MANIPULATION OF AFFERENT FEEDBACK FROM EXTRAOCULAR

MUSCLES VIA JENDRASSIK MANEUVER IN BINOCULARLY INTACT

OBSERVERS AND DEAFFERENTED PATIENTS: EFFECTS ON VERGENCE,

VERSION AND HIGHER ORDER PERCEPTUAL JUDGEMENTS

By

Ewa Niechwiej-Szwedo

A thesis submitted in conformity with the requirements

for the degree of Doctor of Philosophy

Graduate Department of Rehabilitation Science

University of Toronto

Copyright by Ewa Niechwiej-Szwedo (2008)

ii

ABSTRACT

MANIPULATION OF AFFERENT FEEDBACK FROM EXTRAOCULAR

MUSCLES VIA JENDRASSIK MANEUVER IN BINOCULARLY INTACT

OBSERVERS AND DEAFFERENTED PATIENTS: EFFECTS ON VERGENCE,

VERSION AND HIGHER ORDER PERCEPTUAL JUDGEMENTS

Ewa Niechwiej-Szwedo

Doctor of Philosophy, 2008

University of Toronto

Graduate Department of Rehabilitation Science

The central nervous system can use two extraretinal sources to stay informed about

the position of the eyes in the orbit: outflow (efference copy) and inflow (afference). Palisade

endings (PE), found at the myotendinous junction of the multiply innervated fibers of the

global layer of extraocular muscles (EOM), are the putative receptors supplying the inflow

eye position signal. Seminal neuroanatomical tracing studies identified a distinct set of non-

twitch (NT) motoneurons whose activity does not add to the force used to move the eyes. It

has been suggested that NT motoneurons could be involved in modulating the gain of

sensory feedback from EOM analogous to the gamma-efferent fibres which control the

sensitivity of muscle spindles in skeletal muscles. The goal of this thesis was to test the

above hypothesis in humans using behavioural and psychophysical approaches. Jendrassik

Maneuver (JM), which is a forceful muscle contraction that facilitates the amplitude of all

iii

spinal and brainstem reflexes, was used to manipulate afferent feedback. The facilitation is

most likely due to a general up-regulation of the gamma system. It was hypothesized that if

NT motoneurons are analogous to gamma motoneurons, the JM should also increase the

activity of NT neurons and alter the afferent feedback from PE. As hypothesized, the JM

perturbation altered registered vergence eye position when binocularly intact observers

localized targets in depth but did not affect localization in the frontal plane associated with

saccades. Patients with congenital strabismus who have had surgeries on their EOM were not

affected by the JM perturbation. In contrast to the hypotheses, the JM did not affect higher

order perceptual judgments (size and depth constancies). Overall, these studies provide

insight into the putative mechanism involved in the control of sensory feedback from EOM.

In particular, the NT motoneurons might be involved in parametric adjustment of the

proprioceptive feedback loops to match the demands of different types of eye movements.

Understanding the role of proprioceptive feedback loops could have important clinical

implications for surgical treatment of strabismus.

iv

ACKNOWLEGMENTS

I would like to acknowledge the support of many individuals who helped and supported

me along the way in the past few years.

First, my sincere thanks go to my supervisors Marty Steinbach and Molly Verrier for

guiding me towards my goals. I have been extremely fortunate to work with Marty and

Molly, whose passion for neuroscience and life is an inspiration to all their students and

colleagues. Their support and encouragement have made this accomplishment possible.

I would also like to acknowledge the contribution of my committee members Esther

González and Agnes Wong, whose insight and direction enriched my thesis.

I also like to thank my colleagues at the Ocular Motor Lab and University of Toronto for

their support and friendship: Esther González, Linda Lillakas, Lumi Tarita-Nistor,

Olivera Karanovic, Diana Tajik-Parvinchi, Micheal Jurkiewicz and Rosalynn Miller. The

last few years have been an enjoyable and rewarding experience thanks to these new

friendships.

A special mention goes to my family who has always supported and encouraged me,

especially my parents who taught me to always strive for achievement.

I would also like to thank the late Aftab Patla for instigating my curiosity in the field of

visuomotor control.

v

DEDICATION

To my husband, Tomek Szwedo, for constantly reminding me that the journey is just as

important as the final destination.

“The only new voyage of discovery consists not in seeing new landscapes but in

having new eyes”

-Marcel Proust

vi

TABLE OF CONTENTS

Abstract………………………………………………………………………………………..ii

Acknowledgments……………………………………………………………………..……..iv

Dedication …………………………………………………………………………………….v

Table of Contents……………………………………………………..………………………vi

List of Tables……………………………………………………………………………........ix

List of Figures………………………………………………………………………...……….x

List of Abbreviations…………………………………………………………………...…...xiii

Glossary of Terms………………………………………………………………………...….xv

List of Appendices…………………………..………………………………………….......xvii

CHAPTER I: LITERATURE REVIEW

1.1 How does the brain know which way the eye is pointing? Inflow vs. outflow…………...1

1.1.1 Role of inflow during development……………………………………………..4

1.1.2 Role of inflow in oculomotor control…………………………………………...5

1.1.2.1 Evidence from animal studies…………………………………………5

1.1.2.2 Evidence from studies with humans…………………………………..7

1.1.3 Role of inflow in localization…………………………………………………...9

1.1.4 Summary…………………………………………………………………….....12

1.2 Extraocular muscles: anatomy, morphology and innervation……………………………12

1.2.1 Sensory receptors in extraocular muscles……………………………………...16

1.2.2 Dual innervation of extraocular muscles: proprioceptive hypothesis………….19

1.3 Role of gamma innervation in the skeletal system………………………………………21

vii

1.3.1 Anatomy & physiology………………………………………………………..21

1.3.2 Function: alpha-gamma co-activation…………………………………………23

1.3.3 Descending control: implications for function………………………………...23

1.3.4 Summary……………………………………………………………………….26

1.4 Jendrassik Maneuver: possible mechanisms……………………………………………..26

1.4.1 Gamma system………………………………………………………………....27

1.4.2 Alpha motoneuron excitability………………………………………………...28

1.4.3 Presynaptic disinhibition……………………………………………………….29

1.4.4 Time course of the reflex facilitation effect……………………………………30

1.4.5 Summary……………………………………………………………………….31

CHAPTER II: RESEARCH OBJECTIVES & HYPOTHESES………………………..33

CHAPTER III: RESULTS

3.1 Paper 1: Proprioceptive role for palisade endings in extraocular muscles: evidence from

the Jendrassik Maneuver……………………………………………………………………..35

3.2 Paper 2: Localization in the frontal plane is not susceptible to manipulation of afferent

feedback via the Jendrassik Maneuver……………………………………………………....67

3.3 Paper 3: Manipulation of extraocular muscles has no effect on higher order perceptual

judgments…………………………………………………………………………………….90

3.4 Paper 4: Localization in depth is not affected by the Jendrassik Maneuver in patients

operated for strabismus: case studies………………………………………………………118

viii

CHAPTER IV: GENERAL DISCUSSION

4.1 Summary of findings……………………………………………………………………131

4.2 Significance of the project……………………………………………………………...132

4.3 Limitations of the project……………………………………………………………….135

4.4 Future directions………………………………………………………………………..136

REFERENCES……………………………………………………………………………139

ix

LIST OF TABLES

Table 1: Primary and secondary actions of the EOM and their efferent innervation ……...13

Table 2: Parameters obtained from the linear regression model for individual

participants (Experiment 3, Paper 3)……………………………………………………..106

Table 3: Patients’ clinical characteristics and surgical procedures (Paper 4)…………….124

x

LIST OF FIGURES

Paper 1

Figure 1: Schematic illustration of the experimental protocol used in experiment 1: (a)

control task; (b) Task 1: look and point during JM; (c) Task 2: look during JM and point

after JM………………………………………………………………………………………47

Figure 2: Experiment 1: Mean pointing error of the hand. The figure illustrates the significant

difference between Task 2 and the other two conditions (Control and Task1). Error bars show

±1 standard error……………………………………………………………………………..48

Figure 3: Experiment 1: Average vergence-specified distance for near and far targets in all

the tasks. The targets were shown at a distance of 25 cm and 45 cm from the participant

which is shown by the dotted lines…………………………………………………………..49

Figure 4: Schematic representation of the experimental procedures used in

experiment 2………………………………………………………………………………....54

Figure 5: Mean proportion of ‘near’ responses for each comparison target location (at 0 both

targets were presented at the same location). Error bars show +1 standard error. ………….56

Figure 6: Differences in the POE between the Control condition and Tasks 1 and 2 for

individual participants (ID 1 to 20). The y-axis represents the difference in proportion of

‘near’ responses between Task 1 and Control & Task 2 and Control. Positive values indicate

xi

that the comparison target was reported as ‘nearer’ and negative values indicate that the

comparison target was reported as ‘farther’ with respect to the control task……………......57

Figure 7: Summary and interpretation of results for experiment 2……………………….…59

Paper 2

Figure 8: Schematic illustration of the experimental protocol used in experiment 1: (a)

control: no JM; (b) task 1: JM performed during saccade and pointing; (c) task 2: JM

performed during saccade, but not during pointing…………………………………………76

Figure 9: Distribution of pointing responses obtained in experiment 1. The boxplot contains

the middle 50% of the data (the upper edge is the 75th

percentile and the lower edge is the

25th

percentile), the line in the box represents the median. The lines extending from the

boxplot (whiskers) indicate the 1st and 99

th percentile………………………………………78

Figure 10: Schematic illustration of the experimental protocol used in experiment 2: (a)

control: no JM; (b) task 1: JM performed during the presentation of the standard target and

saccade; (c) task 2: JM performed during perceptual localization………………………….82

Figure 11: Mean proportion of ‘left’ responses for the five comparison targets (at 10 both

the standard and comparison target were presented at the same location). (a) standard target

presented in the left hemifield; (b) standard target presented in the right hemifield. Error bars

show +1 standard error of the mean………………………………………………………....84

xii

Paper 3

Figure 12: Schematic illustration of the experimental protocol used in experiment 1………99

Figure 13: Mean proportion responding ‘smaller’ for each of the five sizes of the comparison

square (at 0 both the standard and comparison squares were the same physical size, negative

values indicate that the comparison square was smaller). Error bars show +1 standard

error………………………………………………………………………………………...101

Figure 14: Mean perceived depth for stereoscopically presented stimuli in experiment 2.

Error bars show +1 standard error………………………………………………………….107

Figure 15: Mean values obtained when participants were asked to null the Pulfrich illusion by

adjusting the value of the variable filter. Error bars show +1 standard error………………112

Paper 4

Figure 16: Psychometric functions of the patients tested. The location of the comparison

target with respect to the standard is shown in cm with positive values indicating nearer and

negative values farther positions from the observer. Proportion of ‘near’ responses for each

comparison target location (at 0 both targets were presented at the same location)………..126

xiii

ABBREVIATIONS

Ia: primary afferent

IIa: secondary afferent

: gamma system

2-AFC: two alternative forced choice

APH: active pulley hypothesis

ANOVA: analysis of variance

CNS: central nervous system

EMG: electromyography

EOM: extraocular eye muscles

GTO: golgi tendon organs

IO: inferior oblique

IR: inferior rectus

JM: Jendrassik Maneuver

JND: just noticeable difference

LED: light emitting diodes

LR: lateral rectus

mesADC: mesencephalic area for dynamic control

MIF: multiply innervated fibers

MR: medial rectus

ND: neutral density (filter)

PE: palisade endings

POE: point of objective equality

PSE: point of subjective equality

PSI: presynaptic inhibition

xiv

NT: non-twitch

SIF: singly innervated fibers

SO: superior oblique

SR: superior rectus

SOP: superior oblique palsy

V1: primary visual area (striate cortex)

VOR: vestibulo-ocular reflex

xv

Glossary of Terms

Afference: detection of the activity produced by sensory receptors by the central nervous

system.

Binocular (or retinal) Disparity: the difference in spatial distribution of light in the two

retinal images, basis of stereopsis.

Cyclopean eye: point midway between the eyes which serves as a centre of reference for

headcentric directional judgements (origin of binocular vision) (Howard, 2002).

Depth constancy: observers perceive the depth of an object as relatively constant despite

changes in image size at different fixation distances.

Distal size: the real or physical size of an object.

Efference copy (corollary discharge): a copy of the motor signal generated to distinguish

the reafferent and exafferent signals (McClosky, 1981).

Esotropia: eye(s) deviate towards the nose.

Exafference: sensory activity produced by external factors.

Exotropia: eye(s) deviate temporally.

H-reflex (Hoffman reflex): an electrical counterpart of the tendon reflex elicited by

stimulating the mixed peripheral nerve, thus, activating the Ia afferent directly and bypassing

the activation of muscle spindles.

Horizontal Disparity: an interocular difference in the horizontal angle subtended by a pair

of points.

