcontrol mechanism. Accordingly, it is also possible that controlmechanisms (and ACC activation) recruited for one task serve theother task as well. In other words, failure in the task conflictcontrol mechanism will cause failure to inhibit a prepotent re-sponse, which in turn will cause both delay in the Stroop taskconflict resolution and an erroneous response to a stop signal. LaHeij and Boelens (2011) recently suggested the object interferenceeffect was a result of children’s immature inhibition of a prepotentresponse mechanism. Considering this notion, that the object in-terference effect is caused by task conflict, which was mentionedearlier, gives us another indication of the overlap between taskconflict and stopping.

The current study aimed to expose the task conflict and toexamine the control mechanisms that are recruited to solve thisconflict. We were also interested in the interaction between thetask conflict and stop-signal inhibition. We hypothesized that someoverlap exists between the control mechanism that is responsiblefor the efficient solving of the Stroop task conflict and the controlmechanism that is responsible for stopping. This is in addition toand separate from the overlap that is known to exist between theinformational control mechanism and stopping control (Verbrug-gen et al., 2004).

As mentioned, exposing the overlap between Stroop task con-flict and stopping requires the use of more complex manipulationand analysis. This forces us to find a way to zoom in on trials thatpotentially have some extent of control failure, as did previousstudies that exposed the task conflict (e.g., Goldfarb & Henik,2007; Kalanthroff & Henik, 2011; La Heij et al., 2010). One suchway is to analyze the RT of erroneous responses to stop-signaltrials, but only after finding a stable SSD for each specific partic-ipant. Because participants are known to be able to stop at thisstable SSD, the erroneous responses to stop-signal trials potentiallycontain control failure or at least an incomplete control operation.If erroneous responses represent control failure and some overlapexists between task conflict control and stopping control, thenwhen participants fail to inhibit their response after a stop signal,the task conflict will become visible, namely, we will see a reversefacilitation effect.

It is our belief that exposing the existence of this controlmechanism is important for: (a) understanding of the underratedtask conflict and (b) revealing the nature of the interaction betweenstop-signal inhibition and Stroop interference.

Experiment 1

In Experiment 1, we combined the Stroop task with the stop-signal task, as did Verbruggen et al. (2004), but with two importantdifferences: (a) we analyzed the performance only after identifyinga stable SSD for each participant, and (b) we included a congruentcondition in the Stroop task. Adding a congruent condition, along-side neutral and incongruent conditions, will make it possible to

investigate Goldfarb and Henik’s (2007) suggestion about theexistence of a task conflict in congruent and incongruent trials,which is usually unseen because of a supposedly efficient taskconflict control mechanism. In particular, when stopping fails, sothat, potentially, control is not working properly, we expect areverse facilitation (i.e., a congruent RT longer than a neutral RT),as was found by Goldfarb and Henik. If this pattern of results isobtained, it will also indicate that erroneous response to a stopsignal results from a less efficient control mechanism, at least inmost of the trials. Further support for this will be discussed in theResults section.

Method

Participants. Fourteen, first-year psychology students (11women and 3 men) from Ben-Gurion University of the Negev,Israel, participated in return for partial fulfillment of course re-quirements and credit. All participants had normal or corrected-to-normal vision, were right-handed, had no history of attentiondeficit or dyslexia, were native speakers of Hebrew, and werenaive about the purpose of the experiment. Participants who hadmore than 10% of their responses with long RTs (3 standarddeviations or more above average RTs) were excluded from theanalysis because such RTs indicated the use of a waiting strategy(i.e., they waited to see if a stop signal would appear). Twoparticipants (one woman and one man) were excluded for thisreason.

Stimuli. Participants were presented with a two-color, man-ual Stroop task. Each stimulus consisted of one of two colorwords— (Hebrew for blue), (Hebrew for red)—or afour letter string in Hebrew— (meaningless repetition ofa Hebrew letter, parallel to XXXX in the English version). Themeaningless letter string would not trigger automatic reading, andas such, would cause no task conflict. The ink color was either redor blue. There were 6 combinations of words and ink colors: twocongruent, two neutral, and two incongruent. The words werepresented at the center of a screen on a black background and were0.98 in. high and 2.36 in. wide.

The stop signal was a loud and clear auditory stimulus (750 Hz,85 dB, 50 ms), delivered through earphones.

Procedure. Data collection and stimuli presentation werecontrolled using a computer with an Intel Pentium III centralprocessor. Stimuli were presented on a Compaq S510 monitor. Akeyboard was placed on a table between the participant and themonitor. Participants were tested individually. They sat approxi-mately 23.5 in. from the computer screen.

The experiment included 90 practice trials (which were notanalyzed) and 1,200 experimental trials. The different stimuli werepresented an equal number of times and randomly. During prac-tice, participants received feedback on accuracy and RT. Each trialstarted with a 1,000-ms fixation (a white “�” sign in the center ofa black screen), followed by a visual stimulus. Participants wereinstructed to hit the “b” key on the keyboard with the index fingerof their right hand if the ink color was blue, and to hit the “m” keywith the middle finger of the same hand if the ink color was red.Participants were asked to ignore the meaning of the word and topress the correct key as fast as possible, without making mistakes.The visual stimulus stayed in view for 2,500 ms or until a keypress. RT was calculated from the appearance of the visual stim-

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.


35

ulus to the reaction. Each trial ended with a 500-ms intertrialinterval.

A stop signal was sounded at random a quarter of the time foreach congruency condition, shortly after the appearance of thevisual stimulus. At the beginning of the experiment, the SSD was150 ms. Using a staircase method, after each successful stopping,the SSD was extended by 10 ms and after each unsuccessfulstopping the SSD was shortened by 10 ms. After each stoppingtrial, the SSD was adjusted for each congruency condition sepa-rately (e.g., after a successful stopping in a congruent trial, theSSD was extended only for the next congruent trial, not affectingthe SSD for other congruency conditions). To prevent subjectsfrom applying a “waiting for a stop-signal” strategy and to moti-vate them to work as fast as possible, instructions emphasized thatspeed and accuracy would be rewarded. Participants were intro-duced to a scoring system that awarded more points for a faster,correct response; no points for an incorrect response; and themaximum score for a correct stop. Extra required course credit waspromised to the participant who received the highest score.

Results and Discussion

Stroop and stop-signal data were analyzed separately. Data werecollected only from the moment there were at least two successfulstops and two unsuccessful stops within 10 trials, meaning thatthere were four direction changes in SSD within a window of 10trials (e.g., at least two extensions and two shortenings of SSD).This point was found for each congruency condition separately,and was selected because it indicated the point at which the SSDgraph started to straighten and became stable. Data before thatpoint were excluded from further analysis.

Stroop analysis. In trials without a stop signal, mean RTs ofcorrect responses were calculated for each participant in eachcondition. In many trials with a stop signal, participants failed toinhibit their response; namely, they responded to the Stroop stim-ulus. Mean RTs and accuracy were measured for erroneous re-sponse trials for each participant in each condition. A two-wayanalysis of variance (ANOVA) with repeated measures was ap-plied to RT data with congruency (congruent, neutral, and incon-gruent) and trial type (no stop signal and erroneous response tostop signal) as within-subject factors. As expected, a significantinteraction was found, F(2, 22) � 3.887, mean square error(MSE) � 983.038, p � .05, partial �2 � .261. In addition, maineffects were found for congruency, F(2, 22) � 7.735, MSE �824.185, p � .01, partial �2 � .413, and for trial type, F(1, 11) �10.377, MSE � 15,888.902, p � .01, partial �2 � .485.

Analyzing the no stop-signal trials separately revealed a strongcongruency effect (i.e., congruent RT vs. incongruent RT), F(1,11) � 11.064, MSE � 771.053, p � .01, partial �2 � .501,interference effect (i.e., incongruent RT vs. neutral RT), F(1,11) � 12.071, MSE � 324.217, p � .01, partial �2 � .523, and asmall facilitation effect (i.e., congruent RT vs. neutral RT), F(1,11) � 5.554, MSE � 159.956, p � .05, partial �2 � .336. Thesefindings replicate the common Stroop effect (MacLeod, 1991)(Figure 1 and Table 1).

Analyzing the erroneous response to stop-signal trials separatelyrevealed that participants responded slower to congruent than toneutral trials, F(1, 11) � 9.49, MSE � 851.41, p � .01, partial�2 � .463. In addition, a significant difference was also found

between neutral and incongruent trials, F(1, 11) � 5.046, MSE �1,811.853, p � .05, partial �2 � .314, but not between congruentand incongruent trials, F � 1. These results indicate a reversefacilitation effect. As in Figure 2 and Table 1, the shortest RT waspresent for neutral trials. Note that two different patterns of Stroopeffects were found within the same experiment. The regular Stroopeffect was found when no stopping was required (see Figure 1),whereas the task conflict effect, indicated by a reversed facilita-tion, was found when participants failed to stop (see Figure 2 andTable 1).

Of note, responses to incongruent trials did not differ signifi-cantly from those for congruent trials. Namely, there was noindication for an information conflict, and there was no indicationfor a failure in the informational conflict control. We believe thata likely explanation derives from the possible differentiation be-tween the task control and the informational control, meaning thateven though the task control was less efficient in those trials, theinformational conflict was normally resolved. In other words, thestrong task conflict masked the informational conflict. Usually ina Stroop task, the strong informational conflict masks the quicklyresolved task conflict. Here, we get exactly the opposite.

As in Table 1, RTs for trials without a stop signal are signifi-cantly longer than RTs for erroneous responses to stop-signaltrials. To ensure that our results were not an artifact of using onlypart of the RT distribution, but rather they demonstrated a Stroopeffect in a task control failure condition, we conducted two addi-tional analyses. The first was a one-way ANOVA of correct fastresponses. Similar to the Stroop analysis, for trials without a stopsignal, we analyzed only the faster half of the trials (for eachcongruency condition). Results indicated a Stroop effect (399, 404,and 413 ms for the congruent, neutral, and incongruent conditions,respectively). The results also indicated a strong congruency ef-fect, F(1, 11) � 12.12, MSE � 82.476, p � .01, interference effect,F(1, 11) � 5.19, MSE � 65.95, p � .05, and a marginallysignificant facilitation effect, F(1, 11) � 4.82, MSE � 35.74, p �.051.

The second analysis was a one-way ANOVA for the erroneousslow responses to a stop signal. We used only the slower half of thetrials. Again results were similar to the previous analysis (577,514, and 590 ms for the congruent, neutral, and incongruent

Figure 1. Mean reaction time (RT) for the congruency conditions ofStroop trials with no stop signal in Experiment 1. Error bars are onestandard error of the mean.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.


36

conditions, respectively). Similar to what was found in the full dataanalysis, congruent trials were slower than neutral trials, F(1,11) � 7.67, MSE � 3,111.87, p � .05. Neutral trials were fasterthan incongruent trials (this was only marginally significant), F(1,11) � 4.375, MSE � 7,957.65, p � .06, but no significantdifference was found between congruent and incongruent trials,F � 1.

That our common Stroop effect and our reverse facilitation werestill present in the faster and slower half of the trials, respectively,indicates that our previous results were not artifacts. Moreover,this also demonstrates that the pattern of results is stable evenwhen the go process is fast. This provides an indication that anerroneous response to a stop signal is consistent with a lessefficient stop process (and not just a fast go process), at least inmost of the trials.

Goldfarb and Henik (2007) have suggested the existence of taskconflict in the congruent and incongruent Stroop trials. This iscaused by the conflict between the requested ink recognition taskand the automatic reading task evoked by the strong associationwith the word stimulus. The results of the current experimentreinforce the existence of such a task conflict. Because theseresults refer only to the trials that occurred after the SSD becamestable, it is reasonable to assume that these erroneous responsesindicate a control failure, or at least an incomplete control opera-tion. More specifically, it seems that the manipulation and analysiswe used here enabled us to zoom in on task conflict control failuresituations, and by that, made the task conflict visible.

The number of erroneous responses were subjected to a one-wayANOVA with congruency as a within-subject factor (Figure 3). Ofnote, participants made more errors in the neutral than the con-gruent trials, F(1, 11) � 8.331, MSE � 5.121, p � .05, partial�2 � .431, and in the neutral than the incongruent trials, F(1,11) � 13.553, MSE � 4.439, p � .01, partial �2 � .552. Nosignificant differences were found between responses to congruentand incongruent trials, F � 1 (see Table 1). Even though we useda staircase tacking procedure, there were still more erroneousresponses to a stop signal in neutral trials after the SSD becamestable. The larger number of erroneous responses in the neutralcondition can be understood by assuming that these trials representa failure in the control mechanism, which is common for both theStroop task and the stop-signal task. In the neutral trials, there isonly one attempt to recruit task conflict control (i.e., at the appear-ance of the stop signal) while in nonneutral trials there are twoattempts to recruit control, with the first one occurring earlier thanthe appearance of the stop signal (i.e., when the Stroop stimulusappears). For this reason, it is much more likely that more failuresin control recruiting will occur in the neutral condition, in whichthere is only one late attempt to do so. This will be discussedfurther in Experiment 2, Experiment 3, and in the General Dis-cussion sections.

Stop-signal analysis. As mentioned earlier, the SSD wasadjusted for each congruency condition separately. SSRT wascalculated separately for each congruency condition. As before,data were collected only from the moment the SSD stabilized.

Table 1Results of Experiment 1

Variables

Trial type

Congruent Neutral Incongruent

RT (% errors) 538 (3) 551 (2.98) 576 (3.5)RT for erroneous responses to stop signal 473 434 471Number of erroneous responses to the stop signal 36.08 39.25 36.58SSRT 182 190 211

Note. RT in milliseconds. RT � reaction time; SSRT � stop-signal reaction time.

Figure 2. Mean reaction time (RT) for the congruency conditions oferroneous responses in stop-signal trials in Experiment 1. Error bars areone standard error of the mean.

Figure 3. Number of erroneous responses to the stop signal for thecongruency conditions in Experiment 1. Error bars are one standard errorof the mean.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.


37

Considering the fact that p (response|signal) was not .50, we usedthe integration method to calculate the SSRT (Verbruggen &Logan, 2008). No-signal go RTs were determined by the nth RT,that is, n (number of no-signal trials) � p (response|signal), whichwas done for each congruency condition separately. SSRT wasthen calculated as the nth RT – median SSD. SSRT data weresubjected to a one-way ANOVA, with congruency as a within-subject factor. There was no significant difference between theSSRT of the three congruency conditions (182, 190, and 211 ms,for the congruent, neutral, and incongruent conditions, respec-tively), F � 1 (see Table 1). There was no speed–accuracy trade-off, as indicated by the fact that none of the correlations betweenRTs and number of errors reached significance.

These results were in the same direction as the results reportedby Verbruggen et al. (2004). As in their experiment, we foundlonger SSRT in incongruent trials than in neutral trials (there wasno congruent Stroop condition in their study), but this differencewas not significant. We will discuss this issue after Experiment 3,when we combined the data from both Experiments 1 and 3.

In sum, in Experiment 1, we analyzed only trials that appearedafter the SSD stabilized. We found a common Stroop effect fortrials without a stop signal and a reverse facilitation in erroneousresponses to stop-signal trials. This indicated a common controlmechanism for the task conflict and for stop-signal inhibition. Wereplicated this effect using only the slower half of the trials,indicating that our effect was not an artifact. We also found thatincongruent trials were more difficult to inhibit, though this resultwas not significant. To be sure that our results were due to a failurein task control in erroneous responses to stop-signal trials, weconducted Experiment 2.

Experiment 2

In Experiment 1, we found that when there was a failure in aspecific control mechanism, there was both an erroneous responseto a stop signal and a Stroop reverse facilitation. Experiment 2 wasconducted to verify that nothing but the task conflict control failureproduced the effect observed in Experiment 1. To do this, wereplaced the homogenous letter-string neutral stimulus in Experi-ment 1 with a noncolor word. This would still allow for noinformational conflict in the neutral condition.

According to MacLeod and MacDonald (2000), words raise anautomatic tendency to read, thus a noncolor word should activatean automatic word reading response (the irrelevant task), whereasa string of letters should not. Goldfarb and Henik (2007) found thatthe reverse facilitation effect disappeared when a real noncolorword was used as a neutral. Hence, in Experiment 2, we expectedthat the reverse facilitation effect found in Experiment 1 woulddisappear.

Method

Participants. Twelve, first-year psychology students (10women and 2 men) of Ben-Gurion University of the Negev, Israel,who did not take part in Experiment 1, participated for partialfulfillment of course requirements and credit. All participants hadnormal or corrected-to-normal vision, were right-handed, had nohistory of attention deficit or dyslexia, were native speakers ofHebrew, and all were naive about the purpose of the experiment.

Stimuli. In Experiment 2, the meaningless homogenous letterstring was replaced with the Hebrew noncolor word

(building), which matched by length, word frequency, and didnot begin with the same letter as any of the color words (see alsoGoldfarb & Henik, 2007). Except for this, all the stimuli inExperiment 2 were identical to those in Experiment 1.

Procedure. The procedure was identical to that of Experi-ment 1.


Stroop and stop-signal data were analyzed separately. As inExperiment 1, we only analyzed data from the moment there wereat least two successful stops and two unsuccessful stops within 10trials (i.e., when SSD stabilized).

Stroop analysis. In trials without a stop signal, mean RTs ofcorrect responses were calculated for each participant in eachcondition. As in Experiment 1, mean RTs and accuracy weremeasured for the erroneous response trials for each participant ineach condition. A two-way ANOVA with repeated measures wasapplied to RT data with congruency (congruent, word neutral, andincongruent) and trial type (no stop signal and erroneous responseto stop signal) as within-subject factors. As expected, a significantinteraction was found, F(2, 22) � 10.621, MSE � 2,946.003, p �.001, partial �2 � .491. Additionally, main effects were found forcongruency F(2, 22) � 13.929, MSE � 2,895.902, p � .001,partial �2 � .559, and for trial type, F(1, 11) � 35.741, MSE �477.898, p � .001, partial �2 � .765.

Analyzing the no stop-signal trials separately revealed a strongcongruency effect, F(1, 11) � 8.24, MSE � 688.060, p � .01,partial �2 � .428, and a strong interference effect, F(1, 11) � 9.63,MSE � 74.62, p � .01, partial �2 � .467. The facilitation effect inExperiment 2 was only marginally significant, F(1, 11) � 3.38,MSE � 696.18, p � .09, partial �2 � .235. These findings aresimilar to those of other Stroop studies (MacLeod, 1991) (Table 2and Figure 4).

The most important finding of this experiment was that, whenanalyzing the erroneous response to stop-signal trials separately,there was no significant difference between the RTs of the threecongruency conditions (411, 421, and 422 ms, for the congruent,neutral, and incongruent conditions, respectively), F(2, 22) �1.066, MSE � 447.94, p � .362. The observed power for this nulleffect was .212 (see Table 2 and Figure 5). As expected, using awritten word in the neutral condition eliminated the reverse facil-itation effect. As in Experiment 1, responses in incongruent trialsdid not differ significantly from responses in congruent trials, F(1,11) � 1.22, MSE � 649.82, p � .29. This strengthens our previousassumption that this might occur because in those trials the infor-mational conflict was normally resolved, and was masked by theslowly solved task conflict. This also indicates no failure in theinformational conflict control in these trials. These results fitthe notion that failure in task conflict control is distinguished fromfailure in informational conflict control. This indicates a separatecontrol mechanism for the informational conflict.

The number of erroneous responses were subjected to a one-wayANOVA, with congruency as a within-subject factor (see Figure6), and were not significant (31.3, 30.8, and 31.0, for the congru-ent, neutral, and incongruent conditions, respectively), F(2, 22) �1 (see Table 2). The observed power for this null effect was .082.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.


38

The effect of the number of erroneous responses was also elimi-nated by changing the neutral condition to a real word. Theseresults can be explained using the same explanation as used inExperiment 1. In Experiment 1, nonword neutral trials only hadone late attempt to recruit control (i.e., the appearance of the stopsignal), which caused a larger number of erroneous responses. InExperiment 2, the word neutral trials, like the congruent andincongruent trials, included another earlier attempt to recruit thiscontrol mechanism (the appearance of the written word), whichenabled less errors. This notion also corresponds with our findingsthat show that at least some overlap exists between task controland stopping.

Stop-signal analysis. SSRT was calculated as in Experiment1. SSRT data were subjected to a one-way ANOVA, with congru-ency as a within-subject factor. There was only a marginallysignificant difference between the SSRT of the three congruencyconditions (190, 194, and 218 ms, for the congruent, neutral, andincongruent conditions, respectively), F(2, 22) � 3.12, MSE �295.978, p � .09, partial �2 � .28. Again, we found longer SSRTin incongruent trials than in neutral trials, and in this experimentthis difference was marginally significant, F(1, 11) � 3.96,MSE � 82.254, p � .078, partial �2 � .279 (see Table 2). Therewas no speed–accuracy trade-off, as indicated by the fact that noneof the correlations between RT and number of errors reachedsignificance.

Experiment 3

In the first two experiments, we suggested that a neutral word,but not a neutral meaningless letter string, causes a task conflict

between the relevant color naming task and the irrelevant yetautomatic word reading task. In each experiment, we used onlyone type of neutral. To exclude any alternative explanations relatedto that design, in Experiment 3, we used the two types of neutralsin the same block. This design would allow us to replicate ourresults in a combined block and, by that, demonstrate how robustour effects were.