Inflow: afferent signals produced by the receptors in the muscles.

Jendrassik Maneuver (JM): a muscle contraction during which the amplitude of spinal and

brainstem reflexes is amplified.

Just noticeable difference (JND): is the smallest possible physical difference that can be

detected reliably and it reflects the precision of the judgment.

Myotomy: surgery on the extraocular eye muscles to correct strabismus which involves

weakening the muscle by reducing the number of contractile elements.

Outflow: copy of the motor command sent to structures within the central nervous system.

Phoria: the relative position of the eyes in the absence of binocular fusion.

xvi

Point of objective equality (POE): the point at which the comparison stimulus value

physically equals the value of the standard stimulus.

Point of subjective equality (PSE): the point at which the psychometric function yields a

probability of 0.5 (i.e. the comparison stimulus is perceived as smaller than the standard

stimulus 50% of the time) and it reflects the accuracy of the judgment.

Proximal size: the size (visual angle) of the image subtended on the retina.

Pulfrich illusion: a pendulum objectively swinging in the frontal plane appears to move in

an elliptical orbit in depth when one eye is viewing the stimulus through a neutral density

filter.

Reafference: sensory activity produced by organism’s own movement.

Recession: surgery on the extraocular eye muscles to correct strabismus which involves

weakening the muscle by changing the tangential force. It involves cutting the muscle tendon

close to insertion and reinserting it on the sclera posterior to the original insertion.

Registered eye position: internally generated estimate of the position of the eyes in the

orbits based on visual and non-visual (afferent and efferent) signals.

Resection: surgery on the extraocular eye muscles to correct strabismus to strengthen the

muscle. It involves shortening the length of the muscle by excising a portion of the tendon

and reinserting it to the original insertion.

Size constancy: observers perceive the size of an object as relatively constant despite

changes in retinal size at different fixation distances.

Stereoscopic vision: literally “solid sight”, visual perception of 3D structure of the world

from binocular vision (Howard, 2002).

Strabismus: misalignment of the eye(s).

Trigeminal-oculomotor synkineses: abnormal efferent innervation of the medial rectus by

the trigeminal nerve.

Vergence: disjunctive eye movements (eyes move through equal angles in opposite

directions) to fixate a target closer to the body (convergence) or farther away (divergence).

Vergence Specified Distance: fixation distance in the sagital plane specified by the vergence

angle of the eyes.

xvii

List of Appendices

Appendix 1 …………………………………………………………………………………157

1

CHAPTER I: LITERATURE REVIEW

1.1 How does the brain know which way the eye is pointing? Inflow vs. Outflow

Knowledge of eye position is critical for accurate visuomotor behavior. For instance,

to make an accurate reaching movement to pick-up an object, the central nervous system

(CNS) must combine several signals: the position of the hand and head, as well as the retinal

location of the object and the position of the eyes. The CNS can obtain eye position

information from two non-visual sources: inflow and outflow. Outflow refers to the efferent

motor command that is sent to the eye muscles, a copy of which is also sent to other

structures within the CNS to generate the expected sensory consequences of a given motor

command (efference copy or corollary discharge; for a detailed discussion, see Donaldson,

2000). Inflow refers to the signals that are sent from the eye muscle proprioceptors that

monitor eye position.

Although there is no question that the CNS must obtain eye position information, the

debate between outflow and inflow as the source of the signal has been longstanding and

goes back to Helmholtz and Sherrington (Bach-y-Rita, 1971. For an extensive historical

review see Grusser, 1995). Helmholtz was a proponent of the outflow theory. His support for

this theory was based on a classical experiment which showed that passive displacement of

the eye, by pressing on the outer canthi, results in perceived movement of the visual

surroundings whereas the same movement executed voluntarly does not lead to the illusion.

Helmholtz argued that if the passive movement of the eye changed the afferent signal without

changing the efferent innervation and if feedback is used by the CNS, there should be no

illusory motion of the visual environment. The fact that subjects reported illusory motion

during the experiment suggested that the afferent signals were not being used by the CNS.

2

Helmholtz's view was directly opposed by Sherrington, who was a proponent of the

‗muscle sense‘. Sherrington's support for the inflow theory was based on his own research:

the dicovery of muscle spindles in skeletal muscles and their role in the control of movement

via spinal feedback loops. Sherrington believed that there must be a feedback loop involving

proprioceptors from the eye muscles because they contain a large number of spindles and eye

movements are precisely controlled. Unfortunately, Sherrington never provided any

experimental evidence for the role of eye muscle proprioceptors in oculomotor control.

Helmholtz provided strong experimental evidence supporting the outflow theory.

Other authors looked for additional support using different methods. For instance, Siebeck

(1953, 1954 as described in Matin, 1976), Brindley, Goodwin, Kulikowksi and Leighton

(1976) and Stevens et al. (1976) used curare to paralyze the eye muscles. It was argued that if

outflow were the only source of eye position information, subjects should report illusory

movement of the visual scene when they attempted to move their eyes (i.e., if the eyes could

not move, the afferent signal would not change but the efferent signal would change). The

results of these studies did not fully support the outflow theory: illusory movement was

perceived only with partial parlysis but not with full paralysis, suggesting that the CNS did

use afferent feedback from eye muscles.

There have been two other major arguements against ocular proprioception. First,

unlike skeletal muscles, eye muscles act against a constant load. Consequently, a copy of the

motor command should theoretically provide sufficient information about the state of the

oculomotor plant. Evidence against this assumption was presented by Steinbach and Lerman

(1990) who demonstrated that in the case of the human eye, the centre of rotation is located

behind the center of mass suggesting that orbital proprioception might be important, as it is

3

for other skeletal muscls which are operating under changing loads, because the head‘s

position with respect to gravity changes. Second, an important function of proprioceptors in

the skeletal muscles is regulation of muscle length in response to stretch; however, the

presence of stretch reflexes in extraocular muscles (EOM) is controversial. A classic

experiment by Keller and Robinson (1971) in awake rhesus monkeys found no change in the

activity of the abducens motoneurons in response to muscle stretch. Similar results were

obtained in the oculomotor nucleus of cats by Tomlinson and Schwarz (1977). In contrast,

Dancause and colleagues (Dancause et al., 2007) have recently recorded electromyographic

activity in the horizontal recti muscles in lightly anesthesized rats and squirrel monekys in

response to passive eye rotation. This is the first study to suggest that stretch reflexes might

be present in the EOM but more studies are needed to confirm these findings.

Although the presence of monosynaptic stretch reflexes in the EOM have been

questioned, there is substantial evidence to support the inflow theory. First, neural activity in

response to passive stretch of the EOM has been recorded in many CNS structures: the

cerebellum, superior colliculus, lateral geniculate nucleus, and the primary visual cortex (for

a review see Donaldson, 2000). Second, highly trained observers whose eyes were moved

passively were able to report the correct direction of their eye movements in 70% of trials

(Skavenski, 1972).

In summary, the inflow theory has been neglected for many years but there is now

ample evidence to suggest that both inflow and outflow signals are used by the CNS. The

contribution of these signals might depend on the oculomotor task and the experimental

methodology. The next section will provide an overview of the role of EOM proprioception

during development, in oculomotor control, and in visuomotor behavior.

4

1.1.1 Role of inflow during development

Over the past 20 years many experiments have been conducted in an attempt to

understand the role of eye muscle proprioception in the development of visual function,

oculomotor control and visuomotor coordination using animal models. In animals,

deafferentation can be performed by sectioning the ophthalmic division of the trigeminal

nerve. Buisseret et al. (1995) conducted a series of studies examining the effects of

deafferentation on the development of binocular cells and orientation selective cells in visual

area V1, and reported that the development of these cells is dependent on sensory feedback

from the eye muscles. For example, complete deafferenation precludes the development of

orientation selective cells; specifically, eliminating proprioceptive input, via bilateral

deafferentation, from the horizontal EOM disrupts the development of vertically tuned cells.

The effect of deafferentation on functional deficits, such as depth perception, depends

on the developmental stage of the animal. For example, Graves, Trotter and Freganc (1987)

examined depth perception in kittens who were deafferented between 5-11 weeks of age and

those who were deafferented later, as adults. Depth perception deficits were only evident in

the animals who lost proprioception early during development and not in those who were

deafferented as adults. Overall, the deafferentation experiments suggest that during

development efference copy and visual experience do not provide sufficient information, and

that afferent feedback is critical for the optimal development of the oculomotor and visual

systems.

5

1.1.2 Role of inflow in oculomotor control

1.1.2.1 Evidence from animal studies

EOM afference has been implicated in maintaing eye stability in the dark. For

instance, Fiorentini and Maffei (1977) sectioned the ophthalmic branch of the trigeminal

nerve which resulted in pendular movements (only the horizontal component was measured)

of the ipsilesional eye when the animals were placed in an unstructured visual environment.

In addtion, cats showed asymmetric vestibular nystagmus after deafferentation which was

only present in the dark. Other studies have also found fixation instability and asymmetric

vestibulo-ocular reflex (VOR) in lambs (Pettorossi, Ferraresi, Draicchio, Errico, Santarelli &

Manni, 1995) and altered VOR gain in decerebrate pigeons (Hayman & Donaldson, 1995).

In contrast, several studies reported that deafferentation had no effect on the

programming and execution of saccades, smooth pursuit and VOR. Guthrie, Porter and

Sparks (1983) reported that efference copy provides sufficient information about eye position

for the programming of saccades. In their study, rhesus monkeys performed a double saccade

task: when the saccadic target was extinguished, the superior colliculus was micorstimulated

to change eye position and the animal had to make a compensatory saccade to the original

saccadic target location. Thus, to perform this task accurately, the animal had to take into

account the change in eye position due to stimulation. It was hypothesized that if inflow

contributes to eye position sense, the animal should not be able to perform this task after

deafferentation. Experimental data showed that the surgery did not affect performance, which

supported the outflow theory.

Lewis, Zee, Hayman, Tamargo (2001) also found that bilateral deafferntation did not

affect eye aligment, saccadic amplitude or conjugacy, saccadic adaptation, or the acceleration

6

and gain of smooth pursuit and VOR in rhesus monkeys. No differences in any of these

parameters were found in the five-week follow up. Similiarly, the accuracy and variability of

open-loop pointing was not affected significantly by the surgery (Lewis, Gaymard &

Tamargo, 1998).

An important insight into the role of afference in oculomotor control was provided by

Lewis and colleagues (Lewis, Zee, Gaymard & Guthrie, 1994; Lewis, Zee, Goldstein &

Guthrie, 1999). They conducted two studies which examined the effect of deafferentation on

eye alignment, saccade conjugacy, and post saccadic drift in rhesus monkeys. Prior to

deafferntation, the animals underwent surgery to weaken the inferior rectus muscle which

resulted in vertical muscle palsy and eye misalignment. Following the deafferenting surgery,

animals were tested over a time period of 3 weeks. Deafferentation resulted in a signficant

increase in eye misalignement (from 0.47 to 2.9), and eye conjugacy—indicated by the

pulse/pulse ratio—decreased by 10% (Lewis et al., 1994). Although there were no consistant

changes in post-saccadic drift, it was significantly altered (Lewis et al., 1999).

The critical difference between the studies that found that deafferentation affected

oculomotor control and those that found no difference is the fact that Lewis and colleagues

(1994, 1999) introduced a perturbation (surgery) to the oculomotor plant prior to

deafferentation. The surgery changed the properties of the oculomotor plant so that the

efferent commands were no longer appropriate. Lewis and colleagues proposed that sensory

feedback is necessary to inform the CNS about the peripheral changes in order to maintain

accurate occulomotor control. In contrast, no perturbations were applied to the oculomotor

plant in the studies that found no difference in oculomotor control following deafferentation,

which might in part explain the different results.

7

It is surprising that all the studies reviewed above examined versional eye movements

but negelcted vergence eye movements. Given that precise eye alignment is required in order

to maintain clear binocular vision, it could be hypothesized that deafferentation might have a

more significant effect on vergence than on version. This hypothesis was confirmed by

Guthrie, Porter and Sparks (1982) who reported that cutting the monkeys‘ ophthalmic branch

of the trigeminal nerve altered their vergence responses but had no effect on conjugate eye

movements.

A role for ocular proprioception in improving eye alignment after superior oblique

palsy (SOP) was recently proposed by Shan et al. (Shan et al., 2007). They sought to

determine the ‗ocular motor signature‘ of acquired SOP in rhesus monkeys in acute and

chronic stages. The operated eye was patched for 10 days after the surgery while changes in

vertical alignment were measured with the normal eye fixating. The authors were surprised to

find an improvement in the vertical alignment of the covered eyes during the period of

monocular viewing. They suggested that the CNS might be using a proprioceptive signal

from the operated, patched eye to achieve comitance.