Method

Participants. Eighteen, first-year psychology students (11women and 7 men) of Ben-Gurion University of the Negev, Israel,who did not take part in Experiments 1 and 2, participated forsmall monitory payment. All participants had normal or corrected-to-normal vision, were right-handed, had no history of attentiondeficit or dyslexia, were native speakers of Hebrew, and all werenaive about the purpose of the experiment.

Stimuli. In this experiment, we used both the Hebrew non-color word (building) and the meaningless letter string

(i.e., neutral conditions from both previous experi-ments). Except for these differences, all the stimuli in Experiment3 were identical to those in Experiments 1 and 2.

Procedure. To ensure that the task control would not beincreased by a high proportion of real words, the meaninglesshomogenous letter string appeared in a third of the trials.The remaining conditions—congruent, word-neutral, and incon-gruent—were each represented in 22.2% of the trials. Except forthese changes, the procedure was identical to those in Experiments1 and 2.

Figure 4. Mean reaction time (RT) for the congruency conditions ofStroop trials with no stop signal in Experiment 2. Error bars are onestandard error of the mean.

Figure 5. Mean reaction time (RT) of erroneous responses to the stopsignal for the congruency conditions in Experiment 2. Error bars are onestandard error of the mean.


Variables

Trial type

Congruent Neutral Incongruent

RT (% errors) 493 (4.3) 513 (4.0) 524 (4.7)RT for erroneous responses to stop signal 411 421 422Number of erroneous responses to the stop signal 31.3 30.8 31SSRT 193 194 217

Note. RT in milliseconds. RT � reaction time; SSRT � stop-signal reaction time.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.


39


Stroop and stop-signal data were analyzed separately. As in thefirst two experiments, we only analyzed data from the momentthere were at least two successful stops and two unsuccessful stopswithin 10 trials (i.e., when SSD stabilized).

Stroop analysis. In trials without a stop signal, mean RTs ofcorrect responses were calculated for each participant in eachcondition. As in the first two experiments, mean RTs and accuracywere measured for the erroneous response trials for each partici-pant in each condition. A two-way ANOVA with repeated mea-sures was applied to RT data with congruency (congruent, non-word neutral, word neutral, and incongruent) and trial type (nostop signal and erroneous response to stop signal) as within-subjectfactors. As expected, a significant interaction was found, F(3,51) � 6.104, MSE � 946.710, p � .001, partial �2 � .264. Inaddition, main effects were found for congruency, F(3, 51) �7.901, MSE � 1,282.906, p � .001, partial �2 � .317, and for trialtype, F(1, 17) � 62.592, MSE � 2,388.756, p � .001, partial �2 �.786 (Table 3 and Figure 7).

As in the previous experiments, analyzing the no stop-signaltrials separately revealed a strong congruency effect, F(1, 17) �6.239, MSE � 582.035, p � .05, partial �2 � .268, and aninterference effect (i.e., incongruent RT vs. nonword neutral RT),F(1, 17) � 4.448, MSE � 768.015, p � .05, partial �2 � .215. Thefacilitation effect (i.e., congruent RT vs. nonword neutral RT) inthis experiment was not significant, F(1, 17) � 1. These findingsare similar to those of other Stroop studies (MacLeod, 1991).

There was no significance difference between RTs of nonwordneutral and word neutral conditions, F(1, 17) � 1 (see Table 3 andFigure 8).

Analyzing the erroneous response to stop-signal trials separatelyrevealed the most important finding of this experiment. Similar toExperiment 2, there was no significant difference between the RTsof three congruency conditions (421, 446, and 443 ms, for thecongruent, word neutral, and incongruent condition, respectively),F(2, 34) � 1.714, MSE � 1,976.41, p � .195. Similar to Exper-iment 1, erroneous responses to a stop signal in the nonwordneutral condition were faster than in the congruent condition, F(1,17) � 13.734, MSE � 855.053, p � .005, partial �2 � .447, andincongruent condition, F(1, 17) � 15.014, MSE � 2,034.812, p �.001, partial �2 � .469. In addition, erroneous responses to a stopsignal in the nonword neutral condition were also faster than in theword neutral condition, F(1, 17) � 22.467, MSE � 1,501.603, p �.001, partial �2 � .569 (see Figure 8 and Table 3). As in the firsttwo experiments, responses in incongruent trials did not differsignificantly from responses in congruent trials, F(1, 17) � 1.861,MSE � 2,370.7, p � .19.

As in Experiment 1, we wanted to ensure that our results werenot artifacts of using only part of the RT distribution. Hence, weconducted two more analyses that were similar to those carried outin Experiment 1. The first was a one-way ANOVA for the fasterhalf of the nonsignal trials (for each congruency condition). Re-sults indicated a common Stroop effect (371, 375, 374, and 383 msfor the congruent, nonword neutral, word neutral, and incongruentconditions, respectively). There was a strong congruency effect,F(1, 17) � 13.38, MSE � 85.5, p � .005, interference effect (usingword neutral), F(1, 17) � 5.334, MSE � 106.501, p � .05, and amarginally significant facilitation effect (using word neutral), F(1,17) � 2.701, MSE � 36.956, p � .11. There was no significantdifference between the two neutral conditions, F � 1.

The second analysis was a one-way ANOVA for the slower halfof the erroneous responses to stop-signal trials. Again, results weresimilar to the previous analysis (488, 461, 511, and 510 ms for thecongruent, nonword neutral, word neutral, and incongruent condi-tions, respectively). Similar to what was found in the full dataanalysis, there was no significant difference between the threeword conditions, F � 1. Congruent trials were slower than non-word neutral trials, F(1, 17) � 5.262, MSE � 1,197.114, p � .05.Nonword neutral trials were faster than incongruent trials, F(1,17) � 6.707, MSE � 3,142.486, p � .05, and word neutral trials,F(1, 17) � 9.078, MSE � 2,413.219, p � .01.


Variables

Trial type

Congruent Nonword neutral Word neutral Incongruent

RT (% errors) 482 (2) 484 (1.4) 485 (2.6) 502 (3.1)RT for erroneous responses to stop signal 421 384 446 443Proportions of erroneous responses to the stop signal 41.1 47.6 41.4 40.1SSRT 190 192 194 212SSRT (Exp. 1 � Exp. 3) 187 192 — 212

Note. RT in milliseconds. RT � reaction time; SSRT � stop-signal reaction time; Exp. � Experiment.

Figure 6. Number of erroneous responses to the stop signal for thecongruency conditions of Experiment 2. Error bars are one standard errorof the mean.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.


40

These analyses clarify that our results were not artifacts, andthey strengthen our assumptions that erroneous responses to a stopsignal are consistent with a less efficient stop process (and not justa fast go process), at least in most of the trials.

In Experiment 3, because the number of trials in each congruentcondition was not constant, we used the proportion of erroneousresponses. The proportion of erroneous responses were subjectedto a one-way ANOVA, with congruency as a within-subject factor(see Figure 9). A significant effect was found, F(3, 51) � 6.531,MSE � .003 p � .001, partial �2 � .278. A one-way ANOVA forproportion of erroneous responses of only congruent, word neutral,and incongruent conditions revealed no significant differences,F(2, 34) � 1. In contrast, the proportion of erroneous responses inthe nonword neutral condition were significantly higher than in thecongruent, F(1, 17) � 11.762, MSE � 0.003, p � .005, partial�2 � .409, word neutral, F(1, 17) � 9.534, MSE � 0.004, p � .01,partial �2 � .359, and incongruent conditions, F(1, 17) � 13.92,MSE � 0.003, p � .005, partial �2 � .45 (see Table 3). Again, theeffect of the proportion of erroneous responses was eliminatedwhen a real word was used as a neutral.

Stop-signal analysis. SSRT was calculated as in the previousexperiments. SSRT data were subjected to a one-way ANOVA,with congruency as a within-subject factor. There was a nonsig-nificant trend for the difference between the SSRT of the congru-ency conditions (190, 192, 194, and 212 ms, for the congruent,nonword neutral, word neutral, and incongruent conditions, re-spectively), F(3, 51) � 1.819, MSE � 1,052.88, p � .15. Again,we found a trend for longer SSRT in incongruent than in nonwordneutral trials, F(1, 17) � 2.422, MSE � 1,575.795, p � .13, andin word neutral trials, F(1, 17) � 1.908, MSE � 1,548.178, p �.17 (see Table 3). There was no speed–accuracy trade-off, asindicated by the fact that none of the correlations between RT andnumber of errors were significant.

We combined SSRTs of the congruent, nonword neutral, andincongruent conditions from 30 participants in Experiments 1 and3, and we conducted a one-way ANOVA, similar to the previousanalysis. A main effect for congruency was found, F(2, 58) �3.814, MSE � 1,596.562, p � .05. Post hoc analysis revealed asignificant difference between congruent and incongruent trials,F(1, 29) � 5.912, MSE � 1,642.22, p � .05, partial �2 � .169, a

Figure 7. Mean reaction time (RT) for the congruency conditions of Stroop trials with no stop signal inExperiment 3. Error bars are one standard error of the mean.

Figure 8. Mean reaction time (RT) of erroneous responses to the stop signal for the congruency conditions inExperiment 3. Error bars are one standard error of the mean.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.


41

marginally significant difference between nonword neutral andincongruent trials, F(1, 29) � 3.711, MSE � 1,666.274, p � .06,partial �2 � .113, and no difference between congruent and neutraltrials, F � 1 (see Figure 10 and Table 3). These results correspondwith the finding of Verbruggen et al. (2004). It seems that aninformational conflicting stimulus harms the ability to stop. Theseresults confirm the interaction between inhibition of an initiatedresponse in the stop-signal paradigm and informational conflict.

General Discussion

Many studies have demonstrated that performance of the stop-signal task varies between different situations and different sub-jects (e.g., Logan, Schachar, & Tannock, 1997, found poorerperformance in impulsive subjects). This indicates that perfor-mance in a stop-signal task reflects the performance of a certaincontrol mechanism whose efficiency varies as a function of dif-ferent factors. Our design enabled us to zoom in on trials in which

the efficiency of this mechanism was compromised. Our objectivewas to investigate the Stroop performance in those trials.

A common Stroop effect was found in trials without a stopsignal. In contrast, Stroop performance in erroneous responses tostop-signal trials (i.e., when participants were unable to stop theirresponse) differed completely. When neutral trials were composedof a nonword series of letters (Experiments 1 and 3), a reversefacilitation occurred (RTs for congruent trials were significantlylonger than RTs for neutral trials). Correspondingly, there weremore erroneous responses in the neutral trials than in the congruentand incongruent trials. When the neutral was a real word (Exper-iments 2 and 3), the RTs and number of erroneous responses for allthree Stroop conditions were similar.

The reversed facilitation, found in erroneous response to stop-signal trials for the homogenous letter-string neutral condition(Experiments 1 and 3), suggests the existence of a task conflict. Italso indicates that some overlap exists between control mecha-

Figure 9. Proportion of erroneous responses to the stop signal for the congruency conditions of Experiment 3.Error bars are one standard error of the mean.

Figure 10. Stop-signal reaction time (RT) of all participants for the congruency conditions of Experiments1�3. Error bars are one standard error of the mean.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.


42

nisms that are responsible for inhibition in the stop-signal task andfor managing the task conflict. Once this mechanism failed, anerroneous response to the stop signal occurred and the task conflictbecame visible. These results, coupled with the fact that thereverse facilitation pattern was stable when RTs were fast and inthe slower half of the RT distribution, indicate that erroneousresponses to a stop signal are consistent even with a less efficientstop process (and not just a fast go process), at least in most of thetrials. The similar RTs for all three congruency conditions, foundin erroneous responses to stop-signal trials for real word neutrals(Experiments 2 and 3), strengthens this indication, because theword neutral condition caused a task conflict similar to that createdin the congruent and incongruent conditions.

The comparable performance in the congruent and incongruenttrials suggests differentiation between task conflict and informa-tional conflict. The informational conflict was normally resolvedin those trials and hence was not visible. In other words, the strongtask conflict in these trials, which was slowly solved, masked thenormally and quickly resolved informational conflict. This is con-trary to what is usually seen in Stroop incongruent trials, where astrong informational conflict masks a quickly resolved task con-flict.

The greater number of erroneous responses to the stop signal inthe nonword neutral condition (Experiments 1 and 3, see Figures3 and 9) also suggests the existence of a common control mech-anism for inhibition in the stop-signal task and for the task conflict.This common control results in a primed mechanism that makesstopping easier after this mechanism has been primed by the taskconflict (in all word conditions). One would suspect that, in lightof our limited cognitive resources, the probability of an erroneousresponse would be higher with real word trials. This prediction isbased on the fact that, in word trials, subjects had to engage twotasks that demanded control (i.e., task conflict and inhibition of analready initiated response in the stop-signal task) rather than onlyone in the nonword condition (i.e., inhibition). This descriptionassumes separate control mechanisms for task conflict and inhibi-tion of prepotent responses. Alternatively, a large overlap betweenthe two control mechanisms would predict that operation of one ofthem would result in better performance of the other, especiallyif the two processes were initiated sequentially, because the secondprocess would be primed by the first. This alternative provides adifferent prediction—the higher the probability that control will berecruited, the lower the probability that an erroneous response willoccur. More specifically, this approach predicts a higher number oferroneous responses to a stop signal in the nonword trials. This isbecause, in the nonword neutral trials, there is only one lateattempt to recruit this common control mechanism (i.e., at theappearance of the stop signal) while in nonneutral trials, or in theword neutral trials, there is another earlier attempt (i.e., the ap-pearance of the word stimulus) that primes stopping control. Ourresults support the notion of a primed mechanism. Furthermore,this current pattern of results (i.e., lower probability for erroneousresponse to a stop signal in real word trials) provides furtherevidence for the overlap and interaction between task control andinhibition of prepotent responses.

It seems appropriate to interpret our finding according to thetheoretical framework proposed by Monsell, Taylor, and Murphy(2001). They suggested that at least part of the interference that iscommonly observed in Stroop tasks is due to competition between

the relevant task and the irrelevant automatic task. The irrelevantautomatic task concept can be better understood according toMacLeod and MacDonald (2000), who argued that words raise anautomatic tendency to read a written word. Hence, in any Strooptrial, there is a conflict between two tasks—the relevant colornaming task and the irrelevant word reading task that was triggeredby the word stimulus. La Heij et al. (2010) also found a taskconflict between a relevant color naming task and an irrelevantautomatic object recognition task. Additionally, they suggestedthat this conflict is eliminated in adults by an efficient controlmechanism.

Goldfarb and Henik (2007) suggested that Stroop interference iscomposed of two types of conflict: the informational conflict (i.e.,between the contradicting information provided by the ink and theword) and the task conflict. The information conflict only occursin incongruent trials whereas the task conflict occurs every time awritten word is presented. Their suggestion was inspired by neu-roimaging findings that showed greater ACC activation in partic-ipants during both congruent and incongruent trials compared withactivation that appeared during neutral trials (e.g., Bench et al.,1993; Carter et al., 1995; Milham et al., 2002). The commonnotion that the ACC is recruited for conflict monitoring suggeststhat there is also a conflict in congruent trials, even though behav-ioral findings do not show it. Goldfarb and Henik suggested thatbecause of the efficient control mechanism, the task conflictslowed RTs in congruent trials by just a small amount, and RTswere still short because of the facilitation effect. In healthy adults,a more sophisticated manipulation and analysis is needed to revealthe task conflict. In the current study, we were able to make thetask conflict visible. In many ways, our findings are similar tothose of La Heij et al. (2010). While they demonstrated a taskconflict in children who have an immature control mechanism, wedemonstrated a task conflict in trials characterized by controlfailure or incomplete control operation. In both cases, the taskconflict was due to a deficient control mechanism.

Our findings also indicate a dissociation between two separatecontrol mechanisms; the first is responsible for the inhibition of aprepotent response, and is recruited when a task conflict or a stopsignal appears. The second is responsible for solving the informa-tional conflict, and it is recruited when facing stimuli that produceincongruent information. Our findings showed that failure in thetask control mechanism is not necessarily followed by failure inthe informational control mechanism.

Another important finding concerns the interaction between thestop-signal task and the Stroop informational conflict. When wecombined the results of all 30 participants in Experiments 1 and 3,we found a marginally significant difference in responses betweencongruent and incongruent trials and a significant difference be-tween neutral and incongruent trials. We found no differencebetween congruent and neutral trials. These results replicate pre-vious findings, which found that the ability to stop is influenced bydistracting information (e.g., Chambers et al., 2007; Kramer et al.,1994; Ridderinkhof et al., 1999; Verbruggen et al., 2004, 2006).Stopping seems more difficult when the go stimulus has an inter-fering aspect than when the go stimulus is neutral. Using a com-bined Stroop and stop-signal task, Verbruggen et al. (2004) foundlonger SSRTs for incongruent trials. This suggests an interactionbetween stopping and the informational conflict. The lack ofdifferences between the SSRT in neutral and congruent trials could

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.


43

have indicated no interaction between stopping and the task con-flict. As the results of the current study clearly indicated, this is nottrue. As in the common Stroop task, a closer look revealed the taskconflict and, in our case, its interaction with stopping.

Friedman and Miyake (2004) showed that stopping and Stroopperformance belong to the same individual differences factor.Using structural equation modeling, they suggested that prepotentresponse inhibition and resistance to distractor interference areclosely related. Our results are consistent with these findings, andthey reveal some overlap between Stroop interference and stop-signal inhibition.

Conclusion

The current study implies that two kinds of conflict appear in theStroop task: the informational conflict and the task conflict. Al-though the informational conflict stands out, the task conflict isusually not visible. This is due to a very efficient task controlmechanism and to facilitation in congruent conditions. To behav-iorally demonstrate the task conflict, a more sophisticated manip-ulation and analysis are required. This is usually possible onlywhen task control is damaged. Previous studies have found longerSSRTs for incongruent trials, indicating the existence of an inter-action between stopping and the informational conflict. Zoomingin on control failure trials revealed enhanced task conflict, but didnot indicate increased informational conflict. In other words, taskcontrol failure caused both erroneous responding to a stop signaland reverse facilitation. These findings suggest that at least someoverlap exists between the control mechanisms responsible fortask conflict and stop-signal inhibition; both can be conceptualizedas inhibition of a prepotent response. Furthermore, our resultsindicate that this task control mechanism is dissociable from theinformational conflict.

References

Allport, A., & Wylie, G. (2000). Task-switching, stimulus–response bind-ings and negative priming. In S. Monsell, & J. Driver (Eds.), Control ofcognitive processes: Attention and performance XVIII (pp. 35–70).Cambridge, MA: MIT Press.

Aron, A. R., Behrens, T. E., Smith, S., Frank, M. J., & Poldrack, R. A.(2007). Triangulating a cognitive control network using diffusion-weighted magnetic resonance imaging (MRI) and functional MRI. TheJournal of Neuroscience, 27, 3743–3752. doi:10.1523/JNEUROSCI.0519-07.2007

Aron, A. R., Fletcher, P. C., Bullmore, E. T., Sahakian, B. J., & Robbins,T. W. (2003). Stop-signal inhibition disrupted by damage to rightinferior frontal gyrus in humans. Nature Neuroscience, 6, 115–116.doi:10.1038/nn1003

Bench, C. J., Frith, C. D., Grasby, P. M., Friston, K. J., Paulesu, E.,Frackowiak, R. S., & Dolan, R. J. (1993). Investigations of the func-tional anatomy of attention using the Stroop test. Neuropsychologia, 31,907–922. doi:10.1016/0028-3932(93)90147-R

Botvinick, M., Nystrom, L. E., Fissell, K., Carter, C. S., & Cohen, J. D.(1999). Conflict monitoring versus selection-for-action in anterior cin-gulate cortex. Nature, 402, 179–181. doi:10.1038/46035

Botvinick, M. M., Braver, T. S., Barch, D. M., Carter, C. S., & Cohen, J. D.(2001). Conflict monitoring and cognitive control. Psychological Re-view, 108, 624–652. doi:10.1037/0033-295X.108.3.624

Braverman, A., & Meiran, N. (2010). Task conflict effect in task switching.Psychological Research, 74, 568–578. doi:10.1007/s00426-010-0279-2

Brown, J. W., & Braver, T. S. (2005). Learned predictions of errorlikelihood in the anterior cingulate cortex. Science, 307, 1118–1121.doi:10.1126/science.1105783

Carter, C. S., Botvinick, M. M., & Cohen, J. D. (1999). The contribution of theanterior cingulated cortex to executive processes in cognition. Reviews inNeuroscience, 10, 49–57. doi:10.1515/REVNEURO.1999.10.1.49

Carter, C. S., Braver, T. S., Barch, D. M., Botvinick, M. M., Noll, D., &Cohen, J. D. (1998). Anterior cingulate cortex, error detection, and theonline monitoring of performance. Science, 280, 747–749. doi:10.1126/science.280.5364.747

Carter, C. S., Mintun, M., & Cohen, J. D. (1995). Interference and facilitationeffects during selective attention: An H215O PET study of Stroop taskperformance. NeuroImage, 2, 264–272. doi:10.1006/nimg.1995.1034

Chambers, C. D., Bellgrove, M. A., Gould, I. C., English, T., Garavan, H.,McNaught, E., . . . Mattingley, J. B. (2007). Dissociable mechanisms ofcognitive control in prefrontal and premotor cortex. Journal of Neuro-physiology, 98, 3638–3647. doi:10.1152/jn.00685.2007

Chevrier, A. D., Noseworthy, M. D., & Schachar, R. (2007). Dissociationof response inhibition and performance monitoring in the stop signaltask using event-related fMRI. Human Brain Mapping, 28, 1347–1358.doi:10.1002/hbm.20355

Eimer, M. (1993). Effects of attention and stimulus probability on ERPs ina Go/Nogo task. Biological Psychology, 35, 123–138. doi:10.1016/0301-0511(93)90009-W

Friedman, N. P., & Miyake, A. (2004). The relations among inhibition andinterference control functions: A latent-variable analysis. Journal ofExperimental Psychology: General, 133, 101–135. doi:10.1037/0096-3445.133.1.101

Goldfarb, L., & Henik, A. (2007). Evidence for task conflict in the Stroopeffect. Journal of Experimental Psychology: Human Perception andPerformance, 33, 1170–1176. doi:10.1037/0096-1523.33.5.1170

Jodo, E., & Kayama, Y. (1992). Relation of a negative ERP component toresponse inhibition in a go/no-go task. Electroencephalography & ClinicalNeurophysiology, 82, 477–482. doi:10.1016/0013-4694(92)90054-L

Kalanthroff, E., & Henik, A. (2011). Switching enhances Stroop taskconflict. Manuscript submitted for publication.