1.1.2.2 Evidence from studies with humans

Passive eye deviation using a suction scleral contact lens has been used to examine

whether afferent feedback from eye muscles affects the programming and execution of

saccadic eye movements and smooth pursuit. Knox, Weir and Murphy (2000) reported that

saccadic amplitudes were decreased by approximately 20% and the amplitude-velocity

relationship did not change when the movement of the non-viewing eye was impeded. The

velocity of smooth pursuit was also significantly lower (~15%) when afferent feedback was

8

perturbed. Moreover, these effects were evident in all trials in the epoch 40-80 ms after the

initiation of the eye movement but not within the first 40 ms (Weir & Knox, 2001). These

data suggest that afferent feedback from the impeded eye can be used on-line during the

programming and excution of eye movements.

Insight into the role of EOM proprioception in visuomotor behavior was provided by

vanDonkelaar, Gauthier, Blouin and Vercher (1997). They used an experimental paradigm

which involved adaptation of the smooth pursuit system while the movement of one eye was

impeded via a suction scleral lens. During the adaptation procedure subjects tracked a target

moving at a constant velocity while a proportion of the recorded eye motion signal was

added to the target's motion. The adaptation procedure was performed when the movement of

the non-viewing eye was blocked and the gain of smooth pursuit was examined during the

adaptation as well as during pre- and post-tests. The target was presented for a short duration

(300 ms) during the post test so that smooth pursuit was executed with limited visual

feedback. Results from the study showed a signifcant difference (increase of 50%) in smooth

pursuit gain during the post-test between the control condition and the condition where

afferent feedback was altered during the adaptation. Interestingly, only a small difference in

gain (10%) was found between these condtions during the adaptation procedure. The authors

suggested that during the adaptation procedure the retinal motion signal was the same for

both conditions but this signal was associated with different afferent feedback. Therefore,

afferent information plays a role during adaptation when the relationship between efferent

and afferent signals must be modified.

9

1.1.3 Role of inflow in localization

Many studies have examined the role of inflow in accurate estimation of egocentric

distance. In the following section, studies involving manipulation of afferent feedback in

healthy subjects are discussed first, followed by studies involving patients with disrupted

proprioception.

In the past, two methods have been used to manipulate feedback from the eye

muscles: vibration stimuli applied over the muscle and passive deviation of the eye using a

suction lens. Vibration provides a good stimulus for activating the Ia afferent, which in

skeletal muscles activates the monosynaptic stretch reflex (i.e., contraction of the vibrated

muscle). Roll, Velay and Roll (1991) applied vibration over the inferior rectus muscle while

subjects were fixating a single light in the dark. During the vibration trials, subjects reported

that the target moved up and they also pointed above the target. Similar results were also

obtained by Velay, Roll, Lennerstrand and Roll (1994) where vibration of the right lateral

rectus muscle resulted in an illusory movement of the target to the left. Overall, results from

these studies suggest that vibration of an EOM muscle leads to a perception that the muscle is

lengthening, and participants report that the target is moving in a direction opposite to the

vibrated muscle.

The afferent signals from eye muscles can be altered by passively moving the eye

using a suction lens. This method, introduced by Gauthier, Nommay and Vercher (1990a,b),

offers a way of distinguishing the contributions of inflow and outflow to registered eye

position. The paradigm involves subjects fixating a target with one eye while the other eye is

occluded with a patch. In the first experimental condition, the fixating eye is deviated via a

suction lens, thus, in order to maintian fixation, the amount of innervation sent to the eye

10

muscles must be increased. Since both eyes receive the same amount of innervation during

conjugate eye movements (Hering's law), the occluded eye should deviate by an amount

corresponding to the efferent signal sent to the fixating eye. In this condition, the efferent

signal to the eye muscles must be increased to compensate for the perturbation, but the

afferent feedback from the fixating eye is not changing because the eye is not changing

position. Therefore, this task allows us to examine the effect of efference on registered eye

position. The second experimental condition involves passive deviation of the occluded eye.

In this case, the amount of innervation does not change, but the afferent feedback from the

deviated eye does change; therefore, this condition allows us to examine the contribution of

afference to regitstered eye position. Using the above paradigm, Gauthier et al. (1990a,b)

found that binocularly normal observers mislocalized the target in the direction of the eye

deviation when pointing with their hands or verbally reporting the straight-ahead direction. It

was estimated that the contribution of proprioception to knowledge of eye position was

approximately 32%.

A similiar study, using the eye-press technique instead of a suction lens to deviate the

eye, was conducted by Bridgman and Stark (1991). Results from the study were in

remarkable agreement with the data from Gauthier et al. (1990a,b) and the estimated

contribution of proprioception to eye position sense was 26% with a threshold of

approximately 10. In short, both studies reported greater perceptual localization and

pointing errors when the viewing eye was pressed, indicating that the CNS relied on the

efference to a greater extent but the contribution of afference could not be ignored.

Several studies have shown that patients with pathologies that disrupt proprioceptive

signals from eye muscles have deficits in localization of objects when tested using an open-

11

loop pointing task. For instance, patients with a pathology involving the trigeminal nerve

have increased pointing errors (Steinbach, 1986; Ventre-Dominey, Dominey & Sindou,

1996). Mislocalization errors are also evident immediately after surgery in patients with

strabismus who have undergone multiple surgeries to correct their eye alignment (Steinbach

& Smith, 1981). The Steinbach and Smith study provided a critical insight by identifying the

putative source of proprioceptive feedback. The surgical intervention for starbismus involved

cutting the muscle at the muscle-tendon junction which is where palisade endings (PE), the

putative proprioceptors, are also found. Since the localization deficit was not found in

patients after the first surgery, it is likely that each surgery destroyed the putative EOM

proprioceptors to a certain degree and after multiple surgeries patients were left deafferented.

Evidence highlighting the contribution of proprioception to knowledge of eye

position has also been provided by Lewis and Zee (1993) who studied spatial localization in

a patient with abnormal innervation of the medial rectus muscle. Due to trigeminal-

oculomotor synkineses the patient‘s left eye adducted every time the jaw moved to the right,

which presented a case in which the efferent command to the EOM was dissociated from the

afferent feedback. As hypothesized, the accuracy of open-loop pointing responses was

affected by the synkineses: the responses were shifted in the direction opposite to the rotation

of the affected eye, which is in contrast to what was reported by Gauthier et al. (1990a,b) and

Bridgeman and Stark (1991) (i.e., mislocalization was always in the direction of the deviated

eye). Lewis and Zee suggested that active muscle contraction and passive rotation might

activate the putatitve EOM proprioceptors differentially; specifically, they suggested that PE,

which are in series with the muscle fibres, will be activated to a greater extent during active

contraction compared to passive rotation of the eye.

12

1.1.4 Summary

There is ample evidence from animal and human studies to suggest that afferent

signals from the EOM are used by the CNS for optimal oculomotor control. In particular,

inflow is important during the critical period of development and during adaptation after

properties of the oculomotor plant change. Overall, the CNS relies on efference to a greater

extent, but the contribution of afference to registered eye position is also evident during

motor and perceptual tasks. In particular, while the contribution of inflow to versional eye

movements has been examined extensively, vergence has been largely ignored and should be

examined more carefully.

1.2 Extraocular muscles: anatomy, morphology and innervation

The concept of the ‗oculomotor plant‘ refers to the eye muscles, the motor nuclei and

the cranial nerves, and it is the final common pathway through which cortical centers control

different types of eye movements (Robinson, 1981). The oculomotor plant has been studied

in great detail throughout the years; however, the advent of novel imaging and tracing

techniques has led to an emergence of new concepts and theories of eye movement control.

The purpose of this section is to provide a brief overview of the six extraocular eye muscles

(EOM) focusing on their anatomy, morphology and innervation.

The primary and secondary actions of the EOM and their efferent innervation are

shown in Table 1. The pulling action of eye muscles corresponds to the plane of the

semicircular canals, presumably to simplify the control of eye movements during VOR. The

six EOM are arranged in synergistic pairs—for instance, to execute a conjugate eye

movement in the horizontal plane, the ipsilateral medial rectus (MR), and contralateral lateral

13

rectus (LR) would contract. According to Hering‘s law, synergistic muscles in the two eyes

receive equal innervation to simplify the control of conjugate eye movements.

Table 1: Primary and secondary actions of the EOM and their efferent innervation

EOM Primary action Secondary action Innervation

Medial Rectus Adduction __ Inferior division of

oculomotor nerve

Lateral Rectus Abduction

__ Abducens nerve

Superior Rectus Elevation Intorsion Superior division of

oculomotor nerve

Inferior Rectus Depression Extorsion Inferior division of

oculomotor nerve

Superior Oblique Introsion Depression

Trochlear nerve

Inferior Oblique Extorsion Elevation Inferior division of

oculomotor nerve

Muscle fibers in the EOM can be classified into several types based on their

innervation, morphological, histochemical and contractile properties. The following six fiber

types have been identified: singly and multiply-innervated orbital fibers, singly innervated

red and pale global fibers, multiply-innervated global fibers, and singly innervated global

intermediate fibers (Spencer & Porter, 1988). The classification of fibers as orbital and global

has been made based on their location so that global fibers are closest to the globe of the eye

and the orbital fibers are found near to the orbital wall. A recent study reported that the

global layer contains a larger number of muscle fibers and that the number of fibers among

the recti muscles in the global layer is constant. On the other hand, the number of fibers in

the orbital layer was found to vary up to 58% between the recti muscles, with the largest

number of fibers found in the MR (Oh, Poukens, & Demer, 2001).

14

The orbital layer is composed of approximately 80% singly innervated fibers (SIF),

and the remaining 20% are multiply innervated fibers (MIF). The SIF have a high oxidative

capacity due to numerous, prominent mitochondria and a substantial vascular supply, thus,

these fibers are resistant to fatigue. The MIF of the orbital layer exhibit different

histochemical and electrical properties along the length of the fiber. The mid-region of the

fiber has a larger number of mitochondria in comparison to the proximal and distal ends.

Electrical stimulation of the mid-region also leads to generation of action potentials; in

contrast, only slow, graded potentials can be recorded distal to the end-plate (Jacoby,

Chiarandini, & Stefani, 1989).

The global layer contains four types of SIF which differ in their metabolic capacity.

The red fibers contain a large number of mitochondria and show oxidative metabolism, the

pale fibers have few mitochondria and rely on anaerobic metabolism. The intermediate fibers

have both aerobic and anaerobic metabolic capacity. The global layer contains approximately

10% MIF, which respond to electrical stimulation with slow potentials that have a small

amplitude and long duration. Unlike the orbital MIF, these fibers respond uniformly when

stimulated along the length of the muscle. On average, the motor unit size of global fibers is

smaller than that of the orbital fibers, which suggests that the force can be increased with

greater precision by the global fibers (Oh et al., 2001).

Five of the EOM: MR, LR, superior rectus (SR), superior oblique (SO), and inferior

rectus (IR) originate in the tendinous ring (annulus of Zinn) at the apex of the orbit. The

origin of the inferior oblique (IO) is from the maxillary bone in the medial wall of the orbit

(Spencer & Porter, 1988). The muscles are composed of twenty to thirty thousand fibers and

their length varies between 40 and 50 mm (Carpenter, 1988). The fact that the SO muscle

15

inserts through a pulley, the trochlea, has been known for a quite a long time, but only recent

imaging studies have shown that the other EOM might also have a pulley insertion. Demer

and colleagues (Clark, Miller & Demer, 1997, 2000; Demer, Oh & Poukens, 2000; Demer,

2002) used high resolution magnetic resonance images and histology to demonstrate that the

orbital layers of the muscles insert onto a ring of collagen tissue located near the equator of

the globe in Tenon‘s fascia, which is now being considered as the functional origin of the eye

muscles. The discovery that eye muscles have a dual insertion, with the orbital layer inserting

on a pulley and the global layer inserting on the globe via a tendon, has functional

implications for the control of eye movements and led Demer to propose the active pulley

hypothesis (APH). The APH states that the contraction of fibers in the orbital layer

influences the position of the pulleys, and thereby affects the rotational axis of the EOM,

whereas contraction of the muscle fibers in the global layer directly affects the rotation of the

globe.

The eye muscles receive motor innervation from the motor nuclei in the midbrain and

pons via cranial nerves III (oculomotor), IV (trochlear), and VI (abducens). Two types of

efferent nerve endings are found on the EOM fibers: single, large end-plates (en plaque) and

multiple, small fiber endings (en grappe). Fibers that receive single innervation (i.e., SIF)

have regularly spaced fibrils, large number of sarcoplasmic reticulum, and a well developed

transverse tubule system, which allows these fibers to conduct fast action potentials. In

contrast, the MIF have poorly developed sarcoplasmic reticulum and do not generate action

potentials; instead they generate a prolonged, graded response when stimulated at thresholds

that are 3-6 times greater than the most excitable SIF (Eakins & Katz, 1972). Thus, the MIF

are sometimes referred to as non-twitch fibers.

16

Many studies have shown that the relationship between eye position and firing

frequency of the motoneurons is highly correlated (Carpenter, 1988). However, a study by

Mays and Porter (1984) reported that the relationship between eye position and firing rate is

also dependent on the type of eye movement. In their study, recordings were made from the

motoneurons in the abducens nucleus during conjugate adduction and during convergence.

Data showed that for a given eye position there was increased firing rate when convergence

was compared to conjugate adduction. Expanding on these results, Miller, Bockisch and

Pavlovski (2002) measured the oculorotary forces in the LR and MR muscles to test the

hypothesis that the force in LR should be higher in the converged state. In contrast to their

hypothesis, they found decreased force in both the LR and MR muscles. These results show

that the innervation of the EOM is much more complex than previously acknowledged, and it

is possible that the motor commands to the eye muscles differ during versional and vergence

eye movements.

1.2.1 Sensory receptors in extraocular muscles

The EOM contain several receptors that could potentially provide the CNS with eye

position information: muscle spindles, golgi tendon organs and palisade endings. Muscle

spindles, which are the primary proprioceptors in the skeletal muscles, have been found

predominately in the orbital layer of the EOM in several species: human, sheep, pig and

some primates, but not in other species, such as cat, rabbit, horse and mouse (Maier,

DeSantis & Eldred, 1974). Detailed histological studies of EOM spindles have shown that

they are different from the skeletal spindles. For example, Ruskell (1989) examined spindles

in the EOM of enucleated patients and reported that more than 50% of EOM spindles were

17

indistinguishable from extrafusal fibers as they were not enclosed in a capsule and did not

have a defined equatorial region. He also observed that nuclear bag fibers were virtually

absent, which was confirmed by Lukas, Aigner, Blumer, Heinzl and Mayr (1994) and

Blumer, Lukas, Aigner, Bittner, Baumartner and Mayr (1999). These authors concluded that

due to the morphological differences between spindles in the EOM and those found in the

skeletal muscles, it is unlikely that EOM spindles could provide adequate eye position

information.

Another type of receptor in skeletal muscles is the golgi tendon organ which is found

in series with the muscle fibers and signals muscle tension. Tendon organs have been found

in the ungulate but not in human EOM, therefore, their contribution to knowledge of eye

position cannot be not generalized across species (Ruskell, 1999).

Palisade endings (PE) are receptors unique to EOM, and they are associated with the

MIF of the global layer. They are sometimes referred to as innervated myotendinous

cylinders and have been found in the eye muscles of all species tested to date: cat, sheep, rat,

monkey and human (Alvarado-Mallart & Pinçon-Raymond, 1979; Blumer, Lukas, Wasicky

& Mayr, 1998; Buttner-Ennever, Horn, Scherberger & D'Ascanio, 2001; Eberhorn, Horn,

Eberhorn, Fischer, Boergen & Buttner-Ennever, 2005; Richmond, Johnston, Baker &

Steinbach, 1984). However, Bruenech and Ruskell (2000) did not find any PE in the cadaver

material from human infants. Anatomical studies show that the PE are enclosed in a capsule

at the distal end of the global MIF fiber. A thinly myelinated axon runs along the muscle

fiber and then loops back to enter the capsule as it bifurcates into several branches (Alvardo-

Mallart & Pinçon-Raymond, 1979; Richmond et al., 1984). It has been proposed that PE

18

might be activated by muscle contraction rather than muscle stretch due to their location at

the myotendious junction (Richmond et al., 1984).

Several studies provide morphological and histological evidence suggesting that PE

are sensory receptors which could provide the CNS with proprioceptive information about

eye position. Alvardo-Mallart & Pinçon-Raymond (1979) reported that PE in cats were

associated with the presence of clear vesicles common in other sensory endings, such as

Golgi tendon organs and muscle spindles. They also suggested that the presence of a capsule

is indicative of sensory receptors because the capsule, which is commonly found around

other receptors, protects them from external pressures. Retrograde and anterograde tracing

studies provide corroborating evidence to support the claim that PE are sensory receptors:

Porter and Spencer (1982) injected horseradish peroxide into EOM and neurons were labeled

in the trigeminal ganglion; similarly, Billig, Buisseret-Delmas & Buisseret (1997) reported

that PE were labeled when tracers were injected into the Gasserian (trigeminal) ganglion.

More recently, several studies have shown that in addition to the sensory endings, the

musculotendinous region in the human, rabbit, cat and rhesus monkey EOM also contains

motor endings (Blumer, Wasicky, Hoetzenecker, & Lukas, 2001; Konakci, Streicher,

Hoetzenecker, Blumer, Lukas & Blumer, 2005; Konakci, Streicher, Hoetzenecker, Haberl et

al., 2005; Lukas, Blumer, Denk, Baumgartner, Neuhuber & Mayr, 2000). The motor endings

were identified based on staining of the myoneuronal junction with bungarotoxin, which

labels post-synaptic acytocholinergic neurons. Upon microscopic examination, these authors

also found basal lamina, which is indicative of motor terminals. In contrast, at least one study

did not confirm the presence of motor terminals at the musculotendinous junction of the

19

EOM. Eberhorn, Horn, Fischer and Buttner-Ennever (2005) used bungarotoxin in the EOM

of rats and found no evidence of motor terminals on the PE.

Hertle and colleagues (Hertle, Chan, Galita, Maybodi & Craford, 2002) reported

another type of receptor in the EOM: the enthesial ending and suggested that it might have a

proprioceptive function. The enthesial ending is also associated with fibers of the global

layer, but it is found in the tendino-scleral region of the muscle. Hertle et al showed that the

enthesial area consisted of myelinated and unmyelinated nerve fibers with abundant vascular

supply in healthy control subjects, but neurovascular abnormalities were evident in case of

subjects with congenital nystagmus. The proprioceptive role of enthesial endings has not

been examined to the same extent as that of PE. More studies are needed to confirm the claim

of Hertle and colleagues.

1.2.2 Dual innervation of the extraocular muscles: proprioceptive hypothesis

Although the question whether PE have a sensory or motor function has not been yet

resolved, several authors have proposed the possibility that PE along with the MIF might

have a proprioceptive role in the control of eye movements. Robinson (1991) was the first to

use the term ‗inverted muscle spindle‘ to suggest that the non-twitch MIF fibers and the PE

might be comparable to the gamma ()-spindle system found in the skeletal muscles. This

hypothesis has been further extended by Buttner-Ennever and her colleagues based on their

neuroanatomical tracing studies (Buttner-Ennever, Horn, Graf & Ugolini, 2002), which

demonstrated that the SIF and MIF receive innervation from separate groups of ocular

motoneurons (Buttner-Ennever et al., 2001). The two groups of neurons were identified when

injections of horseradish peroxide were made at different sites of the EOM. Large

20

motoneurons were labeled when the midregion of the muscle fiber close to the end-plate was

injected (i.e., the injection targeted the SIF), whereas smaller motoneurons in a distinct

region around the periphery of the large motoneurons were labeled when the distal

musculotendinous region of the muscle was injected (i.e., the injection targeted the MIF).

These small motoneurons form a cap over the dorsal trochlear nucleus. They are found in the

medial half of the abducens nucleus, bilaterally around the midline of the oculomotor nucleus

to the inferior oblique and the superior rectus (referred to as S-group motoneurons), and at

the dorsal medial border of the oculomotor nucleus to the medial rectus and the inferior

rectus (referred to as C-group motoneurons). The motoneurons of the SIF and MIF identified

by retrograde tracing also exhibit different histochemical properties. For instance, the MIF

motoneurons do not contain nonphosphorylated neurofilaments, calcium binding protein

parvalbumin and perineuronal nets, all of which can be found in the motoneurons of the

twitch fibers (Eberhorn, Ardeleanu, Buttner-Ennever & Horn, 2005). It has been proposed

that parvalbumin and perineuronal nets are usually found in highly metabolically active

neurons, whereas modulatory neurons lack these markers (Bruckner, Brauer, Hartig et al.,

1993; Bruckner, Schutz, Hartig, Brauer, Paulke & Bigl, 1994). Thus, the histochemical

differences between the SIF and MIF motoneurons might reflect the differential involvement

of these neurons in the execution of eye movements.

A subsequent study has shown that the premotor input to the twitch and non-twitch

motoneurons also comes from different premotor areas (Wasicky, Horn & Buttner-Ennever,

2004; Ugolini et al., 2006). The non-twitch motoneurons receive monosynaptic input from

the vestibular (parvocellular) area and nucleus prepositus hypoglossi which are areas

associated with gaze-holding mechanisms, as well as the central mesencephalic reticular

21

formation and the supraoculomotor area which are involved in the programming of vergence

eye movements. In contrast, the twitch motoneurons receive input from classical premotor

saccadic regions and regions involved in programming of VOR, such as the paramedian

pontine reticular formation and the vestibular nuclei (magnocellular zone). In summary, the

discovery of dual innervation of eye muscles provides some support for Robinson‘s original

claim; specifically, that the non-twitch motoneurons of the global MIF might be involved in

regulating the baseline activity of PE, which would be analogous to the gamma-spindle

system in the skeletal muscles (Buttner-Ennever et al., 2002).

1.3 Role of gamma innervation in the skeletal system

Unlike EOM proprioception, the spindle-gamma system in skeletal muscles has been

studied extensively. At the beginning of the last century, Sherrington proposed that the

―muscular sense‖ plays a critical role in motor behavior. The source of feedback has been a

matter of debate for many years with several potential sources proposed, such as joint and

skin receptors, muscles, tendons and corollary discharge. Following the work of Matthews

and others, it was finally accepted in the 1970‘s that activity in muscle spindles is necessary

and sufficient to explain kinesthetic sensations (Matthews, 1982). A basic review the

anatomy and physiology of the muscle spindle-gamma system demonstrates the potential role

of the system in motor control, however, for a detailed review see Hulliger (1984).

1.3.1 Anatomy and physiology in skeletal muscles

Muscle spindles, sometimes referred to as intrafusal fibers, are embedded inside the

muscle belly in parallel with the extrafusal fibers. They are activated by muscle stretch as

22

they are ideally positioned to sense muscle length and changes in muscle length. Three types

of intrafusal fibers have been identified based on their morphology and response to stretch:

dynamic bag, static bag and nuclear chain fibers. Spindles are innervated by two types of

sensory afferents: primary (Ia) which innervate all three types of fibers, and secondary

afferents (IIa) which innervate only the static bag and chain fibers. Each spindle consists of

1–2 primary endings which signal muscle length and changes in muscle length (i.e., position

and velocity sensitivity) and 1–5 secondary endings which signal muscle length (i.e., position

sensitivity) (Gordon & Ghez, 1991).

Spindles in the skeletal muscles receive efferent innervation from gamma

motoneurons which are smaller in size and their axons have longer conduction velocity

compared to the alpha motoneurons. Activation of the gamma motoneurons does not result in

development of significant tension in the corresponding muscle. Two types of efferent

endings have been identified: dynamic and static, which are differentiated based on whether

they decrease or increase the dynamic response of Ia afferents (dynamic response is

calculated based on the firing frequency during the stretch in comparison to the firing rate

when steady length has been achieved). Dynamic gamma motoneurons (d) innervate

dynamic bag fibers and act on primary afferents; in contrast, static gamma motoneurons (s)

innervate static bag and chain fibers and act on both primary and secondary afferents. Each

gamma motoneuron may supply up to seven spindles in the same muscle and individual

spindles may receive innervation from up to 10 gamma neurons (Hulliger, 1984).

23

1.3.2 Function: alpha-gamma co-contraction in skeletal muscles

The complex organization of the spindle-fusimotor system has precluded any

definitive conclusions regarding its functional significance, nonetheless several explanations

have been proposed. The most commonly accepted function of the fusimotor system is its

role in spindle sensitization during muscle contraction. Due to the fact that spindles are

oriented in parallel with muscle fibers, their activity would decrease and actually cease

during muscle contractions if the intrafusal fibers did not have a mechanism that would allow

them to contract simultaneously with the extrafusal fibers. It has been shown that activity in

the gamma system allows the spindles to contract simultaneously with the muscle which

prevents the ‗slacking‘ of the spindles and allows them to remain sensitive during the

contraction. This mechanism has been referred to as alpha-gamma co-activation and is

further supported by the fact that there are motoneurons (termed beta efferents) which

innervate both the extrafusal and intrafusal fibers (Gordon & Ghez, 1991).