Kornblum, S. (1994). The way irrelevant dimensions are processed de-pends on what they overlap with: The case of Stroop- and Simon-likestimuli. Psychological Research, 56, 130 –135. doi:10.1007/BF00419699

Kornblum, S., & Lee, J. W. (1995). Response compatibility with relevantand irrelevant stimulus dimensions that do and do not overlap with theresponse. Journal of Experimental Psychology: Human Perception andPerformance, 21, 855–875. doi:10.1037/0096-1523.21.4.855

Kramer, A. F., Humphrey, D. G., Larish, J. F., Logan, G. D., & Strayer,D. L. (1994). Aging and inhibition: Beyond a unitary view of inhibitoryprocessing in attention. Psychology and Aging, 9, 491–512. doi:10.1037/0882-7974.9.4.491

La Heij, W., & Boelens, H. (2011). Color–object interference: Further testsof an executive control account. Journal of Experimental Child Psychol-ogy, 108, 156–169. doi:10.1016/j.jecp.2010.08.007

La Heij, W., Boelens, H., & Kuipers, J. R. (2010). Object interference inchildren’s colour and position naming: Lexical interference or task-setcompetition? Language and Cognitive Processes, 25, 568–588. doi:10.1080/01690960903381174

Logan, G. D., & Cowan, W. B. (1984). On the ability to inhibit thought andaction: A theory of an act of control. Psychological Review, 91, 295–327. doi:10.1037/0033-295X.91.3.295

Logan, G. D., Cowan, W. B., & Davis, W. B. (1984). On the ability toinhibit simple and choice reaction time responses: A model and amethod. Journal of Experimental Psychology: Human Perception andPerformance, 10, 276–291. doi:10.1037/0096-1523.10.2.276

Logan, G. D. (1981). Attention, automaticity, and the ability to stop aspeeded choice response. In J. Long, & A. D. Baddeley (Eds.), Attention

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.


44

http://dx.doi.org/10.1523/JNEUROSCI.0519-07.2007

http://dx.doi.org/10.1523/JNEUROSCI.0519-07.2007

http://dx.doi.org/10.1038/nn1003

http://dx.doi.org/10.1016/0028-3932%2893%2990147-R

http://dx.doi.org/10.1038/46035

http://dx.doi.org/10.1037/0033-295X.108.3.624

http://dx.doi.org/10.1007/s00426-010-0279-2

http://dx.doi.org/10.1126/science.1105783

http://dx.doi.org/10.1515/REVNEURO.1999.10.1.49

http://dx.doi.org/10.1126/science.280.5364.747

http://dx.doi.org/10.1126/science.280.5364.747

http://dx.doi.org/10.1006/nimg.1995.1034

http://dx.doi.org/10.1152/jn.00685.2007

http://dx.doi.org/10.1002/hbm.20355

http://dx.doi.org/10.1016/0301-0511%2893%2990009-W

http://dx.doi.org/10.1016/0301-0511%2893%2990009-W

http://dx.doi.org/10.1037/0096-3445.133.1.101

http://dx.doi.org/10.1037/0096-3445.133.1.101

http://dx.doi.org/10.1037/0096-1523.33.5.1170

http://dx.doi.org/10.1016/0013-4694%2892%2990054-L

http://dx.doi.org/10.1007/BF00419699

http://dx.doi.org/10.1007/BF00419699

http://dx.doi.org/10.1037/0096-1523.21.4.855

http://dx.doi.org/10.1037/0882-7974.9.4.491

http://dx.doi.org/10.1037/0882-7974.9.4.491

http://dx.doi.org/10.1016/j.jecp.2010.08.007

http://dx.doi.org/10.1080/01690960903381174

http://dx.doi.org/10.1080/01690960903381174

http://dx.doi.org/10.1037/0033-295X.91.3.295

http://dx.doi.org/10.1037/0096-1523.10.2.276

and performance IX (pp. 205–222). Hillsdale, NJ: Lawrence ErlbaumAssociates.

Logan, G. D. (1994). On the ability to inhibit thought and action: A user’sguide to the stop signal paradigm. In D. Dagenbach & T. H. Carr (Eds.),Inhibitory processes in attention, memory and language (pp. 189–239).San Diego, CA: Academic Press.

Logan, G. D., Schachar, R. J., & Tannock, R. (1997). Impulsivity andinhibitory control. Psychological Science, 8, 60–64. doi:10.1111/j.1467-9280.1997.tb00545.x

MacLeod, C. M. (1991). Half a century of research on the Stroop effect: Anintegrative review. Psychological Bulletin, 109, 163–203. doi:10.1037/0033-2909.109.2.163

MacLeod, C. M., & MacDonald, P. A. (2000). Interdimensional interfer-ence in the Stroop effect: Uncovering the cognitive and neural anatomyof attention. Trends in Cognitive Sciences, 4, 383–391. doi:10.1016/S1364-6613(00)01530-8

Meiran, N. (1996). Reconfiguration of processing mode prior to taskperformance. Journal of Experimental Psychology: Learning, Memory,and Cognition, 22, 1423–1442. doi:10.1037/0278-7393.22.6.1423

Milham, M. P., Erickson, K. I., Banich, M. T., Kramer, A. F., Webb, A.,Wszalek, T., & Cohen, N. J. (2002). Attentional control in the agingbrain: Insights from an fMRI study of the Stroop task. Brain andCognition, 49, 277–296. doi:10.1006/brcg.2001.1501

Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A.,& Wager, T. D. (2000). The unity and diversity of executive functionsand their contributions to complex “frontal lobe” tasks: A latent variableanalysis. Cognitive Psychology, 41, 49 –100. doi:10.1006/cogp.1999.0734

Monsell, S. (2003). Task switching. Trends in Cognitive Sciences, 7,134–140. doi:10.1016/S1364-6613(03)00028-7

Monsell, S., Taylor, T. J., & Murphy, K. (2001). Naming the color of aword: Is it responses or task sets that compete? Memory & Cognition,29, 137–151. doi:10.3758/BF03195748

Nieuwenhuis, S., Yeung, N., Wildenberg, W., & Ridderinkhof, K. R.(2003). Electrophysiological correlates of anterior cingulate function ina go/no-go task: Effects of response conflict and trial type frequency.Cognitive, Affective, & Behavioral Neuroscience, 3, 17–26. doi:10.3758/CABN.3.1.17

Prevor, M. B., & Diamond, A. (2005). Color–object interference in youngchildren: A Stroop effect in children 31⁄2–61⁄2 years old. CognitiveDevelopment, 20, 256–278. doi:10.1016/j.cogdev.2005.04.001

Ridderinkhof, K. R., Band, G. P. H., & Logan, G. D. (1999). A study ofadaptive behavior: Effects of age and irrelevant information on theability to inhibit one’s actions. Acta Psychologica, 101, 315–337. doi:10.1016/S0001-6918(99)00010-4

Rogers, R. D., & Monsell, S. (1995). Costs of a predictable switch between

simple cognitive tasks. Journal of Experimental Psychology: General,124, 207–231. doi:10.1037/0096-3445.124.2.207

Rubia, K., Smith, A. B., Brammer, M. J., & Taylor, E. (2003). Rightinferior prefrontal cortex mediates response inhibition while mesialprefrontal cortex is responsible for error detection. NeuroImage, 20,351–358. doi:10.1016/S1053-8119(03)00275-1

Scheres, A., Oosterlaan, J., Swanson, J., Morein-Zamir, S., Meiran, N.,Schut, H., & Sergeant, J. A. (2003). The effect of methylphenidate onthree forms of response inhibition in boys with AD/HD. Journal ofAbnormal Child Psychology, 31, 105–120. doi:10.1023/A:1021729501230

Shallice, T. (1982). Specific impairments of planning. PhilosophicalTransactions of the Royal Society of London, B, 298, 199–209. doi:10.1098/rstb.1982.0082

Steinhauser, M., & Hubner, R. (2009). Distinguishing response conflictand task conflict in the Stroop task: Evidence from ex-Gaussian distri-bution analysis. Journal of Experimental Psychology: Human Percep-tion and Performance, 35, 1398–1412. doi:10.1037/a0016467

Stroop, J. R. (1935). Studies of interference in serial verbal reactions.Journal of Experimental Psychology, 18, 643– 662. doi:10.1037/h0054651

van Boxtel, G. J. M., van der Molen, M. W., & Jennings, J. R. (2005).Differential involvement of the anterior cingulate cortex in performancemonitoring during a stop-signal task. Journal of Psychophysiology, 19,1–10. doi:10.1027/0269-8803.19.1.1

van Veen, V., & Carter, C. S. (2006). Conflict and cognitive control in thebrain. Current Directions in Psychological Science, 15, 237–240. doi:10.1111/j.1467-8721.2006.00443.x

Verbruggen, F., Liefooghe, B., & Vandierendonck, A. (2004). The inter-action between stop signal inhibition and distractor interference in theflanker and the Stroop task. Acta Psychologica, 116, 21–37. doi:10.1016/j.actpsy.2003.12.011

Verbruggen, F., Liefooghe, B., & Vandierendonck, A. (2006). The effect ofinterference in the early processing stages on response inhibition in thestop signal task. The Quarterly Journal of Experimental Psychology, 59,190–203. doi:10.1080/17470210500151386

Verbruggen, F., & Logan, G. (2008). Response inhibition in the stop-signalparadigm. Trends in Cognitive Sciences, 12, 418–424. doi:10.1016/j.tics.2008.07.005

Waszak, F., Hommel, B., & Allport, D. A. (2003). Task switching andlong-term priming: Role of episodic stimulus-task bindings in task-shiftcosts. Cognitive Psychology, 46, 361– 413. doi:10.1016/S0010-0285(02)00520-0

Received May 12, 2011Revision received January 11, 2012

Accepted January 13, 2012 �

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.


45

http://dx.doi.org/10.1111/j.1467-9280.1997.tb00545.x

http://dx.doi.org/10.1111/j.1467-9280.1997.tb00545.x

http://dx.doi.org/10.1037/0033-2909.109.2.163

http://dx.doi.org/10.1037/0033-2909.109.2.163

http://dx.doi.org/10.1016/S1364-6613%2800%2901530-8

http://dx.doi.org/10.1016/S1364-6613%2800%2901530-8

http://dx.doi.org/10.1037/0278-7393.22.6.1423

http://dx.doi.org/10.1006/brcg.2001.1501

http://dx.doi.org/10.1006/cogp.1999.0734

http://dx.doi.org/10.1006/cogp.1999.0734

http://dx.doi.org/10.1016/S1364-6613%2803%2900028-7

http://dx.doi.org/10.3758/BF03195748

http://dx.doi.org/10.3758/CABN.3.1.17

http://dx.doi.org/10.3758/CABN.3.1.17

http://dx.doi.org/10.1016/j.cogdev.2005.04.001

http://dx.doi.org/10.1016/S0001-6918%2899%2900010-4

http://dx.doi.org/10.1016/S0001-6918%2899%2900010-4

http://dx.doi.org/10.1037/0096-3445.124.2.207

http://dx.doi.org/10.1016/S1053-8119%2803%2900275-1

http://dx.doi.org/10.1023/A:1021729501230

http://dx.doi.org/10.1023/A:1021729501230

http://dx.doi.org/10.1098/rstb.1982.0082

http://dx.doi.org/10.1098/rstb.1982.0082

http://dx.doi.org/10.1037/a0016467

http://dx.doi.org/10.1037/h0054651

http://dx.doi.org/10.1037/h0054651

http://dx.doi.org/10.1027/0269-8803.19.1.1

http://dx.doi.org/10.1111/j.1467-8721.2006.00443.x

http://dx.doi.org/10.1111/j.1467-8721.2006.00443.x

http://dx.doi.org/10.1016/j.actpsy.2003.12.011

http://dx.doi.org/10.1016/j.actpsy.2003.12.011

http://dx.doi.org/10.1080/17470210500151386

http://dx.doi.org/10.1016/j.tics.2008.07.005

http://dx.doi.org/10.1016/j.tics.2008.07.005

http://dx.doi.org/10.1016/S0010-0285%2802%2900520-0

http://dx.doi.org/10.1016/S0010-0285%2802%2900520-0

46

Chapter II

Experiment Group 4

Preparation Time Modulates Pro-Active Control and

Enhances Task Conflict in Task Switching

Eyal Kalanthroff & Avishai Henik

Psychological Research

2013

ORIGINAL ARTICLE

Preparation time modulates pro-active control and enhances taskconflict in task switching

Eyal Kalanthroff • Avishai Henik

Received: 3 December 2012 / Accepted: 16 May 2013 / Published online: 28 May 2013

� Springer-Verlag Berlin Heidelberg 2013

Abstract Performance in the Stroop task reflects two

conflicts—informational (between the incongruent word

and ink color) and task (between relevant color naming and

irrelevant word reading). Neuroimaging findings support

the existence of task conflict in congruent trials. A

behavioral indication for task conflict—Stroop reverse

facilitation—was found in previous studies under low task-

control conditions. Task switching also causes reduction in

task control because the task set frequently changes. We

hypothesized that it would be harder to efficiently manage

task conflicts in switching situations and, specifically, as

cue–target interval (CTI) decreases. This suggestion was

examined in two experiments using a combined Stroop

task-switching design. We found a large interference effect

and reverse facilitation that decreased with elongation of

CTI. Results imply that task switching reduces pro-active

task control and thereby enhances the informational and the

task conflicts. This calls for a revision of recent control

models to include task conflict.

Introduction

Cognitive control, essential for pursuing internal goals

(e.g., Banich, 2009; Miller & Cohen, 2001; Miyake et al.,

2000; Shallice & Norman, 1986), is commonly studied

using conflict situations such as the Stroop task (MacLeod,

1991; Stroop, 1935). Incongruent Stroop stimuli (e.g., RED

in green) create an information conflict between the color

of the ink and the meaning of the word. It has been recently

suggested that Stroop-congruent trials (e.g., GREEN in

green) present a task conflict between the automatic ten-

dency to read the word and the requirement to name the

color (Goldfarb & Henik, 2007; Kalanthroff, Goldfarb, &

Henik, 2013). Task conflict seems to appear when cogni-

tive control fails. The current work investigates cognitive

control of task failure in a task-switching situation.

Stroop and task conflict

In a common Stroop task, individuals are asked to identify

the ink color of a color word (i.e., the color the word is

written in) and avoid its meaning. The word color and word

meaning can be either congruent (e.g., RED written in red),

incongruent (e.g., GREEN written in red) or neutral (e.g.,

XXXX written in red). It is commonly assumed that

incongruent trials create a conflict and a control mechanism

is recruited to settle this conflict. Hence, the reaction time

(RT) for incongruent trials is commonly longer than RT for

neutral trials. This is known as the interference effect,

which is a very strong and stable effect (MacLeod, 1991).

In Stroop-congruent trials, both word color and word

meaning activate the same response; hence, RT is usually

faster than for neutral trials. This is known as the facili-

tation effect, which is a much smaller and less stable effect

than the interference effect (e.g., Dalrymple-Alford, &

Budayr, 1966). The low error rate in normal participants

and the increased error rate in participants with executive

and frontal deficits (Cohen & Servan-Schreiber, 1992)

demonstrate the role of executive control and of the pre-

frontal cortex in guiding task-relevant behavior and sup-

pressing automatic responses (Cohen, Dunbar, &

McClelland, 1990).

E. Kalanthroff (&) � A. Henik

Department of Psychology and Zlotowski Center

for Neuroscience, Ben-Gurion University of the Negev,

P.O.B. 653, 84105 Beersheba, Israel

e-mail: [email protected]

123

Psychological Research (2014) 78:276–288

DOI 10.1007/s00426-013-0495-7

47

Neuroimaging data show that the anterior cingulate

cortex (ACC), a brain area thought to monitor conflict

(Botvinick, Braver, Barch, Carter, & Cohen, 2001; Botvi-

nick, Nystrom, Fissell, Carter, & Cohen, 1999; Carter,

Botvinick, & Cohen, 1999; Carter et al., 1998), has greater

activation in incongruent Stroop trials compared to neutral

trials (e.g., Bench et al., 1993; Carter, Mintun, & Cohen,

1995; Milham et al., 2002, for older participants). Inter-

estingly, however, these studies also show that congruent

Stroop stimuli trigger greater ACC activations than neutral

stimuli do. Namely, congruent trials show facilitation in

RT and reverse facilitation in ACC activation. The con-

tradiction between the neuroimaging and the behavioral

data triggered a number of investigations, which led to a

distinction between two types of conflicts: information and

task (Goldfarb & Henik, 2007). In recent years, task con-

flict has drawn increased attention from researchers (e.g.,

Braverman & Meiran, 2010; Haggard, 2008; La Heij &

Boelens, 2011; La-Heij, Boelens, & Kuipers, 2010; Stein-

hauser & Hubner, 2009). Our earlier work (Goldfarb &

Henik, 2007; Kalanthroff, Goldfarb, & Henik, 2013; Kal-

anthroff, Goldfarb, Usher, Henik, 2012; Kalanthroff &

Henik, 2013) showed that the Stroop facilitation effect is

contingent upon task-conflict control. Aarts, Roelofs, and

van Turennouta (2009) found regions in the medial frontal

cortex whose activation is associated with information

conflict as well as with task conflict, and regions in the

lateral prefrontal cortex that are associated specifically with

task conflict.

Task conflict is due to the fact that stimuli acquire

associations with specific tasks. Many stimuli evoke per-

formance of strongly associated tasks (Allport & Wylie,

2000; Rogers & Monsell, 1995; Waszak, Hommel, &

Allport, 2003). For example, MacLeod and MacDonald

(2000) argued that words are strongly associated with the

task of reading and trigger an automatic tendency to read a

written word. This underscores the task conflict in the

Stroop task, namely, a conflict between two tasks—the

relevant color naming task and the irrelevant word reading

task, which was triggered by the stimulus (i.e., the word).

While the information conflict involves the content of the

stimulus and the response needed, and differs between

congruent and incongruent Stroop stimuli, task conflict

involves the task associated with the stimulus, and differs

between congruent and non-word-neutral Stroop stimuli.

Goldfarb and Henik (2007) argued that the reason for the

contradiction between behavioral evidence (i.e., RT facil-

itation) and neuroimaging findings (i.e., reverse facilita-

tion) is the existence of a very efficient active control

mechanism aimed at eliminating the task conflict. This

control mechanism can be considered to be a ‘task-

demand’ control (Botvinick et al., 2001) or a pro-active

control (Braver, 2012; De Pisapia & Braver, 2006). It is our

belief that task conflict may arise in particular in situations

in which the pro-active top–down control mechanism is

diminished, so that the stimulus may trigger a reactive

response of the associated task demands.

Braver’s (2012) dual mechanism of control suggests that

the pro-active control unit keeps constant amplification of

the relevant task units. Hence, pro-active control helps

maintain the task set (MacDonald, Cohen, Stenger, &

Carter, 2000). In contrast, in the absence of pro-active

control, the relevant task will not be strong enough and task

conflict will be high and will depend only on the short-term

reactive control to solve the task conflict. Accordingly, if

task-demand control is low, a very large interference effect

is predicted. Importantly, because task conflict is present in

congruent trials as well, it is reasonable to expect a reverse

facilitation effect (i.e., longer RTs for congruent compared

with neutral trials), which the dual mechanism of control

model does not account for.

One way to reduce task control is by relaxing it. Gold-

farb and Henik (2007) tried to put the task conflict guard

(i.e., pro-active control) to sleep. This was done by

increasing the proportion of non-word-neutral trials to

75 %, and by providing a valid cue for whether the

upcoming trial would be neutral or have a conflict (i.e.,

congruent or incongruent) in 50 % of the trials. They found

a significant reverse facilitation effect only in the non-cued

trials. Because the proportions and cueing were the causes

for the changes in RTs in congruent trials, these results

provide evidence for a task conflict between the word task

and the color task. Another way to reduce the efficiency of

task control is to overload it. To do this, Kalanthroff,

Goldfarb and Henik (2013) combined the Stroop task with

the stop-signal method (Logan, 1994; Logan & Cowan,

1984). They found a reverse facilitation in erroneous

responses to stop-signal trials. Namely, no stop trials

showed interference and facilitation, whereas (erroneous)

responding on stop trials showed interference and reverse

facilitation. A few other studies found indications for task

conflict in populations with low control ability (e.g., for

individual differences study see Kalanthroff & Henik,

2013; for children’s study see La Heij & Boelens, 2011; La

Heij et al., 2010).

Task switching

Another way to reduce task control and to make the task

conflict visible is to use the task-switching paradigm

(Meiran, 1996; Monsell, 2003, for reviews). In the common

procedure, participants are presented with two different

features of stimuli and asked to respond only to one on a

given trial. In the task-cuing method (e.g., Shaffer, 1965),

targets are preceded by a cue, at various intervals (i.e., cue–

Psychological Research (2014) 78:276–288 277

123

48

target interval, CTI; Meiran, 1996), indicating the required

task to carry out. Performance is usually slower on switch

trials (e.g., task in current trial is different from that in the

previous trial) than on repeat trials (e.g., task in current trial

is similar to that in the previous trial). Switch trials incur a

‘switching cost’ (Allport, Styles, & Hsieh, 1994) and are

modulated by CTI (Meiran, 1996). Different explanations

have been provided for the switching cost effect. Schneider

and Logan (2005) argued that in the task-cueing procedure,

faster RTs in repetition trials are caused by priming of the

(repeated) task cue encoding; De Jong (2000) and Rogers

and Monsell (1995) suggested that the cause for the

switching effects is the time taken by control processes to

reconfigure and to establish a changed task set (see also

Steinhauser, Maier, & Hubner, 2007); Allport et al. (1994)

and Allport and Wylie (2000) proposed that switching cost

is caused by carryover effects (interference or facilitation)

from the previous trials and is linked to pro-active inter-

ference; Meiran (1996, 2000) demonstrated that both car-

ryover from the previous trial and advance reconfiguration

and preparation for task shift influence switching cost.

Schuch and Koch (2003) argued that most of the switching

cost should be attributed to response selection and, spe-

cifically, to inhibition of the previous trial’s stimulus–

response association. In spite of the differences between

these various explanations, all agree that the need to switch

between tasks reduces task control or the ability to main-

tain a constant task set. Importantly, the present study does

not aim to determine which of these theories is correct,

although some implications might be suggested. The fre-

quent changes in task identity reduce task control. In the

present study, we use the task-switching paradigm to

reduce pro-active task control.

Task conflict, and the idea that neutral trials cause less

conflict, has been previously discussed in task-switching

theories (Braverman & Meiran, 2010; Gade & Koch, 2007;

Meiran & Daichman, 2005; Meiran & Kessler, 2008;

Rogers & Monsell, 1995; Rubin & Koch, 2006; Waszak

et al., 2003). A few studies found that in a task-switching

design, neutral stimuli lead to better performance than

congruent stimuli (also referred to as bivalency cost)—

evidence for task conflict. Using hierarchical global–local

stimuli, Steinhauser and Hubner (2007) found that both

incongruent and congruent stimuli cause larger conflict

than neutral stimuli, but only in a switching block (com-

pared to a single task block). A few studies combined task

switching with the Stroop task (e.g., Allport et al., 1994;

Gilbert & Shallice, 2002; Ruff, Woodward, Laurens, &

Liddle, 2001; Yeung & Monsell, 2003). However, no

congruent trials were included, making it impossible to

study the Stroop task conflict. Unlike these studies, Aarts

et al. (2009) conducted a Stroop-like task-switching study

with congruent trials and found a reverse facilitation effect

that (in one case) appeared only in switch trials. Steinha-

user and Hubner (2009) also conducted a task-switching

Stroop task and found evidence for task conflict (i.e.,

reverse facilitation). In both these studies, CTI was not a

part of the analysis, thus it is impossible to determine if

CTI can be used to recruit control that will eliminate task

conflict. Furthermore, both Aarts et al. (2009) and Stein-

hauser and Hubner (2009) used the same set of response

keys for both tasks, which makes it impossible to distin-

guish between response-congruent and response-incon-

gruent trials. We will further talk about this issue in

‘‘Discussion’’ of Experiment 1 and in ‘‘Discussion’’ of

Experiment 2.

The fact that reading is more automatic than color

naming underlies the Stroop asymmetry—interference

appears only in the color naming task (MacLeod, 1991;

Stroop, 1935). Most task-switching studies mentioned

above found a reverse Stroop effect (i.e., interference effect

in the word task) in switching blocks. This effect was

attributed to the frequent switching in task sets, which

causes a competition between tasks. Task control (pro-

active control) is reduced due to the fact that no one task

can get constant amplification (Aarts et al., 2009; Wood-

ward, Ruff, & Ngan, 2006). These ideas seem to fit task

conflict reasoning.

The present study

In the present study, we used a combined Stroop and task-

switching task. We used task switching to examine task

conflict in Stroop-congruent trials since previous studies

had shown increased task conflict (and reverse facilitation)

in task switching. The main goal of the present study was

to investigate the influence of different levels of control,

recruited in various preparation times (CTI), on task

conflict management. Participants were asked to respond

manually to the ink color or word meaning of words

and ignore the other dimension. A cue signaled the

required task (ink vs. word) to perform on the upcoming

stimulus. The Stroop stimulus was congruent, neutral, or

incongruent.

As suggested earlier, CTI should modulate difficulty in

dealing with the task conflict. That is, when the cue appears

simultaneously with the Stroop stimulus, task control will

be hardest and will improve as CTI elongates. Accordingly,

we predict a very large interference effect, which will

decrease with the elongation of CTI. We also predict a

reverse facilitation effect that will be larger at shorter CTIs,

possibly reverting to a regular facilitation at the longer CTI

condition. Moreover, we believe that frequently switching

between tasks will cause less available task control (pro-

active control) resources to deal with the Stroop task

278 Psychological Research (2014) 78:276–288

123

49

conflict and that this will apply to the reverse facilitation.

Since the task control (pro-active) unit is a ‘long-term’ or

‘block-wise’ control (Braver, 2012), it is not influenced by

the current trials; hence, we do not predict any differences

between switch (e.g., word task in trial n and color task in

trial n - 1) and repeat (e.g., word task in trial n and in trial

n - 1) trials. Trial-by-trial performance depends more on

reactive (short-term) control. Importantly, because we

made all predictions based on our belief that our design

would cause increased task conflict (or reduced ability to

deal with task conflict), our predictions account for both

the ink color response and word response.

Experiment 1

Method

Participants

Seventeen (11 females and 6 males) undergraduate stu-

dents of Ben-Gurion University of the Negev (Israel) par-

ticipated for partial fulfillment of course requirements and

received course credit for participating. All participants

had normal or corrected-to-normal vision, had no history of

attention deficit or dyslexia, were native speakers of

Hebrew, and were naive as to the purpose of the experi-

ment. The youngest participant was 21 years old and the

oldest was 27 years old (mean 23.3, standard deviation

1.49). One participant was excluded from the analysis due

to extremely slow RT (average RTs were more than 3

standard deviations from the mean in all conditions). Of the

remaining 16 participants, no one had more than 15 %

errors in any condition.

Stimuli

Participants were presented with a two-color, manual

Stroop task. Each stimulus consisted of one of the two

color words—

(Hebrew for green) or

(Hebrew for red). The ink color was either congruent (e.g.,

red written in red) or incongruent to the word meaning

(e.g., red written in green). Two different stimuli were used

as neutrals: for the word task, participants were presented

with each of the words written in white (the background

was always black); for the ink color task, we used a four-

letter string in Hebrew—

(meaningless letter string, parallel to XXXX in the English

version) written in red or green. The meaningless letter

string would not trigger automatic reading, and by that,

would cause no task conflict and no information conflict.

The task cue was ‘word’ or ‘color’ presented shortly

before or simultaneously with the Stroop stimulus. The task

cue was written in white letters (the same size as the Stroop

letters) and was presented above the Stroop stimulus. For

each task, there were 6 combinations of words and ink

colors: 2 congruent, 2 neutral and 2 incongruent. The

Stroop stimuli were presented at the center of a screen on a

black background and were .98 inches high and 2.36 inches

wide.

Procedure

Data collection and stimuli presentation were controlled by

a DELL OptiPlex 760 vPro computer with an Intel core 2

duo processor E8400 3 GHz. Stimuli were presented on a

DELL E198PF 1900 LCD monitor. A keyboard was placed

on a table between the participant and the monitor. Par-

ticipants were tested individually. They sat approximately

24 inches from the computer screen.

The experiment included three practice blocks (which

were not analyzed). The first two practice blocks were

dedicated to practicing the appropriate key-presses; one

block consisted of color cues and the other block consisted

of word cues. The cue and target stimulus appeared

simultaneously in the 12 trials in each of these blocks. The

third practice block included 36 trials and was similar to

the experimental block. Participants received accuracy

feedback for all the practice blocks, and in the third block

they also received RT feedback.

Each experimental trial started with a fixation (a white

plus sign at the center of a black screen), which was pre-

sented for a varied period of time (response–cue interval,

RCI). Following the fixation, a task cue appeared for a

varied period of time (CTI). There were three RCI–CTI

combination options (in ms)—1,800:0, 1,500:300, and

300:1,500. The Stroop stimulus appeared at the end of the

CTI. In all trials, the response–stimulus interval (RSI;

RCI ? CTI) remained the same (1,800 ms). The task cue

was presented up until the time of response so that it

overlapped with the Stroop stimulus. Each trial ended with

a 500-ms inter-trial interval.

There were two tasks, 6 combinations of words and ink

colors, and 3 RCI:CTI ratio options. This created 36


123

50

different trials that were presented randomly 11 times each,

giving a total of 396 trials (22 trials for each task, con-

gruency condition, CTI combination).

Participants were asked to place the index finger of their

right hand on the ‘‘7’’ key on the keyboard and their middle

finger of the same hand on the ‘‘9’’ key. Keys were marked

with colored stickers, which were counterbalanced between

participants. For both the word and the color tasks, par-

ticipants were asked to press the key that matched the

correct response. For example, for a given participant,

pressing 7 could indicate the color green in the color task

and the word green in the word task. Instructions empha-

sized quick and accurate responding. The Stroop stimulus

and the task cue remained in view until a key-press, or until

3,000 ms passed. RT was calculated from the appearance

of the Stroop stimulus to the response.

Results

Mean RTs of correct responses were calculated for each

participant for each condition of congruency (congruent,

neutral, and incongruent), CTI (0, 300, and 1,500 ms),

switching (switch vs. non-switch), and task (color vs.

word).

Repeat (735 ms) versus switch (780 ms) trial response

comparisons revealed significantly longer RTs in switch

trials, F(1, 15) = 16.539, MSE = 975.252, p \ .001, par-

tial eta squared (PES) = .524. This replicated the classic

switching cost effect (e.g., Allport et al., 1994).

The four-way interaction was not significant, F \ 1,

which indicates that the three-way interaction between

switching, congruency and CTI was similar for color and

word tasks.

To reveal the task conflict effect and our a priori

assumption, we conducted two separate three-way ANO-

VAs (analysis of variance) with repeated measures on RT

data for the color and the word tasks. Both analyses were

conducted with switching, congruency and CTI as within

subject factors.

Color task

A significant main effect for CTI was found, F(2,

30) = 205.772, MSE = 6,397.827, p \ .001, PES = .932.

This indicated shorter RTs for the long CTI (1,500 ms)

than for the short CTI (300 ms), and for the short CTI than

for the 0 CTI (0 ms). There was a significant main effect

for congruency, F(2, 30) = 70.210, MSE = 13,491.349,

p \ .001, PES = .824. These results indicated an inter-

ference effect (i.e., incongruent RT minus neutral RT;

246 ms), F(1, 15) = 73.998, MSE = 6,557.557, p [ .001,

PES = .831; and a non-significant reverse facilitation

effect (i.e., neutral RT minus congruent RT; -14), F \ 1.

The interaction between congruency and switching, and the

triple interaction among congruency, CTI, and switching

were not significant, F(2, 30) = 2.047, MSE = 6,831.689,

p = .15, and F \ 1, respectively (see Table 2), indicating

that switching did not affect the congruency effect and that

the interaction between CTI and congruency was similar in

switch and repeat trials. Most importantly, there was a

significant interaction between CTI and congruency, F(4,

60) = 11.171, MSE = 2,768.363, p \ .001, PES = .427.

As can be seen in Fig. 1, there was a reverse facilitation

that decreased as the CTI increased, and which eventually

changed to facilitation for the long CTI condition. Ana-

lyzing the simple effects revealed a significant reverse

facilitation for the 0 CTI condition (-43 ms), F(1,

15) = 4.865, MSE = 2,973.011, p \ .05, PES = .245; a

non-significant reverse facilitation for the short CTI

condition (-9 ms), F \ 1; and a non-significant facilita-

tion for the long CTI condition (15 ms), F \ 1. Comparing

the facilitation effects of the three CTI conditions

revealed a significant difference, F(2, 30) = 3.874,

MSE = 3,480.229, p \ .03, PES = .205. Post-hoc analy-

ses revealed a non-significant trend toward a smaller

facilitation in the 0 CTI condition compared with the 300

CTI condition, F(1, 15) = 2.25, MSE = 3,900.173,

p = .15, PES = .13; a non-significant smaller facilitation

in the short CTI condition compared with the long CTI

condition, F(1, 15) = 1.55, MSE = 3,155.693, p = .232;

and a significant smaller facilitation in the 0 CTI condition

compared with the long CTI condition, F(1, 15) = 7.911,

MSE = 3,384.821, p \ .02, PES = .345.

Contrary to the facilitation effect, the interference

became smaller with elongation of CTI. Analyzing the

simple effects revealed a significant and very large inter-

ference for the 0 CTI condition (334 ms), F(1, 15) =

156.963, MSE = 5,693.865, p \ .001, PES = .913; for the

short CTI condition (251 ms), F(1, 15) = 37.76,

MSE = 13,411.545, p \ .001, PES = .716; and for the

long CTI condition (162 ms), F(1, 15) = 47.588,

MSE = 4,396.064, p \ .001, PES = .76. As can be seen,

results indicated a larger interference at shorter CTIs.

Comparing the interference effects of the three CTI

conditions revealed a significant difference, F(2, 30) =

21.379, MSE = 5,572.638, p \ .001, PES = .588. Post-

hoc analyses revealed a significant larger interference in

the 0 CTI condition compared with the 300 CTI condition,

F(1, 15) = 7.687, MSE = 7,107.427, p \ .02, PES =

.339; a significant larger interference in the short CTI

condition compared with the long CTI condition, F(1,

15) = 9.682, MSE = 6,676.625, p \ .01, PES = .391;

and a significant larger interference in the 0 CTI condition


MSE = 2,933.862, p \ .001, PES = .844 (see Fig. 1;

Table 1).


123

51

Word task


30) = 157.341, MSE = 5,854.348, p \ .001, PES = .913,

indicating shorter RTs for the long CTI than for the short

CTI, and for the short CTI than for the 0 CTI. There was a

significant main effect for congruency, F(2, 30) = 56.533,

MSE = 17,520.591, p \ .001, PES = .79. This reversed

Stroop effect replicates previous findings (Allport & Wylie,

2000; Gilbert & Shallice, 2002; Yeung & Monsell, 2003).

Results indicated an interference effect (248 ms), F(1,

15) = 46.533, MSE = 10,552.761, p [ .001, PES = .756;

and a non-significant reverse facilitation effect (-22), F(1,

15) = 2.316, MSE = 1,683.5, p = .149 (see Fig. 1;

Table 1). The interaction between congruency and switch-

ing, and the triple interaction among congruency, CTI, and

switching were not significant, F \ 1 and F(4, 60) =

1.064, MSE = 6,912.005, p = .382, respectively (see

Table 2), indicating that switching did not affect congru-

ency effects and that the interaction between CTI and

congruency was similar in switch and repeat trials. Most

importantly, there was a marginally significant interaction

between CTI and congruency, F(4, 60) = 2.395, MSE =

3,358.818, p = .06, PES = .138. As can be seen in Fig. 1,

and similar to the color task, there was a reverse facilitation

that decreased as the CTI increased. Analyzing the simple

effects revealed a significant reverse facilitation for the 0

CTI condition (-74 ms), F(1, 15) = 6.266, MSE =

6,927.273, p \ .02, PES = .295; a non-significant reverse

facilitation for the short CTI condition (-27 ms), F(1,

15) = 2.619, MSE = 2,159.12, p = .126; and a non-sig-

nificant reverse facilitation for the long CTI condition

(-19 ms), F(1, 15) = 1.465, MSE = 2,237.844, p = .245.

Comparing the facilitation effects of the three CTI condi-

tions revealed a marginally significant difference, F(2,

30) = 3.046, MSE = 5,658.248, p = .07, PES = .173.

-100

-50

0

50

100

150

200

250

300

350

400

0 300 1500 0 300 1500

Color-Naming Task Word-Reading Task

Eff

ect (

ms)

FacilitationInterference

CTI (ms)

Fig. 1 Facilitation and

interference in the CTI

conditions for color and word

tasks of Experiment 1. Error

bars represent one standard

error from the mean

Table 1 Mean reaction times, [standard error of the mean] and (accuracy rate) for the different conditions of Experiment 1 and Experiment 2

Color task Word task

CTI Congruent Neutral Incongruent Congruent Neutral Incongruent

Experiment 1

0 830 [40] (.95) 788 [32] (.98) 1,122 [49] (.89) 898 [50] (.99) 824 [40] (.98) 1,101 [50] (.89)

300 624 [38] (.95) 614 [34] (.99) 866 [45] (.90) 655 [34] (.99) 629 [26] (.98) 920 [47] (.90)

1,500 523 [27] (.98) 538 [31] (.98) 700 [49] (.91) 615 [35] (.99) 595 [34] (.98) 825 [45] (.92)

Experiment 2

0 957 [41] (.93) 867 [35] (.95) 1,231 [46] (.89) 997 [46] (.93) 904 [32] (.95) 1,094 [50] (.93)

300 705 [43] (.95) 679 [34] (.96) 927 [48] (.92) 744 [36] (.96) 675 [32] (.98) 875 [45] (.94)

1,500 591 [35] (.99) 572 [27] (.99) 722 [54] (.94) 625 [41] (.98) 617 [37] (.99) 778 [47] (.96)

Reaction time and [standard error of the mean] measured in milliseconds


123

52

The interference effect became smaller with elongation of

CTI. Analyzing the simple effects revealed a significant and

very large interference for the 0 CTI condition (291 ms), F(1,

15) = 53.845, MSE = 11,436.618, p \ .001, PES = .782;

for the short CTI condition (277 ms), F(1, 15) = 43.954,

MSE = 15,475.982, p \ .001, PES = .746; and for the long

CTI condition (230 ms), F(1, 15) = 43.68, MSE =

9,713.871, p \ .001, PES = .744. As can be seen, results

indicated a larger interference at shorter CTIs. Comparing the

interference effects of the three CTI conditions revealed a

marginally significant difference, F(2, 30) = 9.944,

MSE = 7,797.653, p = .067, PES = .155 (see Fig. 1;

Table 1).

Discussion

As expected, task switching seems to have decreased the

participants’ ability to manage the task conflict, which

resulted in a reverse facilitation effect. This replicates

previous findings (Aarts et al., 2009; Steinhauser and

Hubner 2009). Most importantly, it seems that CTI was

used by participants to recruit control, and thus in the 0 CTI

condition, we found a reverse facilitation effect and a large

interference effect (Fig. 1). This corresponds with the

suggestion of Meiran (1996) about the nature of the CTI

that is used for control recruitment.

Though we do not have a single task condition to

compare, the large interference stands out in light of the

fact that we used only two colors and manual responding

[where there is a smaller effect for the manual modality

(White, 1969) and for small response-set size (Nielsen,

1974; see also MacLeod, 1991)]. The increased interfer-

ence effect in the task-switching block, compared to what

is commonly found in such designs in a single task block, is

in line with De Pisapia and Braver’s (2006) model, though

the reverse facilitation effect is not predicted by the model

and suggests that congruent Stroop stimuli also require

control. The fact that CTI modulated the interference and

facilitation effects for both tasks (though it was a non-

significant strong trend for the facilitation effect) reinforces

the suggestion that different levels of control modulate task

control (and task conflict effects).

Importantly, we found all of the effects for the word

reading task, including a reverse Stroop effect (interference

effect in the word task). This replicates previous findings

(e.g., Aarts et al., 2009; Woodward et al., 2006) and

strengthens our suggestion that frequent switching increa-

ses the competition between tasks, or in other words, the

task conflict.