1.3.3 Descending control: implications for function

There is no doubt that alpha-gamma co-activation is one of the important functions of

the gamma motoneurons and it is the only function in amphibians and lower vertebrates;

however, there is also evidence that fusimotor activity can be adjusted independently of alpha

motor output. The work of Granit and Kaada (1952), which was based on activation recorded

in muscle spindles and muscles in cats after electrically stimulating structures within the

CNS, showed that stimulation of pontile and mesencephalic tegmentum, dorsal

hypothalamus, caudate and motor cortex increased the firing rate of muscle spindles without

a corresponding muscular contraction. They reported two differences between activation of

24

alpha and gamma efferents in the anesthetized animals: 1) gamma activity could be elicited at

significantly lower levels of stimulation compared to that required for alpha activity, and 2)

stimulation of the brainstem reticular formation led to a facilitation of gamma activity that

lasted up to 30 sec after the stimulus was taken away, in contrast to the response seen after

stimulation of the primary motor cortex when the facilitation was only seen during

stimulation. The authors proposed that gamma motoneurons might be involved in the

regulation of tonic muscular activity because they tend to maintain tonic discharge.

A series of studies by Appelberg and colleagues (Appelberg, 1981; Appelberg &

Jeneskog, 1972) identified a region within the midbrain in proximity to the red nucleus,

which, upon stimulation, selectively increased activity in the d motoneurons. To reflect the

functional significance, the area was referred to as the mesencephalic area for dynamic

control (mesADC). Additional stimulation studies of the mesADC were conducted by Taylor

and Donga (1989) while simultaneously recording from first order neurons of the jaw

muscles in the mesencephalic trigeminal nucleus and from the dorsal root ganglia containing

sensory neurons from the medial gastrocnemius muscle. Their results showed that, depending

on the area of stimulation, dynamic or static activity to either muscle was affected without

concomitant muscle activation demonstrating separate control over gamma and alpha

motoneurons. The area stimulated included the region dorsal and caudal to the red nucleus

extending to fasciculus retroflexus which projects to the interpeduncular nuclei. This area in

turn has pathways connecting with the habenular nuclei which have extensive connections

with the limbic system, thus suggesting a possible function for the gamma system in arousal

in preparation for a motor act. Other studies have also shown that the gamma system might

have an important role when animals are performing tasks which require precision (Tanji,

25

1976) or novel motor tasks (Hulliger, 1993). It is possible that gamma activity is increased in

these tasks due to increased attention and vigilance in anticipation of the upcoming

difficult/novel motor task.

Since activity of the static and dynamic fusimotor systems can be regulated

independently by descending pathways, several authors have proposed that the gamma

system might be ideally suited for parametric control of sensory feedback. Specifically,

dynamic gamma neurons increase the dynamic sensitivity of primary afferents and static

gamma efferents increase spindle bias but reduce dynamic sensitivity; hence, the CNS can

selectively modulate spindle sensitivity to different parameters (velocity or position) during

movement via the gamma system. This hypothesis was investigated in two studies which

examined cyclical jaw movements. First, Appenteng, Morimoto and Taylor (1980) recorded

activity in the masseter motor nerve during jaw movements induced by intra-oral stimulation

(water in mouth). Two types of units were identified: one unit‘s activation was cyclical

together with the activation of alpha motoneuron (modulated activity). The other unit

increased firing at the beginning of the movement and maintained activity throughout the

movement (sustained activity). They proposed that the sustained activity was due to

activation of the dynamic gamma efferents which was set at the beginning of the movement

to determine spindle‘s sensitivity to stretch while the modulated activity was due to the

activation of the static efferents which was set in concert with the alpha motoneuron‘s

activation to make sure that spindles don‘t slack and fall silent during muscle contraction.

These results were replicated and extended by Gottlieb and Taylor (1983), who

simultaneously recorded from the gamma efferents. Overall, these studies highlight a major

26

function of the gamma system: the CNS can set the overall gain of afferent feedback

according to the demands of the motor task.

1.3.4 Summary

Results from a large number of studies suggest that the spindle-gamma system

provides a sophisticated method for the CNS to control sensory transmission. For any given

motor task the CNS can set a particular template for (co)activation of alpha and gamma

motoneurons, thereby creating the appropriate muscle tone and receptor sensitivity for

individual motor tasks. In short, the parametric control of sensory feedback would allow

anticipatory gain control of sensory transmission (Prochazka, 1989). Although the eye

muscles are structurally and morphologically different than skeletal muscles, the principle of

feedback gain control might also be relevant. The six EOM participate in a variety of eye

movements that range from very fast saccades (peak velocity up to 600°/sec) to relatively

slow vergence (peak velocity rarely exceeding 10°/sec) (Hallet, 1986). Previous studies have

shown that manipulations of afferent feedback affect vergence more than version eye

movements (Guthrie et al, 1982), thus, it is possible that the CNS sets the gain of afferent

feedback differently for vergence and saccades.

1.4 Jendrassik Maneuver: possible mechanisms

The Jendrassik Maneuver (JM) refers to a forceful, voluntary muscle contraction in

distant muscle groups which has been shown to facilitate the amplitude of tendon reflexes. It

was first reported by the Hungarian physician Erno Jendrassik (Pasztor, 2004). Although the

reflex reinforcement effect has been reported many times, there is controversy regarding the

27

neurophysiological substrates contributing to the effect. The monosynaptic tendon reflex

involves only two neurons: the Ia afferent and the alpha motoneuron, but there are several

mechanisms that could be involved in reflex facilitation. First, the effect could be mediated

via the gamma system: increased activity of the gamma motoneurons would increase the

sensitivity of the muscle spindles to stretch which would result in a greater afferent volley

when a stimulus is applied. Second, the facilitation effect could be mediated via increased

excitability of the alpha motoneuron or via reduction of presynaptic inhibition of the Ia

afferent. In addition, there could be oligosynaptic facilitation via interneurons in the spinal

cord contributing to the reflex reinforcement effect (Dowman & Wolpaw, 1988; Gregory,

Wood & Proske, 2001; Murthy, 1978; Zehr & Stein, 1999).

1.4.1 Gamma system

Initial studies supported the hypothesis that gamma activity is the predominant

mechanism. This conclusion was based on the results from studies that found an increased

amplitude of tendon reflexes but not H-reflexes [Buller & Dornhorst, 1957; Sommer, 1940

(as described in Zehr & Stein, 1999)]. Since the H-reflex is elicited by electrical stimulation

of the Ia afferent, it bypasses the activation of the receptors (muscle spindles), thus, the lack

of facilitation of the H-reflex would support the gamma hypothesis. In addition, at least two

studies using microneurographic recordings reported that spindle activity increased during

the reinforcement maneuver without concomitant increase in muscle activation (Burg,

Szumski, Struppler & Velho, 1973; Ribot-Ciscar, Rossi-Durand & Roll, 2000).

In direct contrast to the above studies, there are several reports that do not support the

gamma hypothesis. For instance, two studies examined whether the H-reflex is facilitated

28

following a fusimotor block induced using lidocaine (Clare & Landau, 1964) or a pressure

cuff (Bussel, Morin & Pierrot-Deseilligny, 1978). Both studies reported a comparable

increase in tendon and H-reflexes during the reinforcement maneuver when the response

from spindles was blocked. In addition, several studies found no difference in spindle activity

independent of muscle activity as measured by microneurography during the reinforcement

maneuver (Burke, McKeon & Skuse, 1981; Burke, McKeon & Westerman, 1980; Hagbarth,

Wallin, Burke & Lofstedt, 1975).

The controversy regarding the contribution of fusimotor system to reflex

reinforcement is partly due to the lack of a standardized methodology. For instance, studies

have used a variety of reinforcement maneuvers in terms of muscle groups, strength of

contraction and timing when the reflex was elicited. In addition, reflexes are affected by

many other variables, such as the posture of the subject, the state of the muscle during testing

as well as prior to testing (for a review, see Proske, Morgan & Gregory, 1993).

1.4.2 Alpha motoneuron excitability

Another mechanism that has been proposed to explain the reflex reinforcement effect

is the change in alpha motoneuron excitability which was examined by Dowman and

Wolpaw (1988). Electromyography (EMG) was used to record muscle activity while the

soleus H-reflex was elicited by electrical stimulation in two conditions: with JM or without

JM. It was hypothesized that if the JM modifies motoneuron excitability then the EMG

activity should follow the time course of the JM (i.e., there should be an increase in EMG

activity during the JM). Experimental results did not confirm the hypothesis, thus, it is

unlikely that the JM affects the excitability of the alpha motoneurons directly.

29

1.4.3 Presynaptic disinhibition

Another potential mechanism that could mediate the reflex reinforcement effect (i.e.,

JM) is a change in the afferent input to the motoneurons via reduction of presynaptic

inhibition (PSI) of the primary afferents. This hypothesis was tested by Zehr and Stein (1999)

who took advantage of the fact that stimulation of an antagonist nerve increases PSI. The

amplitude of the H-reflex was tested in 3 conditions: 1) while participants performed the JM

(i.e., presumably decreasing PSI), 2) when the antagonist nerve was stimulated (i.e.,

increasing PSI), and 3) when the two experimental conditions were combined. As expected,

the reflex was facilitated during the JM and reduced during nerve stimulation; however,

instead of a complete cancellation during the combined stimulation, data showed a slight

decrease in reflex amplitude. Overall, these results supported the view that one of the

mechanisms involved in the JM might be a reduction in PSI through interneurons in the

spinal pathways.

An elegant experiment that examined the role of PSI and the gamma system was

designed by Gregory et al. (2001). The PSI hypothesis was tested by eliciting the soleus H-

reflex during the JM while electrically stimulating the afferents from heteronymous muscle

(the quadriceps). It was hypothesized that if the JM operates through the PSI mechanism,

combining the two manipulations should result in an additive reduction of the PSI, which

was not confirmed by the results from the study. To test the contribution of fusimotor

activation to reflex reinforcement, the authors examined tendon reflexes after the muscle

conditioning which left the spindles either in a mechanically sensitive state or in a relatively

insensitive state (slack). It was hypothesized that support for the role of gamma activation

could be demonstrated if the JM facilitated reflexes in the condition when spindles were left

30

in a relatively insensitive state after muscle conditioning. Results showed that the JM did not

preferentially facilitate the reflex in the condition when spindles were in a slack state;

instead, reflexes in both conditions were significantly enhanced. The authors concluded that

the JM most likely operates through oligosynaptic spinal pathways which their study was not

able to elucidate.

1.4.4 Time course of the reflex facilitation effect

In an attempt to resolve some of the controversies surrounding the reflex

reinforcement effect due to the JM, Delwaide and Toulouse (1981) conducted a detailed

examination of the time course of the reflex facilitation effect and the factors that affect the

amplitude of the reflex. When the reinforcement maneuver was maintained for 2.5 sec, three

distinct phases could be identified which were affected differently by the intensity of the JM.

Facilitation of the reflex during phase I (150 ms following the signal to contract) was minor

and occurred before any electromyographic (EMG) activity in the muscle involved in the

reinforcement maneuver was detected. The peak facilitation occurred around 300 ms

following the command to perform the maneuver, and the peak was highly dependent of the

intensity of the remote contraction. Facilitation then declined steadily, reached a plateau

around 600 ms and was maintained as long as the muscle contraction was performed.

The factors that were investigated were the strength and type of reinforcement

contraction performed. Results clearly showed that the facilitation effect (i.e., the amplitude

of the reflex) was dependent on the intensity of the voluntary contraction: the higher the

intensity of the contraction, the greater the facilitation effect. In addition, two types of

contractions were examined: fast ballistic movements and slow isometric contractions. The

31

facilitation effect was comparable in phase I and II, but in phase III the effect was only

evident during the isometric contraction. Additional experiments were conducted to

determine the effect of a motor block and vibratory stimulus on reflex facilitation. As

expected, only phase I facilitation was present when the muscle was paralyzed, and only

phase II and III were present when vibration was applied to the muscle (instead of the

voluntary reinforcement maneuver). Based on these results, the authors suggested that

different neurophysiological mechanisms might be involved in the three phases of the

reinforcement maneuver. It was suggested that facilitation seen in phase I is most likely of

supraspinal origin and might be related to the descending motor command and increased

arousal before the onset of movement. Facilitation in phase II and III is only present when

there is afferent feedback from the contracting muscle; however, there is a clear difference

between these two phases, which might be related to the fact that the H-reflex is only

facilitated in phase II and not in phase III (Toulouse & Delwaide, 1980).

1.4.5 Summary

Studies have shown that the neurophysiological mechanism underlying the JM

facilitation effect is complex and most likely involves oligosynaptic spinal pathways and

descending motor pathways that change the excitability of the gamma motoneurons and

primary afferents. Research has shown that JM is a potent manipulation that changes the

excitability of spinal and brainstem reflexes and it does not matter which peripheral muscles

are used to achieve the reinforcing effect. Despite the clear delineation of the three phases

provided by Delwaide and Toulouse (1981), studies by Zehr and Stein (1999) and Gregory et

al (2001) which tried to investigate the mechanism underlying the JM did not try to

32

distinguish these phases or whether a different mechanism might be involved in each phase.

The degree to which decreased PSI and gamma modulation are involved in changing the

excitability of the afferents might be different in each of the phases. Specifically, the gamma

system is more likely to be involved in the facilitation effect in phase III when only tendon

reflexes are facilitated by the JM and the H-reflexes are not affected.

33

CHAPTER II: RESEARCH OBJECTIVES & HYPOTHESES

Despite the surge of interest in EOM proprioception in the last few years, the role of

afferent signals in oculomotor motor control and visuomotor behavior are not yet fully

understood. The aim of this research is to help to elucidate the role of afferent and efferent

signals from the EOM in oculomotor control. Seminal studies by Buttner-Ennever and Horn

(2002), which demonstrated that the singly and multiply innervated fibers receive innervation

from separate groups of ocular motoneurons, have theoretical implications for the control of

eye movements. Specifically, it has been hypothesized that the non-twitch motoneurons

associated with MIF of the global layer might control the gain of proprioceptive feedback

from PE, analogous to the gamma-spindle system found in the skeletal muscles. The goal of

the present study was to test the above hypothesis using behavioral and psychophysical

approaches. The activity of non-twitch motoneurons was altered using the Jendrassik

Maneuver (JM), which was assumed to alter the excitability of the gamma system.

The non-twitch motoneurons receive direct pre-motor input from areas that are known to

be involved in the control of vergence eye movements. Therefore, the first paper examined

manual and perceptual responses while participants executed vergence eye movements and

localized a target presented in the median plane. It was hypothesized that if the non-twitch

motoneurons are analogous to the gamma motoneurons, the JM would also increase the

activity of these neurons and alter the afferent feedback from PE, which would result in

misregistered eye position and localization errors.

Experiments presented in the second paper further extend and clarify findings from the

first paper by examining whether the JM affects registered eye position during localization in

the frontal plane. Since the non-twitch motoneurons do not receive direct premotor input

34

from areas involved in the programming of saccades, we hypothesized that localization

responses associated with the saccadic system would not be affected by the JM.

Accurate estimation of egocenteric distance is not only critical for the performance of

localization tasks, but perceptual constancies, such as size and depth also rely on accurate

registration of absolute distance. The third paper examined whether the JM perturbation

affects perceptual judgments that rely on accurate registration of absolute distance. The

following hypotheses were tested in 3 studies: 1. participants will perceive the size of a

constant retinal stimulus as larger when the feedback from the eye muscles is altered via the

JM; 2. for a given disparity, the perceived depth will be greater when the JM is performed

compared to the condition without JM; 3. for a constant stimulus the perceived depth will be

reported as greater while participants perform the JM.

The fourth paper examined whether patients who have been operated for strabismus are

susceptible to a manipulation of afferent feedback via JM while localizing a target in depth. It

was hypothesized that patients’ responses would not be affected by the JM perturbation

because the surgeries most likely compromised the EOM feedback loops, thus, the gain of

the afferent feedback would not be altered.

35

PAPER 1

PROPRIOCEPTIVE ROLE FOR PALISADE ENDINGS

IN EXTRAOCULAR MUSCLES:

EVIDENCE FROM THE JENDRASSIK MANEUVER

1E. Niechwiej-Szwedo, E. González, S. Bega, M.C. Verrier, A. Wong, M.J. Steinbach

Vision Research (2006); 46:2268-2279

1 See Appendix 1 for contributions of each author

36

ABSTRACT

A proprioceptive hypothesis for the control of eye movements has been recently proposed

based on neuroanatomical tracing studies. It has been suggested that the non-twitch

motoneurons could be involved in modulating the gain of sensory feedback from the eye

muscles analogous to the gamma () motoneurons which control the gain of

proprioceptive feedback in skeletal muscles. We conducted behavioral and

psychophysical experiments to test the above hypothesis using the Jendrassik Maneuver

(JM) to alter the activity of motoneurons. It was hypothesized that the JM would alter

the proprioceptive feedback from the eye muscles which would result in misregistration

of eye position and mislocalization of targets. In the first experiment, vergence eye

movements and pointing responses were examined. Data showed that the JM affected the

localization responses but not the actual eye position. Perceptual judgments were tested

in the second experiment, and the results showed that targets were perceived as farther

when the afferent feedback was altered by the JM. Overall, the results from the two

experiments showed that eye position was perceived as more divergent with the JM, but

the actual eye movements were not affected. We tested this further in experiment 3 by

examining the effect of JM on the amplitude and velocity of saccadic eye movements. As

expected, there were no significant differences in saccadic parameters between the

control and experimental conditions. Overall, the present study provides novel insight

into the mechanism which may be involved in the use of sensory feedback from the eye

muscles. Data from the two first experiments support the hypothesis that the JM alters the

registered eye position, as evidenced by the localization errors. We propose that the

altered eye position signal is due to the effect of the JM which changes the gain of the

37

sensory feedback from the eye muscles, possibly via the activity of non-twitch

motoneurons.

38

1.0 Introduction

Knowledge of eye position is critical for accurate visuomotor behavior. For

instance, to make an accurate reaching movement to pick up an object, the central

nervous system (CNS) must combine several signals including the initial hand position,

head position, eye position, and retinal location of the object. The CNS can obtain eye

position information from two non-visual sources: the efference copy of the motor

command sent to the eye muscles (outflow) and from the eye muscle proprioceptors

(inflow) (Steinbach, 1987). The debate between outflow and inflow theories goes back to

Helmholtz and Sherrington (Bach-y-Rita, 1971), but during the last twenty years ample

studies have provided evidence suggesting that the afferent signals from the extraocular

muscles (EOM) are used during egocentric localization tasks (Bridgeman & Stark, 1991;

Gauthier et al. 1990a,b; Roll et al., 1991; Velay et al., 1994), programming of eye

movements (Knox et al., 2000; Weir & Knox, 2001), and during adaptation of smooth

pursuit (vonDonkelaar et al., 1997). In addition, patients show pointing errors when the

proprioceptive signals from the eye muscles are disrupted, for example, after surgical

interventions that destroy proprioception (Steinbach, Kirshner & Arstikaitis, 1987;

Steinbach & Smith, 1981) or due to pathology involving the trigeminal nerve (Campos et

al.,1989; Ventre-Dominey et al., 1996).

Although it is now recognized that proprioception contributes to registered eye

position, the mechanism of proprioceptive feedback from EOM has not been established.

There are two potential receptors in the human eye muscles that could provide

proprioceptive information: muscle spindles and palisade endings (PEs). Muscle spindles,

which are the primary proprioceptors in the skeletal muscles, have been found in the

39

EOM of several species: human, sheep, pig, and some primates, but not in other species,

such as cat, rabbit, horse, or mouse (Maier et al., 1974). Detailed histological studies of

muscle spindles in the human eye muscles have shown that they are different from the

skeletal spindles. Ruskell (1989) reported that more than 50% of EOM spindles were

indistinguishable from extrafusal fibers as they were not enclosed in a capsule and did not

have a defined equatorial region. He also observed that nuclear bag fibers were virtually

absent, which was also confirmed by others (Blumer et al., 1999; Lukas et al., 1994).

Although the morphological differences between spindles in the EOM and those found in

the skeletal muscles are well documented, the specific function of EOM spindles has not

been established. Thus, it cannot be concluded at the present time whether EOM spindles

can provide adequate proprioceptive signals informing the CNS about changing eye

position.

Another putative source of proprioception from the eye muscles are PEs, which

are receptors that are unique to EOM. PEs are associated with the multiply innervated

fibers (MIFs) of the global layer and they are sometimes referred to as innervated

myotendinous cylinders (Ruskell, 1978). PEs have been found in the EOM of many

species, such as cat, rhesus monkey, sheep, rat, and human (Alvarado-Mallart & Pinçon-

Raymond, 1979; Blumer et al., 1998; Buttner-Ennever et al., 2001; Ebhorn et al., 2005;

Richmond et al., 1984). Anatomical studies show that the PEs are enclosed in a capsule at

the distal end of the MIFs. A thinly-myelinated axon runs along the muscle fiber and then

loops back to enter the capsule as it divides into several branches and makes contact with

the tendon and muscle fibers (Alvarado-Mallart & Pinçon-Raymond, 1979; Richmond et

al., 1984).

40

Although the location of the cell body of PE’s has not been established, several

studies provide morphological and histological evidence suggesting that PEs are among

the sensory receptors which provide the central nervous system (CNS) with

proprioceptive information about eye position. Alvarado-Mallart & Pinçon-Raymond

(1979) reported that PEs in the cat are associated with the presence of clear vesicles

which are common in other sensory endings, such as Golgi tendon organs (GTO) and

muscle spindles. Billig and colleagues (Billig et al., 1997) reported that PEs were labeled

when retrograde tracers were injected into the Gasser’s (trigeminal) ganglion, which

contains only sensory neurons. However, recent histochemical examination of the

musculotendinous junction shows that, in addition to the sensory endings, the

myoneuronal region also contains motor endings (Lukas et al., 2000). These motor

endings were identified based on staining of the myoneuronal junction with

bungarotoxin, which labels acytocholinergic receptors. Lukas and colleagues concluded

that PEs might receive dual, sensory-motor innervation, similar to that found in the

muscle spindles, which are sensory receptors innervated by motoneurons.

In line with the work of Lukas and colleagues (2000), recent anatomical tracing

studies by Buttner-Ennever et al. (2001) demonstrated that the EOM receive dual

innervation from two distinct groups of ocular motoneurons. The EOM of the global

layer can be classified into singly and multiply innervated fibers based on the pattern of

innervation they receive. The singly innervated fibers (SIFs) have a single end-plate zone

located in the midregion of the muscle and respond with fast propagating action

potentials when stimulated, thereby contributing to the force developed by the muscle. In

contrast, the MIFs have multiple end plates distributed along the fiber which are

41

concentrated at the distal end (this is also the region where PEs are found). Upon

electrical stimulation, the MIFs respond with slow graded potentials and do not

contribute to the force developed by the muscle (Fuchs & Luschei, 1971). Due to these

properties, the SIFs are referred to as twitch fibers, whereas the MIFs are referred to as

non-twitch fibers (Buttner-Ennever et al., 2001). When injections of horseradish peroxide

were made at the distal or the midregion of the EOM, two groups of neurons were

identified. Large motoneurons were labeled when the midregion of the muscle fiber close

to the end-plate was injected, whereas smaller motoneurons, in a distinct region around

the periphery of the large motoneurons, were labeled when the distal musculotendinous

region of the muscle was injected. Based on these results, it was concluded that the large

motoneurons innervate the twitch fibers (SIFs), and the smaller motoneurons innervate

the non-twitch fibers (MIFs). Further work has also shown that the twitch and non-twitch

motoneurons receive different premotor input, which sheds light on a possible role of

these fibers in oculomotor control (Wasicky et al., 2004). For instance, the twitch

motoneurons receive projections from the areas within the brainstem that are involved in

the programming of fast eye movments, such as saccades and the vestibulocular reflex.

The non-twitch motoneurons receive pre-motor input from areas that are known to be

involved in gaze-holding mechanisms, vergence eye movements and smooth pursuit.

The role of MIFs and PEs in the control of eye movements remains elusive.

Several authors have proposed that the PEs and MIFs might have a proprioceptive role in

the control of eye movements (Buttner-Ennever et al., 2002; Porter, Baker, Ragusa &

Brueckner, 1995; Robinson, 1991). In particular, the non-twitch motoneurons of the MIF

in the global layer could be involved in modulating the gain of sensory feedback from the

42

PEs, analogous to the motoneurons which control the sensitivity of muscle spindles in

skeletal muscles.

We took a behavioral approach to examine whether the gain of sensory feedback

from the EOM can be altered by a manipulation that affects the activity of the

motoneurons in skeletal muscles. The Jendrassik Maneuver (JM) is an isometric

voluntary contraction of any muscle group. JM is referred to as a reflex reinforcing

maneuver because the amplitudes of skeletal reflexes are facilitated while the JM is

performed (Delwaide & Toulouse, 1981; Murthy, 1978). One of the mechanisms

proposed to explain the reflex reinforcement effect is that the muscle contraction has a

general effect that results in up-regulation of the motoneuron activity which increases

the baseline activity of muscle spindles and, consequently, results in a larger efferent

response when the muscle is stretched.

Stretch reflexes have not been recorded in the EOM muscles (Keller & Robinson,

1971); however, neural responses to EOM stretch have been recorded in several cortical

regions (Donaldson, 2000). The role of proprioception in the control of eye movements is

most likely different than in the control of limb position and movement but the possibility

that proprioceptive feedback might be modulated by the activity of non-twitch

motoneurons should not be dismissed, particularly in light of the new findings that reveal

dual innervation of the EOM from the twitch and non-twitch motoneurons. We

hypothesised that if the non-twitch motoneurons are analogous to the motoneurons, the

JM should also change the activity of these neurons which would alter the afferent

feedback from PEs and result in misregistration of eye position and pointing errors.