In Experiment 1, participants used the same hand for

both tasks (word and color). This design enabled mini-

mizing response conflict. Since only two keys were used,

results could not be attributed to increased response-

selection conflict (see Cooper & Marı-Beffa, 2008; Schuch

& Koch, 2003). However, with this design, trials that were

informational-congruent (e.g., RED written in red) required

the same response for both tasks (e.g., ‘‘7’’ is the correct

response for RED written in red regardless of the task);

Table 2 Mean reaction times

and [standard error of the mean]

for the different conditions of

Experiment 1 and Experiment 2

for repeat and switch trials

separately

Reaction time and [standard

error of the mean] measured in

milliseconds

Color task Word task

CTI Congruent Neutral Incongruent Congruent Neutral Incongruent

Repeat trials

Experiment 1

0 806 [43] 763 [37] 1,079 [55] 874 [49] 799 [41] 1,100 [51]

300 596 [36] 586 [40] 782 [57] 652 [37] 603 [25] 877 [55]

1,500 512 [28] 512 [32] 698 [63] 594 [33] 553 [29] 767 [60]

Experiment 2

0 868 [39] 795 [36] 1,151 [58] 925 [52] 829 [37] 1,027 [55]

300 652 [39] 604 [35] 825 [57] 709 [42] 615 [32] 810 [51]

1,500 529 [24] 543 [28] 667 [55] 595 [39] 585 [34] 727 [47]

Switch trials

Experiment 1

0 851 [45] 829 [36] 1,164 [46] 925 [52] 846 [42] 1,103 [57]

300 644 [39] 627 [31] 947 [77] 665 [39] 654 [31] 935 [69]

1,500 536 [30] 561 [30] 706 [47] 639 [38] 647 [55] 847 [61]

Experiment 2

0 1,066 [41] 934 [40] 1,333 [42] 1,061 [44] 966 [33] 1,156 [53]

300 780 [53] 754 [36] 1,043 [46] 777 [34] 738 [36] 909 [55]

1,500 644 [49] 624 [29] 800 [56] 658 [45] 633 [42] 806 [55]


123

53

hence, they were also response-congruent. This is impor-

tant because in response-congruent trials participants do

not need the task cue to respond correctly. Not using the

task cue would presumably cause faster RTs in congruent

trials (rather than the slower RTs that were obtained in

Experiment 1). Namely, if participants neglected the task

cue in some of the trials, the reverse facilitation effect

obtained in Experiment 1 should have been even larger if

the task cue was always processed. To separate informa-

tional-congruent and response-congruent trials, Experiment

2 employed a separate set of keys for each dimension.

Experiment 2

In Experiment 2 we used separate sets of response keys for

each dimension. Although response conflict (between

hands) is potentially higher, this design will allow us to

investigate the task conflict when informational-congruent

trials are separated from response-congruent trials and will

eliminate the concerns of faster RTs in congruent trials due

to lack of task cue processing.

Method

Participants

Twenty (18 females and 2 males) undergraduate student of

Ben-Gurion University of the Negev (Israel), who did not

take part in Experiment 1, participated for partial fulfill-

ment of course requirements and received course credit for

participating. The same inclusion and exclusion criteria as

in Experiment 1 were used. Two participants (females)

were excluded due to extreme RTs and having more than

15 % errors. From the remaining 18 participants, the

youngest was 22 years old and the oldest was 28 years old

(mean 23.1, SD 1.5).

Stimuli

Stimuli were identical to those used in Experiment 1.

Procedure

In this experiment, we used a different set of response keys,

and different hands, for each task. For the word task, par-

ticipants were instructed to hit the ‘‘n’’ key on the keyboard

with the index finger of their left hand if the word meaning

was red, and hit the ‘‘v’’ key with the middle finger of the

same hand if the word meaning was green (the keys were

marked with stickers labeled with the Hebrew words for

RED and GREEN). For the color task, participants were

instructed to hit the ‘‘7’’ key on the keyboard with the index

finger of their right hand if the ink color was green, and hit

the ‘‘9’’ key with the middle finger of the same hand if the

ink color was red (the keys were marked with colored

stickers). Tasks and hands were counterbalanced across

participants. Except for this, the procedure in Experiment 2

was identical to that of Experiment 1.

Results

Similar to the previous experiment, mean RTs of correct

responses were calculated for each participant for each

condition of congruency, CTI, switching, and task. Repeat

(738 ms) versus switch (860 ms) trial response compari-

sons revealed significantly longer RTs in switch trials, F(1,

17) = 75.592, MSE = 1,779.973, p \ .001, PES = .816.

This replicated the classic switching cost effect (Allport

et al., 1994).

The four-way interaction was not significant, F \ 1,

which indicates that the three-way interaction among

switching, congruency and CTI was similar for color and

word tasks.

Similar to Experiment 1, RT data were subjected to two

separate three-way ANOVAs (for color and word tasks)

with congruency, switching and CTI as within subject

factors.

Color task


34) = 168.839, MSE = 12,448.317, p \ .001, PES = .792,


CTI, and for the short CTI than for the 0 CTI. There was a

significant main effect for congruency, F(2, 34) = 64.586,

MSE = 15,354.82, p \ .001, PES = .792. These results

indicated an interference effect, F(1, 17) = 42.042,

MSE = 9,221.574, p [ .001, PES = .712; and a significant

reverse facilitation effect, F(1, 17) = 8.544, MSE =

1,874.776, p \ .01. PES = .334. The interaction between

congruency and switching, and the triple interaction among

congruency, CTI, and switching were not significant, F(2,

34) = 2.352, MSE = 11,895.657, p = .12, and F \ 1,

respectively (see Table 2), indicating that switching did not

affect congruency effects and that the interaction between

CTI and congruency was similar in switch and repeat trials.

Most importantly, there was a significant interaction

between CTI and congruency, F(4, 68) = 10.118, MSE =

5,461.842, p \ .001, PES = .373. As can be seen in Fig. 2,

there was a reverse facilitation that decreased as the CTI

increased. Analyzing the simple effects revealed a signifi-

cant reverse facilitation for the 0 CTI condition (-91 ms),

F(1, 17) = 12.827, MSE = 5,751.72, p \ .01, PES = .43;

a non-significant reverse facilitation for the short CTI con-

dition (-26 ms), F(1, 17) = 1.91, MSE = 3,229.943,


123

54

p = .185; and a non-significant reverse facilitation for the

long CTI condition (-18 ms), F(1, 17) = 1.288, MSE =

2,449.025, p = .272. Comparing the facilitation effects of

the three CTI conditions revealed a significant difference,

F(2, 34) = 5.988, MSE = 5,755.902, p \ .01, PES = .26.

Post-hoc analyses revealed no significant difference between

the facilitation in the 0 CTI condition compared with the 300

CTI condition, F \ 1; a significant smaller facilitation in the

short CTI condition compared with the long CTI condition,

F(1, 17) = 6.392, MSE = 5,269.236, p \ .03, PES = .273;

and a significant smaller facilitation in the 0 CTI condition


MSE = 5,528.727, p \ .01, PES = .408.

Contrary to the facilitation effect, the interference

became smaller with elongation of CTI. Analyzing the

simple effects revealed a significant and very large inter-

ference for the 0 CTI condition (364 ms), F(1,

17) = 61.298, MSE = 19,550.032, p \ .001, PES = .783;

for the short CTI condition (247 ms), F(1, 17) = 58.136,

MSE = 9,477.822, p \ .001, PES = .774; and for the long


9,799.351, p \ .001, PES = .548. As can be seen, results

indicate a larger interference for shorter CTIs. Comparing

the interference effects of the three CTI conditions revealed

a significant difference, F(2, 34) = 15.288, MSE =

13,667.569, p \ .001, PES = .473. Post-hoc analyses

revealed a significant larger interference in the 0 CTI

condition compared with the 300 CTI condition, F(1,

17) = 12.77, MSE = 9,725.631, p \ .01, PES = .429; a

significant larger interference in the short CTI condition


MSE = 10,198.553, p \ .01, PES = .331; and a signifi-

cant larger interference in the 0 CTI condition compared

with the long CTI condition, F(1, 17) = 19.77,

MSE = 21,078.522, p \ .001, PES = .538 (see Fig. 2;

Table 1).

Word task


34) = 173.616, MSE = 8,744.043, p \ .001, PES = .911,


one, and for short CTI than for 0 CTI. There was a signif-

icant main effect for congruency, F(2, 34) = 68.765,

MSE = 6,908.779, p \ .001, PES = .802. This reversed

Stroop effect replicates the findings of Experiment 1 and

previous studies (Allport & Wylie, 2000; Gilbert & Shal-

lice, 2002; Yeung & Monsell, 2003). Results indicated an

interference effect, F(1, 17) = 105.561, MSE = 2,865.885,

p [ .001, PES = .861; and a significant reverse facilitation

effect, F(1, 17) = 12.496, MSE = 2,115.26, p \ .01,

PES = .424. Similar to the color task analysis, the inter-

action between congruency and switching, and the triple

interaction among congruency, CTI, and switching were not

significant, F \ 1, for both analyses (see Table 2), indi-

cating that switching did not affect congruency effects and

that the interaction between CTI and congruency was sim-

ilar in switch and repeat trials. Most importantly, there was

a significant interaction between CTI and congruency, F(4,

68) = 3.1, MSE = 3,016.905, p \ .03, PES = .154. As

can be seen in Fig. 2, and similar to the color task, there was

a reverse facilitation that decreased as the CTI increased.

Analyzing the simple effects revealed a significant reverse

facilitation for the 0 CTI condition (-93 ms), F(1,

17) = 16.053, MSE = 4,841.301, p \ .001, PES = .486; a

significant reverse facilitation for the short CTI condition

(-69 ms), F(1, 17) = 14.692, MSE = 2,943.017, p \.001; and a non-significant reverse facilitation for the long

-100

-50

0

50

100

150

200

250

300

350

400

0 300 1500 0 300 1500

Color-Naming Task Word-Reading Task

Eff

ect

(ms)

FacilitationInterference

CTI (ms)

Fig. 2 Facilitation and

interference in the CTI

conditions for color and word

tasks of Experiment 2. Error

bars represent one standard

error from the mean


123

55

CTI condition (-8 ms), F \ 1. Comparing the facilitation

effects of the three CTI conditions revealed a significant

difference, F(2, 34) = 5.988, MSE = 5,755.902, p \ .01,

PES = .26.

Analyzing the simple effects for the interference effect

revealed a significant and very large interference (yet

smaller than the similar effect for the color task) for the 0


6,117.782, p \ .001, PES = .756; an interference effect

for the short CTI condition that unexpectedly did not

decrease (199 ms), F(1, 17) = 69.781, MSE = 5,138.706,

p \ .001, PES = .804; and an interference for the long

CTI condition that was shorter (161 ms), F(1, 17) =

50.756, MSE = 4,576.207, p \ .001, PES = .749. Over-

all, comparing the interference effects of the three CTI

conditions revealed a smaller interference for the longest

CTI, though the difference between the three interference

effects failed to reach significance, F(2, 34) = 1.045,

MSE = 7,035.219, p = .363 (see Fig. 2; Table 1).

Discussion

The results of Experiment 2 replicated the findings of

Experiment 1. Again we found that task switching

decreased the participants’ ability to manage the informa-

tional and task conflicts. Most importantly, we found that

when no preparation time was available (i.e., 0 CTI con-

dition), the reverse facilitation effect was largest. The

interference effect was very large and this was true espe-

cially for the short CTI condition. Again, we found a

reverse Stroop effect. The results of Experiment 2 suggest

that task conflict occurs when information-congruent and

response-congruent trials are completely separated and that

the task conflict effect is contingent upon different levels of

task control (i.e., CTI). The fact that the current results

replicated those of Experiment 1 means that the previous

results were not due to the use of the same response keys

for the two tasks. The replication in Experiment 2 suggests

that results obtained in Experiment 1 should be attributed

to task conflict management.

General discussion

We used a combined two-color manual Stroop task in a

cued task-switching design with variable CTIs to investi-

gate task conflict with different levels of task-demand

control. In Experiment 1, we asked participants to respond

with one set of keys and in Experiment 2 we used a dif-

ferent set of keys for each task to investigate task conflict in

informational-congruent trials that were not response-con-

gruent. Importantly, Experiment 2 allowed us to ensure that

participants processed the task cue in the congruent trials of

Experiment 1. We used a meaningless letter string (XXXX)

as a neutral for the color task and a white ink color as a

neutral for the word task. Table 1 reveals that overall,

neutral trials were a bit slower for the word task—possibly

due to the fact that white causes a small task conflict

(though much smaller than the conflict caused by colors in

the response set).

We found all of the effects for the word reading task,

including a reverse Stroop effect, that is, an interference

effect in the word task. This effect replicates earlier find-

ings (e.g., Aarts et al., 2009; Ruff et al., 2001) that have

previously been attributed to competition between tasks

(e.g., Aarts et al., 2009; Woodward et al., 2006). This

strengthens our assumption that task conflict is high (or it is

difficult to deal with) in a task-switching design. Though

some studies found a reverse Stroop effect in a manual

Stroop task (e.g., Blais & Besner, 2006), it is important to

mention that with translated response (like color stickers

that were used in the present study) it is not clear how

crucial response mode is because tasks investigated are

rather different from the basic Stroop task (e.g., Ikeda,

Hirata, Okuzumi, & Kokubun, 2010) and the stimuli are

different from the basic Stroop stimuli as well (MacLeod,

1991). Moreover, it is important to stress the fact that all

effects from the color task were replicated in the word task,

which strengthens our suggestion that the results should be

attributed to competition between tasks caused by the fre-

quent task switching.

In both experiments, we found a very large interference

effect (considering we used a two-color manual Stroop

task) that was reduced with elongation of preparation time

(CTI). This corresponds with Braver’s (2012) and De

Pisapia and Braver’s (2006) model of cognitive control.

This model predicts that if a task is not constant (and

possibly pro-active control is either low or occupied else-

where), a high interference effect occurs, and that larger

preparation time could reduce the time needed to solve the

conflict. These results also correspond with the suggestion

that CTI is used to recruit control (Meiran, 1996).

Most importantly, we found that responding to congru-

ent trials is contingent upon control levels. In both exper-

iments, we found a reverse facilitation effect (replication of

Aarts et al., 2009; Steinhauser and Hubner 2009) that was

modulated by CTI. In the 0 CTI condition (i.e., less time to

recruit control), we found a significant and large reverse

facilitation effect in both experiments and for both color

and word tasks. The reverse facilitation reduced with

increase in CTI and sometimes even changed to a regular

facilitation, in both tasks in both experiments. CTI lengths

were taken from the classic paper of Meiran (1996), which

found smaller switching cost with the elongation of CTI. In

both his and our experiments, differences in control levels

between CTIs of 0 and 300 ms were found, though the


123

56

larger change occurred only after 1,500 ms—it seems that

the control recruitment processes start early though they

are not completed until much later. Apart from Meiran’s

study, there are other indications in the literature for con-

trol recruitment in very short periods of time. One classic

example is Neely’s (1977) study on semantic priming that

shows voluntary priming (and inhibition of automatic

priming) is possible after 250 ms but before 750 ms.

There was no difference in facilitation and interference

effects between switch and repeat trials. Nevertheless, it is

important to mention that though the interaction between

congruency and switch was not significant for either task in

neither of the experiments, there was a trend in some of the

analyses. Importantly, we do not rule out the possibility

that task conflict is stronger and task control is lower in

switching trials. Though, the fact that this interaction was

not significant is in line with the assumption that the

reverse facilitation effect is caused by processes that are

due to the frequent changes in task sets (that force keeping

both tasks in mind at the same time), which are long-term

processes and not trial-by-trial processes. In other words,

the fact that switching did not affect the facilitation effect

in the same trial (i.e., switch trial) indicates that our results

can be attributed to changes in task (pro-active) control,

which is a long-term control (Braver, 2012). Further

research is needed to investigate the influences of switch

and repeat conditions in single trials.

Monsell, Taylor and Murphy (2001) suggested that at

least part of the interference that is commonly observed in

Stroop tasks was due to competition between the relevant

task and the irrelevant automatic task. The irrelevant

automatic task concept can be better understood according

to MacLeod and MacDonald (2000), who argued that

words raise an automatic tendency to read them. Hence, in

any Stroop trial there is a conflict between two tasks—the

relevant color task and the irrelevant word task, which was

triggered by the word stimulus. Usually this conflict is not

visible due to an efficient task-control mechanism (Gold-

farb & Henik, 2007; Kalanthroff, Goldfarb, Usher et al.,

2012; Kalanthroff & Henik, 2013; La Heij et al., 2010).

Task switching enhances the task conflict and makes it

visible. It seems that preparation time was used for efficient

recruitment of cognitive control that could compensate for

the enhanced task conflict caused by frequent changes in

the task.

Several researchers have emphasized the need for task

demands or task-set maintenance control (e.g., Braver,

2012; Cohen et al., 1990; De Pisapia & Braver, 2006;

MacDonald et al., 2000; Monsell, 2003; Rogers & Monsell,

1995). Specifically, these researchers focused on difficul-

ties in Stroop incongruent trials when task control (pro-

active control) was overloaded. Our results indicate that an

overloaded task control will not only cause incongruent

trials to be more difficult (and slow) but will also cause

congruent trials to be more difficult due to task conflict.

This calls for revaluation of the current models so that they

will deal with the task conflict that exists whenever there is

a written word (i.e., Stroop-congruent and word-neutral

trials).

It is worthwhile mentioning that although in the present

study manipulations that affected the task conflict (mainly

presented by the facilitation effect) also affected the

informational conflict (presented by the interference

effect), there is some evidence indicating that pro-active

control of the informational conflict and pro-active control

of the task conflict are distinguishable at least to some

extent. In Kalanthroff, Goldfarb, and Henik’s (2013) study

that was mentioned earlier, we found that in erroneous

responding when a stop signal appeared, facilitation was

reversed although the interference effect was not signifi-

cantly larger than in no stop-signal trials. This indicates

that a specific manipulation can affect mostly the task

conflict. Similarly, Bugg, McDaniel, Scullin, and Braver

(2011) manipulated congruent, neutral, or incongruent

proportions in separate Stoop blocks using non-color words

as neutrals. These researchers found a decreased interfer-

ence effect (increased pro-active control) in the high

incongruent and high neutral-word proportion blocks,

although the facilitation effect was similar in all conditions

(looking at their descriptive statistics reveals even a bit

smaller facilitation in the mostly incongruent block). This

is an example for a way to affect informational conflict and

not task conflict since all trials contained real words and

thus generated task conflict.

We employed different neutrals for each task (i.e.,

XXXX for the color task and a white color for the word

task). This might mean that under specific circumstances,

participants ignored the task cue and responded according

to the target stimulus only. This strategy could be applied

in the 0 CTI condition. Although if this was true, RTs for

neutral trials in the 0 CTI condition should have been faster

than RTs for neutral trials in the longer CTI conditions.

This was not the case. RTs for neutral trials in the 0 CTI

condition are slower than RTs for neutral trials in the two

other CTI conditions.

The present study did not aim to explain switching

effects. However, our results contribute to the ongoing

debate on the subject. In particular, our results suggest that

CTI is important for the operation of executive control (De

Jong, 2000; Meiran, 1996; Rogers & Monsell, 1995).

Furthermore, our previous study (Kalanthroff, Goldfarb, &

Henik, 2013; Kalanthroff & Henik, 2013) indicated some

overlap between task control and inhibition of a prepotent

response. Accordingly, the current results suggest that

inhibitory processes are greatly reduced in a switching

situation. This can potentially support the suggestion that


123

57

switching cost is due to inhibition of the previous trial task.

Most importantly, the present study suggests that any task-

switching experiment should take into consideration the

influences on pro-active control and, specifically, the pos-

sibility for task conflict.

Conclusions

To conclude, our findings indicate that task switching

enhances the informational conflict (evidenced as larger

interference) and the task conflict (evidenced as reverse

facilitation). The most important finding of the present

study is that preparation time is needed to recruit task

control and can, to some extent, alleviate the reduction in

task control caused by frequent task switching. The current

results call for a revision of recent models of control (e.g.,

Braver, 2012; De Pisapia & Braver, 2006) so that they

include the task conflict that appears in congruent trials.

Acknowledgments We thank Prof. Nachshon Meiran for his help-

ful insights, Dr. Julie Bugg and another anonymous reviewer for their

helpful comments, and Ms. Desiree Meloul for her useful input on

this article.

References

Aarts, E., Roelofs, A., & van Turennouta, M. (2009). Attentional

control of task and response in lateral and medial frontal cortex:

Brain activity and reaction time distributions. Neuropsychologia,

47, 2089–2099.

Allport, D. A., Styles, E. A., & Hsieh, S. (1994). Shifting intentional

set: Exploring the dynamic control of tasks. In C. Umilta & M.

Moscovitch (Eds.), Attention and performance XV: Conscious

and nonconscious information processing (pp. 421–452). Hills-

dale, NJ: Erlbaum.

Allport, A., & Wylie, G. (2000). Task switching, stimulus–response

bindings and negative priming. In S. Monsell & J. Driver (Eds.),

Control of cognitive processes: Attention and performance XVIII

(pp. 35–70). Cambridge, MA: MIT Press.

Banich, M. T. (2009). Executive function: The search for an

integrated account. Current Directions in Psychological Science,

18, 89–94.

Bench, C. J., Frith, C. D., Grasby, P. M., Friston, K. J., Paulesu, E.,

Frackowiak, R. S., et al. (1993). Investigations of the functional

anatomy of attention using the Stroop test. Neuropsychologia,

32, 907–922.

Blais, C., & Besner, D. (2006). Reverse Stroop effects with

untranslated responses. Journal of Experimental Psychology:

Human Perception and Performance, 32, 1345–1353.

Botvinick, M. M., Braver, T. S., Barch, D. M., Carter, C. S., & Cohen,

J. D. (2001). Conflict monitoring and cognitive control.

Psychological Review, 108, 624–652.

Botvinick, M. M., Nystrom, L. E., Fissell, K., Carter, C. S., & Cohen,

J. D. (1999). Conflict monitoring versus selection-for-action in

anterior cingulate cortex. Nature, 402, 179–181.

Braver, T. S. (2012). The variable nature of cognitive control: A dual

mechanisms framework. Trends in Cognitive Sciences, 16,

106–113.

Braverman, A., & Meiran, N. (2010). Task conflict effect in task

switching. Psychological Research, 74, 568–567.

Bugg, J. M., McDaniel, M. A., Scullin, M. K., & Braver, T. S. (2011).

Revealing list-level control in the Stroop task by uncovering its

benefits and a cost. Journal of Experimental Psychology: Human

Perception and Performance, 37, 1595–1606.

Carter, C. S., Botvinick, M. M., & Cohen, J. D. (1999). The

contribution of the anterior cingulated cortex to executive

processes in cognition. Reviews in Neuroscience, 10, 49–57.

Carter, C. S., Braver, T. S., Barch, D. M., Botvinick, M. M., Noll, D.,

& Cohen, J. D. (1998). Anterior cingulate cortex, error detection,

and the online monitoring of performance. Science, 280,

747–749.

Carter, C. S., Mintun, M., & Cohen, J. D. (1995). Interference and

facilitation effects during selective attention: An H215O PET

study of Stroop task performance. NeuroImage, 2, 264–272.

Cohen, J. D., Dunbar, K., & McClelland, J. L. (1990). On the control

of automatic processes: A parallel distributed processing account

of the Stroop effect. Psychological Review, 97, 332–361.

Cohen, J. D., & Servan-Schreiber, D. (1992). Context, cortex and

dopamine: A connectionist approach to behavior and biology in

schizophrenia. Psychological Review, 99, 45–77.

Cooper, S., & Marı-Beffa, P. (2008). The role of response repetition

in task switching. Journal of Experimental Psychology: Human


Dalrymple-Alford, E. C., & Budayr, B. (1966). Examination of some

aspects of the Stroop color-word test. Perceptual and Motor

Skills, 23, 1211–1214.

De Jong, R. (2000). An intention-activation account of residual switch

costs. In S. Monsell & J. S. Driver (Eds.), Control of cognitive

processes: Attention and performance XVIII (pp. 357–376).

Cambridge, MA: MIT Press.

De Pisapia, N., & Braver, T. S. (2006). A model of dual control

mechanisms through anterior cingulate and prefrontal cortex

interactions. Neurocomputing, 69, 1322–1326.

Gade, M., & Koch, I. (2007). Cue–task associations in task switching.

The Quarterly Journal of Experimental Psychology, 60,

762–769.

Gilbert, S. J., & Shallice, T. (2002). Task switching: A PDP model.

Cognitive Psychology, 44, 297–337.

Goldfarb, L., & Henik, A. (2007). Evidence for task conflict in the

Stroop effect. Journal of Experimental Psychology: Human


Haggard, P. (2008). Human volition: Towards a neuroscience of will.

Nature Reviews Neuroscience, 9, 934–946.

Ikeda, Y., Hirata, S., Okuzumi, H., & Kokubun, M. (2010). Features

of Stroop and reverse-Stroop interference: Analysis by response

modality and evaluation. Perceptual and Motor Skills, 110,

654–660.

Kalanthroff, E., Goldfarb, L., & Henik, A. (2013). Evidence for

interaction between the Stop-Signal and the Stroop task conflict.

Journal of Experimental Psychology: Human Perception and

Performance, 39, 579–592.

Kalanthroff, E., Goldfarb, L., Usher, M., & Henik, A. (2012). Stop

interfering: Stroop task conflict independence from informa-

tional conflict and interference. Quarterly Journal of Experi-

mental Psychology. Advance online publication. doi:10.1080/

17470218.2012.741606.

Kalanthroff, E., & Henik, A. (2013). Individual but not fragile:

Individual differences in task control predict Stroop facilitation.

Consciousness and Cognition, 22, 413–419.

La Heij, W., & Boelens, H. (2011). Color–object interference: Further

tests of an executive control account. Journal of Experimental

Child Psychology, 108, 156–169.

La Heij, W., Boelens, H., & Kuipers, J. R. (2010). Object interference

in children’s colour and position naming: Lexical interference or


123

58

http://dx.doi.org/10.1080/17470218.2012.741606

http://dx.doi.org/10.1080/17470218.2012.741606

task-set competition? Language and Cognitive Processes, 25,

568–588.

Logan, G. D. (1994). On the ability to inhibit thought and action: A

user’s guide to the stop signal paradigm. In D. Dagenbach & T.

H. Carr (Eds.), Inhibitory processes in attention, memory and

language (pp. 189–239). San Diego, CA: Academic Press.

Logan, G. D., & Cowan, W. B. (1984). On the ability to inhibit

thought and action: A theory of an act of control. Psychological

Review, 91, 295–327.

MacDonald, A. W., Cohen, J. D., Stenger, V. A., & Carter, C. S.

(2000). Dissociating the role of dorsolateral prefrontal cortex and

anterior cingulate cortex in cognitive control. Science, 288,

1835–1838.

MacLeod, C. M. (1991). Half a century of research on the Stroop

effect: An integrative review. Psychological Bulletin, 109,

163–203.

MacLeod, C. M., & MacDonald, P. A. (2000). Interdimensional

interference in the Stroop effect: Uncovering the cognitive and

neural anatomy of attention. Trends in Cognitive Sciences, 10,

383–391.

Meiran, N. (1996). Reconfiguration of processing mode prior to task

performance. Journal of Experimental Psychology. Learning,

Memory, and Cognition, 22, 1423–1442.

Meiran, N. (2000). Modeling cognitive control in task-switching.

Psychological Research, 63, 234–249.

Meiran, N., & Daichman, A. (2005). Advance task preparation

reduces task error rate in the cuing task-switching paradigm.

Memory & Cognition, 33, 1272–1288.

Meiran, N., & Kessler, Y. (2008). The task rule congruency effect in

task switching reflects activated long-term memory. Journal of

Experimental Psychology: Human Perception and Performance,

34, 137–157.

Milham, M. P., Erickson, K. I., Banich, M. T., Kramer, A. F., Webb,

A., Wszalek, T., et al. (2002). Attentional control in the aging

brain: Insights from an fMRI study of the Stroop task. Brain and

Cognition, 49, 277–296.

Miller, E. K., & Cohen, J. D. (2001). An integrative theory of

prefrontal cortex function. Annual Review of Neuroscience, 24,

167–202.

Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H.,

Howerter, A., & Wager, T. D. (2000). The unity and diversity of

executive functions and their contributions to complex ‘‘frontal

lobe’’ tasks: A latent variable analysis. Cognitive Psychology,

41, 49–100.

Monsell, S. (2003). Task switching. Trends in Cognitive Sciences, 7,

134–140.

Monsell, S., Taylor, T. J., & Murphy, K. (2001). Naming the color of

a word: Is it responses or task sets that compete? Memory and

Cognition, 29, 137–151.

Neely, J. H. (1977). Semantic priming and retrieval from lexical

memory: Roles of inhibitionless spreading activation and

limited-capacity attention. Journal of Experimental Psychology:

General, 106, 226–254.

Nielsen, G. D. (1974). The locus and mechanism of the Stroop color

word effect. Dissertation Abstracts International, 35, 5672.

Rogers, R. D., & Monsell, S. (1995). Costs of a predictable switch

between simple cognitive tasks. Journal of Experimental Psy-

chology: General, 124, 207–231.

Rubin, O., & Koch, I. (2006). Exogenous influences on task set

activation in task switching. The Quarterly Journal of Experi-

mental Psychology, 59, 1033–1046.

Ruff, C. C., Woodward, T. S., Laurens, K. R., & Liddle, P. F. (2001).

The role of the anterior cingulate cortex in conflict processing:

Evidence from reverse Stroop interference. NeuroImage, 14,

1150–1158.

Schneider, D. W., & Logan, D. G. (2005). Modeling task switching

without switching tasks: A short-term priming account of

explicitly cued performance. Journal of Experimental Psychol-

ogy: General, 134, 343–367.

Schuch, S., & Koch, I. (2003). The role of response selection for

inhibition of task sets in task shifting. Journal of Experimental

Psychology: Human Perception and Performance, 29, 92–105.

Shaffer, L. H. (1965). Choice reaction with variable S-R mapping.

Journal of Experimental Psychology, 70, 284–288.

Shallice, T., & Norman, D. (1986). Attention to action: Willed and

automatic control of behavior. In R. Davidson, G. Schwartz, &

D. Shapiro (Eds.), Consciousness and self-regulation: Advances

in research and theory (Vol. 4, pp. 2–18). New York: Plenum

Press.

Steinhauser, M., & Hubner, R. (2007). Automatic activation of task-

related representations in task shifting. Memory and Cognition,

35, 138–155.

Steinhauser, M., & Hubner, R. (2009). Distinguishing response

conflict and task conflict in the Stroop task: Evidence from ex-

Gaussian distribution analysis. Journal of Experimental Psy-

chology: Human Perception and Performance, 35, 1398–1412.

Steinhauser, M., Maier, M., & Hubner, R. (2007). Cognitive control

under stress: How stress affects strategies of task-set reconfig-

uration. Psychological Science, 18, 540–545.

Stroop, J. R. (1935). Studies of interference in serial verbal reactions.

Journal of Experimental Psychology, 18, 643–662.

Waszak, F., Hommel, B., & Allport, D. A. (2003). Task switching and

long-term priming: Role of episodic stimulus-task bindings in

task-shift costs. Cognitive Psychology, 46, 361–413.

White, B. W. (1969). Interference in identifying attributes and

attribute names. Perception and Psychophysics, 6, 166–168.

Woodward, T. S., Ruff, C. C., & Ngan, E. T. (2006). Short- and long-

term changes in anterior cingulate activation during resolution of

task-set competition. Brain Research, 1068, 161–169.

Yeung, N., & Monsell, S. (2003). Switching between tasks of unequal

familiarity: The role of stimulus-attribute and response-set

selection. Journal of Experimental Psychology: Human Percep-

tion and Performance, 29, 455–469.


123

59

60

Chapter II

Theoretical model

What Should I (Not) Do? Control Over Irrelevant Tasks

in Obsessive-Compulsive Disorder Patients

Eyal Kalanthroff, Gideon E. Anholt, Rotem Keren, & Avishai Henik

Clinical Neuropsychiatry

2013

Clinical Neuropsychiatry (2013) 10, 3, Suppl. 1, 61-64

WHAT SHOULD I (NOT) DO? CONTROL OVER IRRELEVANT TASKS IN OBSESSIVE-COMPULSIVE DISORDER PATIENTS

Eyal Kalanthroff, Gideon E. Anholt, Rotem Keren and Avishai Henik

Abstract

Obsessive compulsive disorder (OCD) is characterized by intrusive and anxiety evoking thoughts followed by repetitive behaviors (compulsions). Accumulative evidence revealed neuropsychological deficits in executive functions, especially in inhibitory mechanisms, in OCD patients. The connection between inhibitory control and the onset and maintenance of OCD is yet unclear. Task control—a mechanism responsible for promoting and maintaining goal directed actions and suppressing irrelevant actions that stimuli associatively and automatically evoke—was found to be contingent upon inhibitory control. Specifically, task control was found to be inadequate in OCD patients. We propose here that deficient task control might function as a mediator between inhibitory control deficit and the development of OCD. The difficulty to inhibit irrelevant behaviors related to intrusive thoughts inflates the perceived importance of these thoughts, which eventually are interpreted as catastrophic and thus should be suppressed by committing compulsive behavior. Paradoxically, these repetitive behaviors increase the anxiety first aroused by the obsessions and a vicious circle is perpetuated.

Key words: OCD, executive functions, inhibitory control, task conflict

Declaration of interest: the authors declare that they do not have any conflict of interest and that APA ethical standards were followed.

Eyal Kalanthroff, Gideon E. Anholt, Rotem Keren and Avishai HenikDepartment of Psychology and Zlotowski Center for Neuroscience, Ben-Gurion University of Negev, Beer Sheva, Israel.

Corresponding authorEyal Kalanthroff, Department of Psychology, Ben-Gurion University of the Negev, P.O.B. 653, Beer Sheva, Israel 84105. Phone: 972-8-6477209, Fax: 972-8-6472072, E-mail: [email protected]

Obsessive-compulsive disorder (OCD) is a highly debilitating anxiety disorder with a lifetime prevalence of 2%-3% (Weisman et al. 1994, Huppert et al. 2009). OCD is characterized by recurrent intrusive thoughts or impulses (obsessions), and repetitive, irresistible behaviors (compulsions) aimed to prevent the feared consequences from happening and to avoid anxiety (American Psychiatric Association 2000). OCD patients’ typical behaviors tend to inflict paradoxical effects—increasing rather than decreasing the anxiety caused by obsessions— effectively perpetuating compulsions (Salkovskis 1999, van den Hout and Kindt 2003, van den Hout et al. 2008). Intrusive thoughts occur in at least 90% of the general population and are not pathological per se (Rachman and de Silva 1978). The cognitive behavioral model for OCD (Salkovskis 1999) indicates that the vicious cycle of obsessions and compulsions begins with giving catastrophic interpretations to such intrusive thoughts.

In a previous paper, we suggested that “impaired response inhibition is related to the development of OCD metacognitive beliefs. The experience of difficulty to inhibit behavior tendencies related to intrusive

thoughts may lead to the perceptions of these thoughts as important and likely to occur. As a consequence, these patients may engage in thought suppression as well as compulsive behavior, and become entangled in a vicious circle” (Anholt et al. 2012, p. 74). This suggestion was based on two lines of research: The first refers to the accumulating evidence suggesting that OCD patients demonstrate neuropsychological deficits in executive functions (e.g., Lucey et al. 1997, Penades et al. 2005, for reviews see also Greisberg and McKay 2003, Kuelz et al. 2004,) and specifically in inhibitory control (Bannon et al. 2002, Penades et al. 2007). Several researchers even suggested inhibitory control deficit to be a cognitive endophenotype of OCD (Menzies et al. 2007, Morein-Zamir et al. 2010, de Wit et al. 2012). Recently, we found that people with efficient inhibitory control are less likely to experience memory distrust (a core symptom of OCD) following repeated checking (Linkovski et al. 2012). The second line of research refers to a classic psychological theory by William James, who argued that people tend to give emotional meaning to their own behaviors, “we feel sorry because we cry, angry

© 2013 Giovanni Fioriti Editore s.r.l. 6161

Eyal Kalanthroff et al.

62 Clinical Neuropsychiatry (2013) 10, 3, Suppl. 1

OCD and task control – integrating basic cognitive science with cognitive behavior theory

The idea that OCD patients suffer from deficient task control makes sense since the disorder is mainly characterized by repeatedly performing irrelevant, though automatically triggered, tasks. Indeed, in recent studies we found indications for deficient task control in OCD patients (e.g., Kalanthroff, Henik et al. 2013). Prevor and Diamond (2005) showed that naming the color of a neutral object takes more time than naming the color of an abstract form and La Heij et al. (2010) showed that this effect occurs due to task conflict between the relevant task (i.e., color naming) and an automatic irrelevant task triggered by the objects. Importantly, this effect is commonly obtained only in young children, under the age of 7 years old, and it was suggested that this is due to their undeveloped task control mechanism. In our study (Kalanthroff, Anholt et al. 2013) we found object interference in adult OCD patients but not in a matched control group. This indicates that OCD is characterized by an inefficient task control mechanism.

We propose here that a pro-active\task control deficit is a potential link between poor inhibitory control and OCD. We suggest that OCD patients experience increased difficulties to inhibit tasks that are associatively evoked. In turn, performing these tasks starts a vicious cycle of compulsive behavior, intrusive thoughts, and attempts to suppress these thoughts (which causes a paradoxical effect). Lhermitte et al. (1986) suggested that people have action plans that are evoked by stimuli. These researchers found that some frontal lobe patients showed a utilization effect—a condition in which patients are unable to suppress these action plans and thus they operate on each stimulus they see (e.g., if they see a syringe they try to use is to inject the physician). Similarly, Cisek (2006) suggested that multiple motor plans are generated automatically across visuo-motor regions of the cortex in response to attended stimuli. Makris et al. (2011) aimed to investigate the “affordances theory” hypothesis that implies that visual objects can potentiate motor responses even in the absence of an intention to act. These researchers showed that physical properties of objects automatically activated specific motor codes in the brain. We suggest that OCD patients suffer from a similar situation on a much smaller scale—action plans are too potent and inhibition is too deficient; hence, some irrelevant behavior occurs.

In recent years, a few researchers have suggested that goal-directed action may be compromised in OCD patients and compulsions may be driven by maladaptive habits (Boulougouris et al. 2009). This is in line with the findings mentioned earlier concerning an executive control deficit in OCD (e.g., Greisberg and McKay 2003, Kuelz et al. 2004). Gillan et al. (2011) tested OCD patients and found that they had a deficit in goal-directed control and an overreliance on habits. This led the researchers to suggest that OCD patients’ urge to perform compulsive acts is mediated by a disruption in the balance between flexible, goal-directed action control and habitual behavior. Moreover, the authors concluded that due to their control deficit, patients with OCD are forced to rely on habits that can be triggered by stimuli, regardless of the desirability of the consequences.

Cognitive behavioral therapy (CBT) is the most efficient psychological treatment for OCD (Rosa-

because we strike, afraid because we tremble” (James 1884/1969, 1890/1950). A growing body of research supports the notion that a variety of motor movements can influence individuals’ thoughts and feelings (e.g., Buck 1980, Stepper and Strack 1993, Förster and Strack 1997, Chandler and Schwartz 2009). Based on these findings, we suggested that people who suffer from deficient inhibitory control have difficulties to inhibit irrelevant automatic responses related to their intrusive thoughts and that these behaviors facilitate fearful appraisals of these intrusions. However, the mechanism by which deficient inhibitory control causes these irrelevant behaviors is not yet fully clear.

Task control Back in 1979, Gibson proposed the “affordances

theory” that suggests that people perceive directly what tools afford in terms of meaningful actions. In that theory, Gibson suggested that when one sees an object in the environment one immediately perceives not only the external features but also its affordances—a feature of the object that clearly identifies how the object could be used. Examples include the grasp-ability of a stick, the lift-ability of an object, the click-ability of a light switch and so on. In other words, we perceive directly what tools afford in terms of meaningful actions. Gibson’s ideas corresponds with Monsell’s (2003) suggestion that task sets can be activated by the perception of a stimulus attribute that is strongly associated with a particular task set (“exogenous activation”). In sum, converging evidence suggests that a stimulus can trigger performance of a task that has acquired a strong association with it (Rogers and Monsell 1995, Allport and Wylie 2000, Waszak et al. 2003).

Automatically triggered irrelevant and unwanted tasks cause task conflict (Goldfarb and Henik 2007, Haggard 2008, Steinhauser and Hubner 2009) and a specific control mechanism is needed in order to suppress those tasks—task control, which has been suggested to be very efficient in healthy adults (La Heij et al. 2010, Kalanthroff, Goldfarb et al. 2013). In other words, stimuli automatically trigger tasks; when these stimuli trigger an irrelevant task, task control is needed in order to suppress these tendencies. Because this control mechanism is usually very efficient, task conflict can only be behaviorally studied in children, since it is not yet fully developed at an early age (La Heij and Boelens 2011), or by using procedures that reduce task control efficiency (e.g., Braverman and Meiran 2010, Kalanthroff et al. 2012, Kalanthroff, Goldfarb et al. 2013, Kalanthroff and Henik 2013a). Task control can be conceptualized as ‘pro-active’ control that is responsible for keeping constant amplification of the relevant task units and constant inhibition of the irrelevant tasks (see also Braver 2012, MacDonald et al. 2000).

In several recent studies we found that task control is highly contingent upon inhibition control (e.g., Kalanthroff, Goldfarb et al. 2013). In one study, we found significant behavioral evidence for task conflict in adults who performed poorly in a response inhibition task (Kalanthroff and Henik 2013b). As mentioned earlier, OCD is a disorder characterized by deficient inhibitory control. Thus, it is reasonable to assume that patients suffering from OCD will also experience difficulties in effectively activating the task control mechanism.

62

Control over irrelevant tasks in OCD

Clinical Neuropsychiatry (2013) 10, 3, Suppl. 1 63

Psychology 71, 421-430.Gibson JJ (1979). The ecological approach to visual

perception. Houghton Mifflin, Michigan. Gillan CM, Papmeyer M, Morein-Zamir S, Sahakian BJ,

Fineberg NA, Robbins TW, de Wit, S (2011). Disruption in the balance between goal-directed behavior and habit learning in obsessive-compulsive disorder. American Journal of Psychiatry 168, 7, 718-726.‏

Goldfarb L, Henik A (2007). Evidence for task conflict in the Stroop effect. Journal of Experimental Psychology: Human Perception and Performance 33, 1170-1176.