Furthermore, if the JM affects the activity of the non-twitch motoneurons, the actual eye

43

position should not be different between the conditions because the non-twitch

motoneurons do not add to the force used to move the eyes (Fuchs & Luschei, 1971).

It has been reported that the non-twitch motoneurons receive monosynaptic input

from the pre-motor centers located in caudal mesencephalic reticular formation and the

supraoculomotor area, which are involved in the control of vergence eye movements

(Wasicky et al., 2004). Therefore, localization responses were examined while

participants performed vergence eye movements in the first 2 experiments. Saccadic eye

movements were examined in experiment 3 which served as a control because non-twitch

motoneurons do not receive direct premotor input from areas involved in programming of

saccadic eye movements.

2.0 Experiment 1

2.1 Method

2.1.1 Observers

Participants in all three studies had normal or corrected-to-normal visual acuity of

20/20 and stereopsis of at least 40 seconds of arc as measured with the Titmus test

(Titmus Optical Co., Inc., Petersburg, Virginia 23805). All experimental protocols were

approved by the Ethics Review Boards at the University of Toronto and the University

Health Network. All participants gave their informed consent prior to participating. Ten

healthy adults with no history of any ocular disorders, mean age 30.8±7.2 years,

participated in the first experiment.

2.1.2 Stimuli

44

The stimuli were 2 green light emitting diodes (LEDs) embedded in a custom-

made black board and controlled by the experimenter via a trigger box. The stimuli were

in an earth-horizontal plane and aligned with the participant's midline, slightly below eye

level, and the viewing distance was 25 cm to the near target, and 45 cm to the far target.

2.1.3 Apparatus

Horizontal and vertical position of both eyes was monitored and recorded using

an infra-red eye-tracker system (El-Mar series 2020, Toronto, Ontario, Canada). The

horizontal and vertical eye positions were obtained from the relative positions of multiple

corneal reflections and center of pupil. The system accuracy is 0.5° with a linear visual

range of ±40° horizontally and ±30° vertically. The system is free from drift and has a

resolution of 0.1°. Eye position data were sampled at 120 Hz and stored on a computer

for further analysis. Prior to data collection, the eye tracker was calibrated. The

calibration procedure involved fixating fourteen points displayed along the horizontal and

vertical axes (7 fixation points along each axis), separated by 3.3° visual angle. The

participant’s head was stabilized using a chin rest and adjusted so that the eyes were in

the central position when looking at the center of the array.

Arm movement data were recorded at 60 Hz using an electromagnetic device

(Flock of Birds, Ascension Technology Co, Burlington, Vermont, USA). The resolution

of the system is 0.5 mm. The receiver was placed on the participant's thumb of the

dominant hand, which was used for pointing. The calibration involved passively placing

the participant’s thumb at the targets’ location, which was performed at the end of the

experimental session in order to avoid any bias or learning effect.

45

JM involved an isometric, voluntary muscle contraction which was performed

with the abductor muscles of the legs against resistance. The device used for resistance

was a Thigh Master™. Participants were asked to perform each contraction at a 75%

level of their maximal voluntary contraction, which was determined prior to the initiation

of the experiment. To ensure that the isometric contraction was performed at a consistent

level throughout the experiment, a string was tied around the Thigh Master™ which was

pulled taut when the muscle contraction was executed. Participants were instructed to

hold the string taut when performing the JM.

2.1.4 Procedure

Participants were seated in total darkness with their head stabilized by a chin rest

and performed an open-loop pointing task. During the experimental procedure,

participants were instructed to look and point by raising the thumb to be exactly

underneath the target (green LED) as accurately as possible when cued by the

experimenter. All extraneous visual cues were removed to ensure that participants had to

use a non-visual source of information to localize the target. There were three

experimental conditions randomized in 5 blocks of 6 trials as to order: (a) Control: look

and point to target; (b) Task 1: look and point to the target while performing a muscle

contraction (JM) with the lower limbs; (c) Task 2: look at the target while performing a

muscle contraction and point 2–3 sec after the contraction has been released (see Figure 1

for illustration of the protocol).

46

2.1.5 Data analysis

Data were analyzed using a custom software program and focused on the end-

point accuracy of vergence eye movements and hand movements. Vergence angle (µ)

was obtained by subtracting the right horizontal eye position from the left horizontal eye

position. Vergence-specified distance (D) was calculated using the vergence angle and

the individual interocular distances (I): D = I / µ. Pointing error in the median plane was

calculated by subtracting real target position from the hand position data.

Vergence-specified distance and pointing error data were submitted to a repeated

measures, two-way analysis of variance (ANOVA) with condition (control, task 1, task 2)

and target position (far, near) as the independent variables. Post-hoc analysis was

performed using Tukey’s HSD test which was considered significant when p<0.05.

2.2 Results

Participants systematically overshot the target with the hand (Figure 2) and with

the eyes (i.e. converged beyond the target) in all the conditions (Figure 3). There was a

significant effect of condition (F(2,18)=11.94, p=0.0005). Results from the 2-way

ANOVA showed no significant interaction effect between condition and target position

(far or near) (F(2,18)=0.44, p>0.05). Post-hoc analysis revealed that pointing responses

were significantly less accurate in task 2 (mean pointing error 6.93±5.0 cm) compared to

47

Figure 1: Schematic illustration of the experimental protocol used in experiment 1: (a)

control task; (b) Task 1: look and point during JM; (c) Task 2: look during JM and point

after JM.

48

the control condition (mean pointing error 5.32±5.06 cm) and task 1 (mean pointing error

5.51±4.93 cm).

There were no significant differences between the mean vergence-specified

distance of any of the conditions (F(2,18)=0.26, p>0.05) and the interaction effect was

also non significant (F(2,18)=1.02, p>0.05). On average, participants looked beyond the

target by 62±24% (mean ± standard deviation).

Figure 2: Experiment 1: Mean pointing error of the hand. The figure illustrates the

significant difference between Task 2 and the other two conditions (Control and Task1).

Error bars show ±1 standard error.

49

Figure 3: Experiment 1: Average vergence-specified distance for near and far targets in

all the tasks. The targets were shown at a distance of 25 cm and 45 cm from the

participant which is shown by the dotted lines.

2.3 Discussion

We hypothesized that the JM would affect the localization performance by

altering the proprioceptive signal from the eye muscles, possibly via the activity of non-

twitch motoneurons. Data from the study provided partial support for the hypothesis, but

cannot be interpreted unambiguously. In particular, results showed that when participants

first made an eye movement to the target while the JM was performed and executed the

pointing response 2–3 sec after the contraction has been released, the pointing response

50

was significantly less accurate compared to the control condition or to the task when the

JM was performed throughout the trial. It might be surprising, at first, to find no

difference in pointing accuracy when JM was performed throughout the trial compared to

the control condition. One possible explanation for this effect is by considering that the

CNS continually monitors the afferent feedback from the EOM and that the JM alters the

signal sent to the CNS. The larger pointing error was found in the condition when the eye

movements and the hand movement were executed under different afferent feedback (i.e.

eye movement with JM, hand movement without JM). On the other hand, no significant

difference was found between the control condition and when JM was performed

throughout the trial because the movements of the eyes and the hand were programmed

and executed under the same afferent feedback.

Another explanation that must be considered is that the effects obtained in the

present study were due to the effect of JM on motoneurons of the arm muscles used for

pointing. Presumably, the JM has a general effect on all motoneurons (Delwaide &

Toulouse, 1981), and it is possible that the activity of muscle spindles in the arm muscles

was also altered and might have influenced the localization response. This limitation was

addressed in the next experiment.

A critical finding from this study was that the vergence eye movements and the

vergence-specified distance were not affected by the JM as shown by the lack of

differences between any of the conditions. These data provide support for the fact that JM

does not affect the actual eye position and, consequently, the differences in localization

response must be due to an altered registered eye position signal. This notion is consistent

with the fact that JM should alter the proprioceptive feedback from EOM via the non-

51

twitch neurons without altering the actual eye position because eye movements are

controlled by the twitch neurons.

On average participants converged beyond the target in all the tasks, which is a

finding consistent with a previous study by Malinov, Epelboim, Herst and Steinman

(2000). In that study, participants under-converged by 20-45% while looking and tapping

to targets under natural viewing conditions (i.e. head was not restrained and with full

visual feedback). In the present study, participants converged even farther beyond the

target, which is most likely due to methodological differences between the two studies:

participants in our study had restrained head movement and no visual reference.

3.0 Experiment 2

The purpose of the second experiment was to further examine whether the eye

position signal is indeed altered by the JM. The major caveat in experiment 1 was that the

JM could have affected the accuracy of the pointing response of the hand by altering the

spindle activity of the arm muscles. This limitation was addressed in experiment 2 by

using an entirely visual task, which involved a criterion-free perceptual judgment task.

Based on our results from the previous experiment, we hypothesized that that the

perceptual judgments would be significantly affected by the temporal order of the JM. In

other words, it was expected that target localization would be significantly affected when

one of the targets, either the first (standard target) or the second (comparison target), is

shown during altered eye muscle afferent feedback.

3.1 Method

52

3.1.1 Observers

Twenty one healthy adults with no history of any ocular disorders, mean age

33.4±10.6 years, participated in the second experiment (the sample included 10

participants who also took part in experiment 1).

3.1.2 Stimuli

The stimuli were white dots (visual angle 0.24 min arc) displayed on a flat CRT

monitor (refresh rate 85 Hz). The display was programmed using VPixx (VPixx

Technologies, Inc., Montreal, QC), a graphics generation and psychophysics testing

software, controlled by a MacIntosh G4 computer. Targets were shown in the earth

horizontal plane in the participant's midline, approximately 15.5 cm below eye level and

the viewing distance for the 5 targets ranged between 67.6 cm to 71.7 cm. The standard

target was shown at a constant location at a viewing distance of 69.7 cm. The vergence

angle required to converge on the 5 targets ranged between 5 to 5.5 degrees. One of the

comparison targets was shown in the same location as the standard target and the other

four were shown closer or farther than the standard.

3.1.3 Apparatus

JM involved an isometric, voluntary muscle contraction against resistance

performed with the shoulder abductor muscles (10 participants) or with the abductor

muscles of the legs (11 participants). A custom made device, based on a spring loaded

scale, was used to provide resistance when participants used the shoulder muscles to

perform the JM. Participants performed the maneuver by pulling their arms apart while

53

holding the device in their hands. The device used for resistance with the lower limbs and

the JM procedure was the same as in experiment 1.

3.1.4 Procedure

Participants were seated in total darkness and performed a two-alternative forced

choice task using the method of constant stimuli. At the beginning of each trial

participants were instructed to look at the standard target, which was shown for 2.5 sec,

and to remember its location when it disappeared. The comparison target was then shown

at 1 of 5 possible locations, determined randomly by the computer. Participants made a

judgment by saying whether the comparison target appeared ‘nearer’ or ‘farther’ than the

standard target. There were four experimental conditions: (a) Control: standard and

comparison targets were shown with no JM; (b) Task 1: standard target appeared during

the JM, and comparison target appeared after the JM was released; (c) Task 2: standard

target appeared when the JM was not performed, and the comparison target appeared

during the JM; (d) Task 3: standard and comparison targets appeared while the JM was

performed (see Figure 4 for illustration of the protocol). The experimental conditions

were completely randomized. In each experimental condition the comparison target was

shown 10 times at each of the 5 locations for a total of 200 trials per participant.

Our prediction was that participants’ judgments would be affected by the order of

JM. In particular, we expected the largest difference between Task 1 and Task 2 because

one of the targets, either the standard or the comparison, was presented while the

feedback from EOM was altered. Task 3 served as another control condition because

both targets were shown with the same, altered feedback.

54

Figure 4: Schematic representation of the experimental procedures used in experiment 2.

3.1.5 Data analysis

The proportion of ‘near’ responses was calculated for each participant and task at

the five locations where the comparison target was shown and a psychometric function

fitted. All psychometric curves were visually inspected to determine whether the type of

muscle contraction (shoulder or leg abductor muscles) resulted in any qualitative

differences. Subsequently, an overall psychometric function based on the mean of all

participants was fitted for each task.

The point of objective equality (POE) was defined as the proportion of ‘near’

responses when the comparison target was shown at the same location as the standard

55

target. The POEs for each participant and task were submitted to a one-way ANOVA

with task (control, task 1, task 2, task 3) as the independent variable.

Data for each participant and task was fitted using a logistic regression (SAS, ver

8.1). The goodness of fit of the model was tested using the Hosmer-Lemeshow statistic

and a non-significant result was used to verify that the logistic model was appropriate.

The point of subjective equality (PSE) was calculated using the estimated parameters

(slope and intercept) from the logistic model. The PSE is the point at which the logistic

function yields a probability of 0.5 (i.e. the comparison target is perceived as nearer than

the standard target 50% of the time). Cook’s distance was used to identify influential

observations (outliers) in the dataset. The PSE, intercept and slope were submitted to a

one-way ANOVA with task (control, task 1, task 2, task3) as the independent variable.