Greisberg S, McKay D (2003). Neuropsychology of obsessive-compulsive disorder: a review and treatment implications. Clinical Psychology Review 23, 95-117.

Haggard P (2008). Human volition: towards a neuroscience of will. Nature Reviews Neuroscience 9, 934-946.

Huppert JD, Simpson HB, Nissenson KJ, Liebowitz MR, Foa E (2009). Quality of life and functional impairment in obsessive-compulsive disorder: A comparison of patients with and without comorbidity, patients in remission, and healthy controls. Depression and Anxiety 26, 39-45.

James W (1950). The principles of psychology. Dover, New York. (Original work published 1890.)

James W (1969). What is an emotion? In J William (ed) Collected essays and reviews. Russell and Russell, New York. (Original work published 1884.)

Kalanthroff E, Anholt EG, Henik A (2013). Always on guard: Test of high vs. low control conditions in Obsessive-Compulsive Disorder patients. Manuscript submitted for publication.

Kalanthroff E, Goldfarb L, Henik A (2013). Evidence for interaction between the Stop-Signal and the Stroop task conflict. Journal of Experimental Psychology: Human Perception and Performance 39, 579-592.

Kalanthroff E, Goldfarb L, Usher M, Henik A (2012). Stop interfering: Stroop task conflict independence from informational conflict and interference. Quarterly Journal of Experimental Psychology. Advance online publication. doi: 10.1080/17470218.2012.741606

Kalanthroff E, Henik A (2013a). Preparation time modulates pro-active control and enhances task conflict in task switching. Psychological Research. Advance online publication. 10.1007/s00426-013-0495-7.

Kalanthroff E, Henik A (2013b). Individual but not fragile: Individual differences in task control predict Stroop facilitation. Consciousness and Cognition 22, 413-419.

Kalanthroff E, Henik A., Anholt EG (2013). To do or not to do: Is a task control Deficit the missing link between a response inhibition deficit and Obsessive-Compulsive Disorder? Manuscript submitted for publication.

Kuelz AK, Hohagenb F, Voderholzer U (2004). Neuropsychological performance in obsessive-compulsive disorder: a critical review. Biological Psychology 65, 185-236.

La Heij W, Boelens H (2011). Color–object interference: Further tests of an executive control account. Journal of Experimental Child Psychology 108, 156-169.

La Heij W, Boelens H, Kuipers JR (2010). Object interference in children’s colour and position naming: Lexical interference or task-set competition? Language and Cognitive Processes 25, 568-588.

Lhermitte F, Pillon B, Serdaru M (1986). Human autonomy and the frontal lobes. Part I: imitation and utilization behavior: A neuropsychological study of 75 patients. Annals in Neurology 19, 326-334.

Linkovski O, Kalanthroff E, Henik A, Anholt G (2013). Did

Alcázar et al. 2008). CBT for OCD mainly consists of exposures (to stimuli and situations that evoke compulsions) and response prevention—ERP. In a way, it can be said that ERP is a kind of training program for improving task conflict control; an associative task is evoked and patients are to suppress the automatic behavior. Rosa-Alcázar et al. found that there are no indications that cognitive therapy using ERP treatment improves the outcome of OCD patients. This finding is in line with our suggestion that task control deficiency, which relies on inhibitory control, is a core element in the development and maintenance of OCD.

To conclude, we propose that OCD patients, who are known to suffer from inhibitory deficit, experience difficulties in solving task conflicts. Namely, automatically triggered tasks or action plans are not efficiently suppressed. In turn, this causes activation of irrelevant behaviors or compulsions, which lead to irrelevant interpretations that are coupled with intrusive thoughts in a way that perpetuates the compulsive behavior.

References Allport A, Wylie G (2000). Task-switching, stimulus–response

bindings and negative priming. In S Monsell J Driver (Eds) Control of cognitive processes: Attention and performance XVIII. MIT Press, Cambridge.

American Psychiatric Association (2000). Diagnostic and statistical manual of mental disorders, 4th. ed. text rev. APA, Washington.

Anholt G, Linkovski O, Kalanthroff E, Henik A (2012). If I do it, it must be important: Integrating basic cognitive research findings with cognitive behavior theory of obsessive-compulsive disorder. Psicoterapia Cognitivo Comportamentale 18, 69-77.

Bannon S, Gonsalvez CJ, Croft RJ, Boyce PM (2002). Response inhibition deficits in obsessive-compulsive disorder. Psychiatry Research 110, 165-174.

Boulougouris V, Chamberlain SR, Robbins TW (2009). Cross-species models of OCD spectrum disorders. Psychiatry Research 170, 15-21.

Braver TS (2012). The variable nature of cognitive control: a dual mechanisms framework. Trends in Cognitive Sciences 16, 106-113.

Braverman A, Meiran N (2010). Task conflict effect in task switching. Psychological Research 74, 568-567.

Buck R (1980). Nonverbal behavior and the theory of emotion: The facial feedback hypothesis. Journal of Personality and Social Psychology 38, 811-824.

Chandler J, Schwartz N. (2009). How extending your middle finger affects your perception of others: Learned movements influence concept accessibility. Journal of Experimental Social Psychology 45, 123-128.

Cisek P (2006). Integrated neural processes for defining potential actions and deciding between them: A computational model. The Journal of Neuroscience 26, 9761-9770.

de Wit SJ, de Vries FE, van der Werf YD, Cath DC, Heslenfeld DJ, Veltman EM, et al. (2012). Presupplementary motor area hyperactivity during response inhibition: A candidate endophenotype of obsessive-compulsive disorder. American Journal of Psychiatry 169, 1100-1108.

Förster J, Strack F (1996). Influence of overt head movements on memory for valenced words: A case of conceptual-motor compatibility. Journal of Personality and Social

63

Eyal Kalanthroff et al.

64 Clinical Neuropsychiatry (2013) 10, 3, Suppl. 1

obsessions. Behaviour Research and Therapy 16, 233-248.Rogers RD, Monsell S (1995). Costs of a predictable switch

between simple cognitive tasks. Journal of Experimental Psychology: General 124, 207-231.

Rosa-Alcázar AI, Sánchez-Meca J, Gómez-Conesa A, Marín-Martínez F (2008). Psychological treatment of obsessive–compulsive disorder: A meta-analysis. Clinical Psychology Review 28, 1310-1325.

Salkovskis PM (1999). Understanding and treating obsessive-compulsive disorder. Behaviour Research and Therapy 37, s29-s52.

Steinhauser M, Hubner R (2009). Distinguishing response conflict and task conflict in the Stroop task: Evidence from ex-Gaussian distribution analysis. Journal of Experimental Psychology: Human Perception and Performance 35, 1398-1412.

Stepper S, Strack F (1993). Proprioceptive determinants of emotional and nonemotional feelings. Journal of Personality and Social Psychology 64, 211-220.

van den Hout M, Kindt M (2003). Repeated checking causes memory distrust. Behaviour Research and Therapy 41, 301-316.

van den Hout MA, Engelhard IM, de Boer C, du Bois A, Dek E (2008). Perseverative and compulsive-like staring causes uncertainty about perception. Behaviour Research and Therapy 46, 12, 1300-1304.

Waszak F, Hommel B, Allport DA (2003). Task switching and long-term priming: Role of episodic stimulus-task bindings in task-shift costs. Cognitive Psychology 46, 361–413.

Weisman MM, Bland RC, Canino GJ, Greenwald S, Hwu HG, Lee LK, et al. (1994). The cross-national epidemiology of obsessive compulsive disorder. Journal of Clinical Psychiatry 55, suppl. 3, 5-10.

I turn off the stove? Good inhibitory control can protect from influences of repeated checking. Journal of Behavior Therapy and Experimental Psychiatry 44, 30-36.

Lucey JV, Burness CE, Costa DC, Gacinovic S, Pilowsky LS, Ell PJ, et al. (1997). Wisconsin Card Sorting Task (WCST) errors and cerebral blood flow in obsessive-compulsive disorder (OCD). British Journal of Medical Psychology 70, 403-411.

MacDonald AW, Cohen JD, Stenger VA, Carter CS (2000). Dissociating the role of dorsolateral prefrontal cortex and anterior cingulate cortex in cognitive control. Science 288, 1835-1838.

Makris S, Hadar AA, Yarrow K (2011). Viewing objects and planning actions: On the potentiation of grasping behaviors by visual objects. Brain and Cognition 77, 257-264.

Menzies L, Achard S, Chamberlain SR, Fineberg N, Chen CH, del Campo N, et al. (2007). Neurocognitive endophenotypes of obsessive-compulsive disorder. Brain 130, 3223-3236.

Monsell S (2003). Task switching. Trends in Cognitive Sciences 7, 134-140.

Morein-Zamir S, Fineberg NA, Robbins TW, Sahakian BJ (2010). Inhibition of thoughts and actions in obsessive-compulsive disorder: extending the endophenotype? Psychological Medicine 40, 263-272.

Penades R, Catalan R, Andres S, Salamero M, Gasto C (2005). Executive function and nonverbal memory in obsessive-compulsive disorder. Psychiatry Research 133, 81-90.

Penades R, Catalan R, Rubia K, Andres S, Salamero M, Gasto C (2007). Impaired response inhibition in obsessive compulsive disorder. European Psychiatry 22, 404-410.

Prevor MD, Diamond A (2005). Color-object interference in young children: A Stroop effect in children 3½-6½ years old. Cognitive Development 20, 256-274.

Rachman S, de-Silva P (1978). Abnormal and normal

64

65

CHAPTER III: General Discussion

This work examined the idea of task conflict and the task control mechanism.

First I will summarize the main results from the four experiment groups, then I will

elaborate on the theoretical importance of the findings and I will present a model that

can account for the main findings of this work, and finally, I will discuss the possible

implications for OCD that were presented in the theoretical model at the end of

Chapter 2.

In Experiment Group 1 we used a version of the Stroop task similar to

Goldfarb and Henik (2007). In this task 75% of the trials were neutrals and in 50% of

the trials there was a valid prime-cue (X) indicating whether the upcoming trial would

be neutral or not. Unlike in Goldfarb and Henik's study, the current study did not

include incongruent trials, thus there was no chance of making a mistake by initiating

the reading task. We found a reverse facilitation effect (an indication for task conflict)

in the non-informative cue condition. The results suggest that our reverse facilitation

effect was smaller than the one obtained in Goldfarb and Henik’s original study. This

indicates that the magnitude of task conflict does indeed depend on informational

conflict to some extent (though no direct comparison between studies was conducted).

The fact that the results did not replicate when a color word was used as neutral,

instead of a meaningless letter string (e.g., XXXX), strengthen the suggestion that the

reverse facilitation effect occurred due to task conflict. In Experiment 2, 136

participants conducted the Stroop and stop-signal tasks (in a counter-balanced order).

We found a significant negative correlation between the Stroop facilitation effect and

the SSRT (stop-signal reaction time; a measure of inhibition efficiency). More

specifically, we found a correlation between Stroop congruent trial RT and SSRT

even when Stroop neutral RT (as a measurement for general processing and

responding speed) was controlled for. When dividing participants into six equal

groups, according to their SSRTs, we found a significant reverse facilitation effect in

the longest SSRT group (lowest inhibition efficiency). In Experiment Group 3 we

combined the Stroop and stop-signal tasks and found a reverse facilitation effect in

erroneous responding to stop-signal trials. This finding suggests that when inhibitory

control fails (due to random fluctuations), task control also fails and a reverse

66

facilitation appears. Thus, this finding strengthens the finding from Experiment 2

indicating that task control is largely based on inhibitory control. Similar to

Experiment Group 1, we suggest that the results can be explained by task conflict

alone. This suggestion is reinforced by the fact that we did not replicate the reverse

facilitation effect when a non-color word was used as a neutral. In Experiment Group

4 we aimed to investigate task conflict in a situation in which the task was not

constant―task switching. This was encouraged by Braver’s (2012) dual mechanisms

of control theory that suggested control can be divided into two separate components:

the reactive control that is recruited in the face of a conflict and pro-active control that

is recruited throughout the performance of the task and is mainly responsible for

keeping the task set and instruction activated. In a combined task-switching Stroop

task, we found a reverse facilitation effect. Most importantly, we found that this

reverse facilitation effect was contingent upon preparation time―the time between

the appearance of the task cue (word reading or color naming) and the target stimulus

(the Stroop stimulus).

In what follows I will elaborate on the theoretical importance of our findings.

It seems appropriate to interpret our findings according to the theoretical framework

proposed by Monsell, Taylor, and Murphy (2001). These researchers suggested that at

least part of the interference that is commonly observed in Stroop tasks is due to

competition between the relevant task and the irrelevant automatic task, that is, task

conflict. The underlying idea for the notion of task conflict is that a stimulus may

trigger performance of a task that acquired a strong association with it (Allport &

Wylie, 2000; Rogers & Monsell, 1995; Waszak et al., 2003). Thus, while in Stroop

incongruent trials the effects of the response conflict and the task conflict add up,

leading to a robust slowing down effect, in Stroop congruent trials, they are in

opposite directions (the information facilitates while the task interferes) resulting is

non-robust effects with marked variability. It has been suggested that commonly in

healthy adults, task control is very efficient (e.g., Goldberg & Henik, 2007; La Heij &

Boelens, 2011; La Heij et al., 2010), hence, the task conflict is not behaviorally

evident under standard conditions—perhaps this is the reason why task conflict is

barely referred to in Stroop papers and in the executive control literature in general.

Nevertheless, as this work demonstrates, task conflict can be observed in special

67

conditions or with specific individuals. Note that in this work, we suggested that

Stroop neutral trials do not contain a conflict at all, although it is possible that even

these (neutral) trials contain a conflict but at a very low level because they hardly

engage attention (since the system can filter the neutral stimuli as non-words very

quickly). This alternative interpretation should be investigated in future works.

To summarize, the current work demonstrates that variability in Stroop reverse

facilitation arises in response to task contingencies within the same individual, or

between individuals in the same task. The interpretation proposed is that this variation

is due to systematic changes in the pro-active task control that the participants can

recruit. This control process can vary as a result of task contingencies (such as the

time from the task cue to the stimulus) or the effectiveness of the control system (and

specifically, the inhibitory control system). We suggest that when pro-active control

fails, the participant experiences not only response conflict but also task conflict.

In order to better understand the reverse facilitation effect, we have developed

a neural network computational model of the Stroop task (Kalanthroff, Davelaar,

Henik, & Usher, in preparation). Here I briefly highlight the operation of the model

and predictions that emanate from it, as a summary of the results obtained in the

current work.

The model follows previous connectionist Stroop models (e.g., Botvinick et

al., 2001; Cohen et al., 1990; De Pisapia & Braver, 2006), by assuming a set of task

control units that bias information processing towards the relevant task dimension;

importantly, these task units (component C in Figure 1) are connected by bilateral

connections to the input stimuli (components A & B in Figure 1). Under high pro-

active task control conditions (component E in Figure 1), the effect of the bottom-up

connections is negligible, as the active color-demand (component C in Figure 1)

inhibits the competing word-reading unit (component B in Figure 1), and thus we

obtain the standard result: RTs are faster for congruent and slower for incongruent

trials, compared with neutral trials. However, under low pro-active task control (when

the observer does not engage the task control unit in advance of the stimulus

presentation), the relevant task control unit (color naming) is only weekly activated.

In this situation, the bottom-up connection from stimuli to the task units triggers

68

activation in those units. For neutral stimuli only the color unit is activated, as there is

a color but no word reading material. In both congruent and incongruent trials, the

two-task demand units are activated bottom-up. In the absence of strong inhibition

from the color-demand unit (week pro-active task control), the color-word unit is also

activated, resulting in task conflict (component C in Figure 1). This model is

consistent with the idea that stimuli can activate associated tasks and with the idea

that congruent Stroop trials contain a conflict. Finally, the model predicts that task

conflict inhibits responses (component D in Figure 1). This inhibition is maintained

until the task conflict is resolved via the re-active control. This model can account for

the findings showing that under low pro-active task control conditions there is an

increase in both the congruent and the incongruent Stroop RT, compared with

neutrals. This is because in both types of trials the response is slowed down until the

task conflict is resolved. This delay adds to the regular Stroop RT, resulting in a

reversal of the facilitation effect.

Figure 1. A schematic model demonstrating the work of two control units (component

C & E in the figure) in the Stroop task. Arrows represent activation; the solid point

BLUE GREEN

BLUE GREEN BLUE GREEN

Color Word

PC

Color feature layer (A) Lexical layer (B)

Response layer (D) Task demand layer (C)

2 2

2.5

2.5

(E)

69

represents inhibition; wider arrows represent stronger connections between units;

numbers represent the weight of activation; the star represents conflict.

In the last part of this discussion I will address the possible implications of the

current work to the understanding and treating of OCD. In recent years there has been

a debate in the literature regarding the importance of inhibitory deficits to

understanding OCD. I will start by describing this debate. OCD was considered to be

untreatable until the 1960s, when Meyer (1966) developed a new behavioral

intervention, which later became known as "exposure and response prevention" (ERP)

intervention. With time this turned into the treatment of choice for OCD (Olatunji,

Davis, Powers, & Smits, 2012). Since there were still some disadvantages to this

treatment (Clark, 1999), the interest in cognitive theories of OCD increased from the

1980s onwards. It was believed that adding cognitive interventions to the treatment of

OCD would increase treatment effectiveness and alleviate some of the behavioral

therapy disadvantages. There are no doubts that research driven from the cognitive

approach has led to some very important findings and advanced the understating of

OCD (for a review see Calkins et al., 2013). Nevertheless, many argue that analysis of

current research findings indicates that cognitive theory and treatment of OCD have

yet to achieve their set goals (for a review see Anholt & Kalanthroff, 2013). Recently,

findings showing inhibitory deficits in OCD patients have altered researchers'

attention again to an inhibitory deficit reflected in a behavior approach (e.g.,

Chamberlain et al., 2005; de Wit et al., 2012; Lennertz et al., 2012; Menzies et al.,

2007; Morein-Zamir, Fineberg, Robbins, & Sahakian, 2010). For example, in a recent

paper (Linkovski, Kalanthroff, Henik, & Anholt, 2013) we found that people with

poor inhibitory control are more prone to become entangled in an OCD “vicious

cycle” of doubts and compulsions. In a different study, which was described in the

last part of section 2 (Kalanthroff, Anholt, & Henik, 2014), we conducted a Stroop

task with high- and low-control conditions (33% vs. 75% neutrals in a block,

respectively) on OCD patients and healthy controls. In the high-control condition we

found a reduced facilitation in OCD patients, which strengthens the task control

deficit theory (e.g., Anholt et al., 2012). Moreover, while healthy controls showed

larger interference and smaller facilitation in the low-control condition, OCD patients

showed no difference between the two conditions, indicating impairment in cognitive

70

control flexibility. Although inhibitory deficit seems to be a very promising candidate

for the core symptom of OCD, there are still some problems with it (for a meta-

analysis see Abramovitch et al., 2013). Mainly, two problems were raised: (1) many

psychopathologies are characterized by inhibitory deficit and (2) correlations between

inhibitory deficit and symptom severity have yet to be found. For those reasons, some

researchers refer to inhibitory deficit as simply an epiphenomenon of OCD (e.g.,

Abramovitch et al., 2012).

In light of all of the above, we proposed a model in which task control is the

“missing link” between inhibitory deficit and OCD. The theory we proposed, "if I do

it, it must be important" (Anholt et al., 2012), has many cognitive parts but mainly the

theory suggests that a deficit in task control is at the foundation of the process that

will develop into (and later on maintain) OCD. Essentially, we suggest that due to

poor task control (and poor inhibition), OCD patients engage in irrelevant tasks.

Subsequent to engaging in these tasks, they will have to explain these behaviors to

themselves and this could lead to distorted cognitive beliefs. Further research is

needed in order to fully understand task control importance in OCD and how

understanding task control could be put to use in clinical practice.

71

References

Aarts, E., Roelofs, A., & van Turennouta, M. (2009). Attentional control of task and

response in lateral and medial frontal cortex: Brain activity and reaction time

distributions. Neuropsychologia, 47, 2089-2099.

Abramovitch, A., Abramowitz, J. S., & Mittelman, A. (2013). The neuropsychology

of adult obsessive–compulsive disorder: A meta-analysis. Clinical Psychology

Review, 33, 1163-1171.

Abramovitch, A., Dar, R., Schweiger, A., & Hermesh, H. (2011). Neuropsychological

impairments and their association with obsessive-compulsive symptom

severity in obsessive-compulsive disorder. Archives of Clinical

Neuropsychology, 26, 364-376.

Abramovitch, A., Dar, R., Schweiger, A., & Hermesh, H. (2012). Comparative

neuropsychology of adult obsessive-compulsive disorder and attention

deficit/hyperactivity disorder: Implications for a novel executive overload

model of OCD. Journal of Neuropsychology, 6, 161-191.

Allport, A. (1980). Attention and performance. In G. Claxton (Ed.), Cognitive

psychology: New directions (pp. 112-153). London: Routledge & Kegan Paul.

Allport, A., & Wylie, G. (2000). Task switching, stimulus–response bindings and

negative priming. In S. Monsell, & J. Driver (Eds.), Control of cognitive

processes: Attention and performance XVIII (pp. 35–70). Cambridge, MA:

MIT Press.

American Psychiatric Association. (2000). Diagnostic and statistical manual of mental

disorders (4th ed., text rev.). Washington, DC: Author.

Anholt, G. E., & Kalanthroff, E. (2013). Letter to the editor: Recent advances in

research on cognition and emotion in OCD: A review. Current Psychiatry

Reports, 15:416, 1-4.

Anholt, G. E., Linkovski, O., Kalanthroff, E., & Henik, A. (2012). If I do it, it must be

important: Integrating basic cognitive research findings with cognitive

behavior theory of obsessive-compulsive disorder. Psicoterapia

Comportamental e Cognitiva, 18, 69-79.

Aron, A. R., Behrens, T. E., Smith, S., Frank, M. J., & Poldrack, R. A. (2007).

Triangulating a cognitive control network using diffusion-weighted magnetic

72

resonance imaging (MRI) and functional MRI. The Journal of Neuroscience,

27, 3743–3752.

Aron, A. R., & Poldrack, R. A. (2006). Cortical and subcortical contributions to stop

signal response inhibition: role of the subthalamic nucleus. The Journal of

Neuroscience, 26, 2424-2433.

Bannon, S., Gonsalvez, C. J., Croft, R. J., & Boyce, P. M., (2002). Response

inhibition deficits in obsessive-compulsive disorder. Psychiatry Research 110,

165-174.

Bench, C. J., Frith, C. D., Grasby, P. M., Friston, K. J., Paulesu, E., Frackowiak, R.

S., & Dolan, R. J. (1993). Investigations of the functional anatomy of attention

using the Stroop test. Neuropsychologia, 31, 907–922.

Botvinick, M., Nystrom, L. E., Fissell, K., Carter, C. S., & Cohen, J. D. (1999).

Conflict monitoring versus selection-for-action in anterior cingulated cortex.

Nature, 402, 179–181.

Botvinick, M. M., Braver, T. S., Barch, D. M., Carter, C. S., & Cohen, J. D. (2001).

Conflict monitoring and cognitive control. Psychological Review, 108, 624–

652.

Braver, T. S. (2012). The variable nature of cognitive control: a dual mechanisms

framework. Trends in Cognitive Sciences, 16, 106-113.

Calkins, A. W., Berman, N. C., & Wilhelm, S. (2013). Recent advances in research on

cognition and emotion in OCD: A review. Current Psychiatry Reports, 15, 1-

7.

Carter, C. S., Botvinick, M. M., & Cohen, J. D. (1999). The contribution of the

anterior cingulated cortex to executive processes in cognition. Reviews in

Neuroscience, 10, 49–57.

Carter, C. S., Braver, T. S., Barch, D. M., Botvinick, M. M., Noll, D., & Cohen, J. D.

(1998). Anterior cingulate cortex, error detection, and the online monitoring of

performance. Science, 280, 747–749.

Carter, C. S., Mintun, M., & Cohen, J. D. (1995). Interference and facilitation effects

during selective attention: An H215O PET study of Stroop task performance.

NeuroImage, 2, 264–272.

Chamberlain, S. R., Blackwell, A. D., Fineberg, N. A., Robbins, T. W., & Sahakian,

B. J. (2005). The neuropsychology of obsessive compulsive disorder: the

importance of failures in cognitive and behavioural inhibition as candidate

73

endophenotypic markers. Neuroscience & Biobehavioral Reviews, 29, 399-

419.

Cisek, P. (2006). Integrated neural processes for defining potential actions and

deciding between them: A computational model. The Journal of Neuroscience,

26, 9761-9770.

Clark, D. A. (1999). Cognitive behavioral treatment of obsessive-compulsive

disorders: a commentary. Cognitive and Behavioral Practice, 6, 408–415.

Cohen, J. D., Dunbar, K., & McClelland, J. L. (1990). On the control of automatic

processes: A parallel distributed processing account of the Stroop effect.

Psychological Review, 97, 332-361.

Cohen, J. D., & Servan-Schreiber, D. (1992). Context, cortex and dopamine: A

connectionist approach to behavior and biology in schizophrenia.

Psychological Review, 99, 45–77.

De Pisapia, N., & Braver, T. S. (2006). A model of dual control mechanisms through

anterior cingulate and prefrontal cortex interactions. Neurocomputing, 69,

1322-1326.

de Wit, S. J., de Vries, F. E., van der Werf, Y. D., Cath, D. C., Heslenfeld, D. J.,

Veltman, E. M., … & van den Heuvel, O. A. (2012). Presupplementary motor

area hyperactivity during response inhibition: A candidate endophenotype of

obsessive-compulsive disorder. American Journal of Psychiatry, 169, 1100-

1108.

Gava, I., Barbui, C., Aguglia, E., Carlino, D., Churchill, R., De Vanna, M., &

McGuire, H. (2007). Psychological treatments versus treatment as usual for

obsessive compulsive disorder (OCD). Cochrane Database of Systematic

Reviews, 2 (article no. CD005333).

Gibson, J. J. (1979). The ecological approach to visual perception. Boston, MA:

Houghton Mifflin.

Goldfarb, L., & Henik, A. (2007). Evidence for task conflict in the Stroop effect.

Journal of Experimental Psychology: Human Perception and Performance,

33, 1170–1176.

Grabill, K., Merlo, L., Duke, D., Harford, K.L., Keeley, M.L., Geffken, G.R., &

Storch, E.A. (2008). Assessment of obsessive-compulsive disorder: A review.

Journal of Anxiety Disorders, 22, 1-17.

74

Greisberg, S., & McKay, D. (2003). Neuropsychology of obsessive-compulsive

disorder: a review and treatment implications. Clinical Psychology Review, 23,

95-117.

Haggard, P. (2008). Human volition: towards a neuroscience of will. Nature Reviews

Neuroscience, 9, 934-946.

Harnishfeger, K. K. (1995). The development of cognitive inhibition. In F. N.

Dempster and C. J. Brainerd (Eds.), Interference and inhibition in cognition

(pp. 175-204). New York: Academic Press.

Huppert, J. D., Simpson, H. B., Nissenson, K. J., Liebowitz, M. R., & Foa, E. (2009).

Quality of life and functional impairment in obsessive-compulsive disorder: A

comparison of patients with and without comorbidity, patients in remission,

and healthy controls. Depression and Anxiety, 26, 39–45.

Kalanthroff, E., Anholt, G. E., & Henik, A. (2014). Always on guard: Test of high vs.

low control conditions in obsessive-compulsive disorder patients. Psychiatry

Research, 219, 322-388

Kuelz, A. K., Hohagen, F., & Voderholzer, U. (2004). Neuropsychological

performance in obsessive-compulsive disorder: a critical review. Biological

Psychology, 65, 185-236.

La Heij, W., & Boelens, H. (2011). Color–object interference: Further tests of an

executive control account. Journal of Experimental Child Psychology, 108,

156–169.

La Heij, W., Boelens, H., & Kuipers, J. R. (2010). Object interference in children’s

colour and position naming: Lexical interference or task-set competition?

Language and Cognitive Processes, 25, 568–588.

Lennertz, L., Rampacher, F., Vogeley, A., Schulze-Rauschenbach, S., Pukrop, R.,

Ruhrmann, S., ... & Wagner, M. (2012). Antisaccade performance in patients

with obsessive–compulsive disorder and unaffected relatives: further evidence

for impaired response inhibition as a candidate endophenotype. European

Archives of Psychiatry and Clinical Neuroscience, 262, 625-634.

Lhermitte, F., Pillon, B., & Serdaru, M. (1986). Human autonomy and the frontal

lobes. Part I: Imitation and utilization behavior: A neuropsychological study of

75 patients. Annals of Neurology, 19, 326-334.

Linkovski, O., Kalanthroff, E., Henik, A., & Anholt, G. E. (2013). Did I turn off the

stove? Good inhibitory control can protect from influences of repeated

75

checking. Journal of Behavior Therapy and Experimental Psychiatry, 44, 30-

36.

Lucey, J. V., Burness, C. E., Costa, D. C., Gacinovic, S., Pilowsky, L. S., Ell, P. J., …

Kerwin, R. W. (1997). Wisconsin Card Sorting Task (WCST) errors and

cerebral blood flow in obsessive-compulsive disorder (OCD). British Journal

of Medical Psychology, 70, 403-411.

MacDonald, A. W., Cohen, J. D., Stenger, V. A., & Carter, C. S. (2000). Dissociating

the role of the dorsolateral prefrontal and anterior cingulate cortex in cognitive

control. Science, 288, 1835-1838.

MacLeod, C. M. (1991). Half a century of research on the Stroop effect: An

integrative review. Psychological Bulletin, 109, 163–203.

MacLeod, C. M., & MacDonald, P. A. (2000). Interdimensional interference in the

Stroop effect: Uncovering the cognitive and neural anatomy of attention.

Trends in Cognitive Sciences, 4, 383–391.

Makris, S., Hadar, A. A., & Yarrow, K. (2011). Viewing objects and planning actions:

On the potentiation of grasping behaviors by visual objects. Brain and

Cognition, 77, 257-264.

Meiran, N., Diamond, G. M., Toder, D., & Nemets, B. (2011). Cognitive rigidity in

unipolar depression and obsessive compulsive disorder: Examination of task

switching, Stroop, working memory updating and post-conflict adaptation.

Psychiatry Research, 185, 149–156.

Menzies, L., Achard, S., Chamberlain, S. R., Fineberg, N., Chen, C. H., del Campo,

N., … Bullmore, E. (2007). Neurocognitive endophenotypes of obsessive-

compulsive disorder. Brain, 130, 3223–3236.

Meyer, V. (1966). Modification of expectations in cases with obsessional rituals.

Behaviour Research and Therapy, 4, 273–280.

Milham, M. P., Erickson, K. I., Banich, M. T., Kramer, A. F., Webb, A., Wszalek, T.,

& Cohen, N. J. (2002). Attentional control in the aging brain: Insights from an

fMRI study of the Stroop task. Brain and Cognition, 49, 277–296.

Miller, E. K., & Cohen, J. D. (2001). An integrative theory of prefrontal cortex

function. Annual Review of Neuroscience, 24, 167-202.

Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., & Wager,

T. D. (2000). The unity and diversity of executive functions and their

76

contributions to complex “frontal lobe” tasks: A latent variable analysis.

Cognitive Psychology, 41, 49-100.

Monsell, S. (2003). Task switching. Trends in Cognitive Sciences, 7, 134–140.

Monsell, S., Taylor, T. J., & Murphy, K. (2001). Naming the color of a word: Is it

responses or task sets that compete? Memory & Cognition, 29, 137–151.

Morein-Zamir, S., Fineberg, N. A., Robbins, T. W., & Sahakian, B. J. (2010).

Inhibition of thoughts and actions in obsessive-compulsive disorder: extending

the endophenotype? Psychological Medicine, 40, 263-272.

Muller, J., & Roberts, J. E., (2005). Memory and attention in Obsessive–Compulsive

Disorder: a review. Journal of Anxiety Disorders, 9, 1-28.

Nachev, P., Wydell, H., O’Neill, K., Husain, M., & Kennard, C. (2007). The role of

the pre-supplementary motor area in the control of action. NeuroImage, 36,

T155-T163.

Nigg, J. T. (2000). On inhibition/disinhibition in developmental psychopathology:

views from cognitive and personality psychology and a working inhibition

taxonomy. Psychological Bulletin, 126, 220-246.

Norman, D. A., & Shallice, T. (1986). Attention to action: willed and automatic

control of behaviour. In R. Davidson, G. Schwartz, & D. Shapiro (Eds.),

Consciousness and self-regulation: Advances in research and theory (vol. 4,

pp. 1-18). New York: Plenium Press.

Olatunji, B. O., Davis, M. L., Powers, M. B., & Smits, J. A. (2012). Cognitive-

behavioral therapy for obsessive-compulsive disorder: A meta-analysis of

treatment outcome and moderators. Journal of Psychiatric Research, 47, 33-

41.

Penades, R., Catalan, R., Andres, S., Salamero, M., & Gasto, C. (2005). Executive

function and nonverbal memory in obsessive-compulsive disorder. Psychiatry

Research, 133, 81-90.

Penades, R., Catalan, R., Rubia, K., Andres, S., Salamero, M., & Gasto, C. (2007).

Impaired response inhibition in obsessive compulsive disorder. European

Psychiatry, 22, 404-410.

Rafal, R., & Henik, A. (1994). The neurology of inhibition: integrating controlled and

automatic processes. In D. Dagenbach, & T. H. Carr (Eds.), Inhibitory

processes in attention, memory and language (pp. 1–52). New York:

Academic Press.

77

Roelofs, A., van Turennout, M., & Coles, M. G. (2006). Anterior cingulate cortex

activity can be independent of response conflict in Stroop-like tasks.

Proceedings of the National Academy of Sciences, 103, 13884-13889.

Rogers, R. D., & Monsell, S. (1995). Costs of a predictable switch between simple

cognitive tasks. Journal of Experimental Psychology: General, 124, 207–231.

Rubia, K., Smith, A. B., Brammer, M. J., & Taylor, E. (2003). Right inferior

prefrontal cortex mediates response inhibition while mesial prefrontal cortex is

responsible for error detection. NeuroImage, 20, 351–358.

Rubia, K., Smith, A. B., Taylor, E., & Brammer, M. (2007). Linear age‐correlated

functional development of right inferior fronto‐striato‐cerebellar networks

during response inhibition and anterior cingulate during error‐related

processes. Human Brain Mapping, 28, 1163-1177.

Salkovskis, P. M. (1999). Understanding and treating obsessive-compulsive disorder.

Behaviour Research and Therapy, 37, s29-s52.

Stein, D. J. (2002). Obsessive-compulsive disorder. Lancet, 360, 397-405.

Steinhauser, M., & Hubner, R. (2009). Distinguishing response conflict and task

conflict in the Stroop task: Evidence from ex-Gaussian distribution analysis.

Journal of Experimental Psychology: Human Perception and Performance,

35, 1398–1412.

Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of

Experimental Psychology, 18, 643– 662.

Tzelgov, J., Henik, A., & Berger, J. (1992). Controlling Stroop effects by

manipulating expectations for color words. Memory & Cognition, 20, 727-735.

van den Hout, M. A., Engelhard, I. M., de Boer, C., du Bois, A., & Dek, E. (2008).

Perseverative and compulsive-like staring causes uncertainty about perception.

Behaviour Research and Therapy, 34, 889–898.

van den Hout, M. A., & Kindt, M. (2003). Repeated checking causes memory distrust.

Behaviour Research and Therapy, 41, 301-316.

van Veen, V., & Carter, C. S. (2006). Conflict and cognitive control in the brain.

Current Directions in Psychological Science, 15, 237–240.

Verbruggen, F., & Logan, G. (2008). Response inhibition in the stop-signal paradigm.

Trends in Cognitive Sciences, 12, 418–424.

78

Waszak, F., Hommel, B., & Allport, D. A. (2003). Task switching and long-term

priming: Role of episodic stimulus-task bindings in task-shift costs. Cognitive

Psychology, 46, 361– 413.

Weisman, M. M., Bland, R. C., Canino, G. J., Greenwald, S., Hwu, H. G., Lee, L. K.,

… Yeh, E. K. (1994). The cross-national epidemiology of obsessive-

compulsive disorder. Journal of Clinical Psychiatry, 55 (suppl. 3), 5-10.

v

תקציר

הראו זה מכבר כי מטלות יכולות להיות, הדמיה התנהגותיות וטכניקות משימות באמצעות

עם תכונית של אובייקט אשר מקושרת אסוציאטיבית ומזוהה מאודתפיסה של ידי על מופעלת

המערכת הקוגניטיבית תופשת מיד לא בסביבה אדם רואה אובייקט כאשר —מסוימת משימה

את הפעולות האפשריות לביצוע על ועם רק את התכונות החיצוניות של אותו האובייקט אלא גם

ושמירה לקידום אחראימנגנון בקרת המטלה .משמעותיות פעולות של אותו האובייקט במונחים

המתעוררות באופן רלוונטיות לא אך אוטומטיות פעולות דיכויהתנהגות מכוונת מטרה תוך על

של מנגנון ותהחשוב ותוההשלכשל בקרת המטלה לרעיון ניתנה לב תשומת הרבה לא. אוטומטי

של העיקרית המטרה. ולוגיהירופסיכוזה לפסיכולוגיה הקוגניטיבית, הקלינית ולתחום של נ

ולאפיין את מנגנון בקרת המטלה ולפרט את ההשלכות החשובות לזהות הייתה הנוכחי המחקר

צבע. מטלה זה -התפקיד העיקרי בו נעשה שימוש במחקר זה הינו מטלת הסטרופ מילה. שלו

וצבע הדיו אשר אינם תואמים) המילה בין( קונפליקט המידע—משקפת שני קונפליקטים

מאפשרת טרופהס מטלת ,לפיכך). שיום צבע לבין קריאת המילה בין(וקונפליקט המטלה

כי תימצא הראשון בניסויאפקט ההאצה ההפוך. ─קונפליקט המטלה ל התנהגותית אינדיקציה

כאשר המטלה גם יתרחשכך שהראשון מידעמקונפליקט ה עצמאי ינוהקונפליקט המטלה

אותה התגובה כמו המטלה הרלוונטית (כך שאין אפשרות כלל תוביל ל תמיד האוטומטית

ו של יעילות כי , ערכתי מחקר הבדלים אינדיבידואליים ומצאתיהשני בניסוילקונפליקט מידע).

הניסוי עיכוב תגובה. –מנגנון בקרת המטלה נמצאת במתאם עם היעילות של מנגנון בקרה אחר

לקונפליקט המטלה בצעדים בהם היה כשל ציהאינדיק ומצא הזה הרעיון את המשיך השלישי

, השתמשי בגרסה של הרביעי בניסוי רגעי (כתוצאה מתנודות מקריות) במנגנון עיכוב התגובה.

כי בקרת המטלה תהיה אפקטיבית הרבה פחות כאשר סט תימצאהמטלה ו-פרדיגמת החלפת

שנו אפקט חיובי על ושלזמן ההכנה י )תכופות לעתים משתנהכאשר המטלה ( קבוע ואינ המטלה

מחקר על ל אפשריות השלכותבמספר דון, אבחלק החמישי. היכולת לנהל את קונפליקט המטלה

חלק זה יתבסס על ).OCD( כפייתית טורדנית מהפרעה הסובלים אנשים עבור המטלה בקרת

מטלה מוגבר ושהיכולת קונפליקטישנו OCD-תוצאות ניסוי המראות כי לאנשים הסובלים מ

קונפליקט זה נמצאת במתאם עם חומרת הסימפטומים. לסיום, בחלק הדיון אתאר מודל לפתור

מסביר את התוצאות של כל הניסויים בעבודה זאת. ממוחשב של רשתות עצביות אשר

הפרעה טרדנית כפייתית, תפקודים ניהוליים, עיכוב תגובה, קונפליקט המטלה, מילות מפתח:

סטרופ

iv

הגשת עבודת הדוקטור לשיפוטהצהרת תלמיד המחקר עם

אני אייל קלנטרוף מצהיר בזאת:

.חיברתי את חיבורי בעצמי, להוציא עזרת ההדרכה שקיבלתי מאת מנחה/ים

מתקופת היותי תלמיד/ת מחקרהחומר המדעי הנכלל בעבודה זו הינו פרי מחקרי.

ה טכנית הנהוגה בעבודה בעבודה נכלל חומר מחקרי שהוא פרי שיתוף עם אחרים, למעט עזר

ניסיונית. לפי כך מצורפת בזאת הצהרה על תרומתי ותרומת שותפי למחקר, שאושרה על ידם

ומוגשת בהסכמתם.

שם התלמיד: אייל קלנטרוף תאריך: ________________

חתימה: _________________

iii

יקהנ אבישי פרופ' של ותמיכתו בהדרכתו נעשתה העבודה

לפסיכולוגיה במחלקה

והחברה ברוח למדעי בפקולטה

ii

אפיונים –מנגנון בקרה לקונפליקט המטלה

והשלכות

מחקר לשם מילוי חלקי של הדרישות לקבלת תואר "דוקטור לפילוסופיה"

מאת

אייל קלנטרוף

הוגש לסנאט אוניברסיטת בן גוריון בנגב

אישור המנחה

די מחקר מתקדמים ע"ש קרייטמןאישור דיקן בית הספר ללימו

July 16, 2014 י"ח תמוז תשע"ד

באר שבע

אפיונים –מנגנון בקרה לקונפליקט המטלה

והשלכות

מחקר לשם מילוי חלקי של הדרישות לקבלת תואר "דוקטור לפילוסופיה"

מאת

אייל קלנטרוף

הוגש לסנאט אוניברסיטת בן גוריון בנגב

July 16, 2014 תשע"די"ח תמוז

באר שבע

Task Conflict Control Mechanism – Characteristics and ...

Documents

Transcript of Task Conflict Control Mechanism – Characteristics and ...