Post-hoc analysis was performed using Tukey’s HSD test which was considered

significant when p<0.05.

3.2 Results

Preliminary inspection of the individual psychometric curves did not reveal any

differences in the performance of participants who used the shoulder abductor muscles as

compared to those who used the leg abductor muscles to perform the JM. Therefore, the

data was collapsed and the mean performance of all participants in each condition is

shown in Figure 5. The individual data of twenty of the participants showed a consistent

trend which is evident in the mean data shown in Figure 5. Participants consistently

perceived the target as farther when the JM was performed during the presentation of the

second target (Task 2). The results of one of the participants were a mirror-image of those

56

of the rest of the group (i.e. the comparison target was perceived as nearer on Task 2),

which was most likely due to a misinterpretation of the instructions. These data were not

included in the statistical analysis2.

Figure 5: Mean proportion of ‘near’ responses for each comparison target location (at 0

both targets were presented at the same location). Error bars show ± 1 standard error.

Figure 6 shows the differences in the POE between the Control condition and

Tasks 1 and 2 for individual participants (POE for Task 3 is not shown). The mean POEs

across conditions were: Task 1=0.61, Task 2=0.34, Task 3=0.51, and Control=0.45

[F(3,57)=10.62, p<0.0001]. Post-hoc comparisons showed that performance was

2 Including this subject’s data did not change the overall statistical results (i.e. both PSE and POE results

57

significantly different between Task 1 and Task 2. Overall, the data showed that

participants perceived the location of the comparison target as nearer when the JM was

performed during the presentation of the standard target (Task 1) as compared to when

the JM was performed when the comparison target was shown (Task 2) or when the JM

was not performed (control).

Figure 6: Differences in the POE between the Control condition and Tasks 1 and 2 for

individual participants (ID 1 to 20). The y-axis represents the difference in proportion of

‘near’ responses between Task 1 and Control & Task 2 and Control. Positive values

indicate that the comparison target was reported as ‘nearer’ and negative values indicate

that the comparison target was reported as ‘farther’ with respect to the control task.

remained significant).

58

The logistic model fitted the experimental data well for the majority of the

psychometric curves (76 out of 80), which was supported by the non-significant result

from the Hosmer-Lemeshow test. Although in four cases (one in the control condition

and three in Task 3) the test was statistically significant, the logistic model was still used

to fit the data. Two outliers were detected using the Cook’s test in the PSE dataset (one

observation in Task 2 and one in Task 3). These two observations were twice the

magnitude of the recommended cut-off value (4/n) and they were replaced by the

geometric mean obtained from the 19 observations for a given task.

Analysis performed on the parameters obtained from the logistic regression model

showed statistically significant differences between conditions for the PSE

[F(3,57)=13.18, p<0.0001] and intercept [F(3,57)=8.70, p<0.0001], but not for the slope

[F(3,57)=0.42, p=0.7360]. Post hoc analysis revealed that the PSE was significantly

higher in Task 2 compared to the other conditions (Task 2=10.0 mm, Task 1=3.5 mm,

Control=5.0 mm, Task 3=3.4 mm), which means that in Task 2 the comparison target had

to be presented significantly nearer in order to be perceived at the same location as the

standard target. The value of the intercept was significantly higher in Task 1 (0.89) than

in Task 2 (0.29), the Control condition (0.61), and Task 3 (0.62). These results suggest

that the JM influenced the gain but not the sensitivity of the perceptual judgments.

3.3. Discussion

Results from the second experiment provided support for our hypothesis that JM

affects the registered position of the eyes and shed more light on the effect of the JM. A

schematic diagram summarizing the results is shown in Figure 7. In the case when both

59

targets (standard and comparison) were shown at the same location and the JM was

performed when the standard target was presented, participants reported that the

comparison target was ‘nearer’. This result suggests that participants perceived the

location of the standard target as farther with the JM. In contrast, when the JM was

performed while the comparison target was presented, the comparison target was reported

as ‘farther’, which again suggests that during JM the location of the target is perceived as

farther. In summary, results from the second experiment provide strong evidence that eye

position is registered as more divergent when the JM is performed.

Figure 7: Summary and interpretation of results for experiment 2.

60

4.0 Experiment 3

The critical finding from experiment 1 was that the JM manipulation did not

affect the actual eye position. Experiment 3 was conducted to further examine whether

JM has any effect on eye movements by examining a different type of eye movements:

the saccadic system. We chose saccadic eye movements for two reasons. First, saccades

are fast eye movements programmed by different cortical and subcortical areas

(Carpenter, 1988) than the vergence eye movements which were examined in experiment

1. Secondly, a neuroanatomical tracing study has shown that the twitch and non-twitch

motoneurons receive premotor input from distinct brainstem areas, which are associated

with the saccadic and vergence systems, respectively (Wasicky et al., 2004). Thus, we

hypothesized that if the JM acts via the non-twitch motoneurons, the parameters of the

saccadic eye movement should not be affected.

4.1 Method

4.1.1 Observers

Ten healthy adults with no history of any ocular disorders, mean age 32.2±12.9

years, participated in the experiment (3 participants also took part in experiments 1 and

2).

4.1.2 Stimuli

The stimulus was a white dot which subtended 0.25 degrees of visual angle. The

stimulus was rear-projected onto a black background and displayed at 10 eccentricity to

the left and right of the fixation. The stimulus presentation was controlled by VPixx

61

(VPixx Technologies, Inc., Montreal, QC), a graphics generation and psychophysics

testing software, controlled by a MacIntosh G4 computer.

4.1.3 Apparatus

The method of eye movement recording and the JM manipulation procedure were

the same as described in the methods section of Experiment 1.

4.1.4 Procedure

Participants were seated in a dimly lighted room and performed saccadic eye

movements to randomly presented stimuli (±10° to the left and right of the fixation

point). In the experimental condition participants started the JM while looking the central

fixation point and performed the JM during the saccadic eye movement. In the control

condition eye movements were performed without the JM. The stimuli were shown 10

times at each location for a total of 40 trials in the control and experimental condition.

4.1.5 Data analysis

Saccades that followed the presentation of the stimulus were detected using a

custom software program using the velocity criterion of 30/sec. All saccades identified

by the program were visually confirmed by the experimenter. Peak velocity and

amplitude of the first saccade for each trial were determined using a custom software

program. Data were submitted to repeated-measures ANOVA with condition (control,

experimental) as the independent variable.

62

4.2 Results

As expected, the data showed no significant differences for peak velocity

(F(1,9)=0.89, p>0.05) and amplitude (F(1,9)=0, p>0.05) between the conditions. The

mean peak velocity in the control and experimental conditions were 302.6161.97/s and

306.7563.95/s, respectively. The mean amplitude of the first saccade in the control and

experimental conditions were 9.301.56and 9.371.16, respectively.

4.3 Discussion

Overall, the results from experiment 3 suggest that the JM does not affect the

actual eye movements as shown by the lack of differences in saccadic parameters

between the control and experimental conditions. The negative findings from this

experiment provide additional support to our hypothesis that the JM acts via the activity

of non-twitch motoneurons and has no effect on the twitch motoneurons.

5.0 General Discussion

The results from the present study provide novel insight into the mechanism

which may be involved in the use of sensory feedback from the EOM. Behavioral and

psychophysical data support the hypothesis that the JM alters the registered eye position,

but not the actual eye position. We propose that the altered eye position signal is due to

the effect of the JM which changes the gain of the sensory feedback from the eye

muscles, possibly via the activity of non-twitch motoneurons.

The EOM fibers can be classified into several types based on their innervation,

morphological, histochemical, and contractile properties (for a review see Spencer &

63

Porter, 1988). Two types of efferent nerve endings are found on the EOM fibers: single,

large end-plates (en plaque) and multiple, small fiber endings (en grappe). Fibers that

receive single innervation (SIF) have regularly spaced fibrils, large number of

sarcoplasmic reticulum, and a well developed transverse tubule system, which allows

these fibers to conduct fast action potentials. In contrast, the MIF have poorly developed

sarcoplasmic reticulum and do not generate action potentials, instead, they generate a

prolonged graded response when stimulated at thresholds that are 3-6 times greater than

the most excitable SIF (Eakins & Katz, 1972). Thus, the MIF are sometimes referred to

as non-twitch fibers. Given that the MIF do not contribute to the tension developed by the

muscle (Fuchs and Luschei, 1971), the question that arises is, what role could these non-

twitch fibers play in oculomotor processes?

Some insight to this question comes from recent anatomical tracing studies by

Buttner-Ennever and colleagues (2001) who demonstrated that the EOM receive dual

innervation from separate groups of ocular motoneurons. The close association between

non-twitch motoneurons, the MIF and the PEs has led several authors to propose a

proprioceptive hypothesis for the control of eye movements (Buttner-Ennever & Horn,

2002; Porter et al., 1995; Robinson, 1991). In particular, more than a decade ago

Robinson (1991) referred to PE and MIF as the inverted muscle spindles, and recently

Buttner-Ennever & Horn (2002) suggested that the non-twitch motoneurons might have a

role analogous to the efferent fibers which control the gain of the intrafusal fibers in the

skeletal muscles. The control of sensory feedback from EOM has been demonstrated in

an ungulate by Whitteridge (1959). However, at the present time there is no direct

anatomical evidence confirming that non-twitch motoneurons modulate the sensory

64

feedback in primates, which is partly due to the fact that the sensory pathway and the

location of the somata of the PEs have not been established. The non-twitch motoneurons

share similarities with the motoneurons in that they are both smaller then their

corresponding alpha motoneurons, and their activity does not generate fast action

potentials or contribute to changes in muscle tension directly. Recording from the cell

body of the EOM sensory neuron while stimulating the non-twitch motoneurons would

provide unequivocal evidence for a gain control regulation of proprioceptive feedback

from the eye muscles.

In the present study we used a proxy method (JM) to alter the activity of the

motoneurons. While the JM is performed, the amplitude of all stretch reflexes is

facilitated, which was first reported by the Hungarian physician Ernst Jendrassik

(Delwaide & Toulouse, 1981). The monosynaptic tendon reflex involves only two

neurons: the Ia afferent and the alpha motoneuron, but there are several mechanisms that

could be involved in the facilitation of the reflex. First, the effect could be mediated via

the gamma feedback loop: increased activity of the motoneurons would increase the

gain of the muscle spindle (i.e. increased discharge rate of the spindle), which would

result in a greater afferent volley when the muscle is stretched. Second, the facilitation

effect could be mediated via supraspinal control which can decrease the presynaptic

inhibition of the Ia afferent or increase the excitability of the alpha motoneuron. In

addition, there could be polysynaptic facilitation via interneurons in the spinal cord

contributing to the effect (Dowman & Wolpaw, 1988; Gregory et al., 2001; Murthy,

1978; Zehr & Stein, 1999). A detailed examination of the factors that affect the amplitude

of the reflex suggested that all the above mechanisms might contribute to the reflex

65

reinforcement effect of the JM (Delwaide & Toulouse, 1981). In particular, the

contribution of the motoneurons to reflex reinforcement might be more relevant when

the contraction is maintained longer than 600 ms, which was the case in the present

study.

As mentioned previously, the JM has been studied extensively in the context of

reflex reinforcement. Although stretch reflexes have never been recorded in the EOM

(Keller & Robinson, 1971), neural activity in response to passive stretch of the EOM has

been reported in cortical and subcortical areas (for review, see Donaldson, 2000). Clearly,

the proprioceptive signals from the EOM are being used by the CNS despite the lack of

reflex responses in the eye muscles.

Behavioral studies have shown that proprioceptive signals from the EOM are used

during localization tasks (Bridgeman & Stark, 1991; Gauthier et al. 1990a; Roll et al.,

1991; Velay et al., 1994) and, as we show in the present study, a pointing task and a

perceptual judgment task. Since, presumably, JM has a general effect that up-regulates

the activity of the system, we hypothesized that the eye position signal would be altered

if proprioceptive feedback from the EOM is affected by the activity of non-twitch

motoneurons. Our study provides preliminary support for the hypothesis. In particular,

the JM affected the bias of the judgment but not its sensitivity (slope), which is consistent

with the action of the motorneurons on muscle spindles (Prochazka, 1989).

In experiments 1 and 2, participants performed vergence eye movements and the

task involved judgments of absolute depth. Since all visual cues were removed,

participants had to rely on the eye position signal to perform the task. A vergence task

was chosen because it has been reported that the pre-motor input to the non-twitch

66

motoneurons comes from caudal supraoculomotor area, central mesencephalic reticular

formation, medial vestibular nuclei (parvocellular division), and nucleus prepositus

hypoglossi (Wasicky et al, 2004), which are brainstem regions involved in vergence eye

movements, ocular following, and gaze holding mechanisms. A critical finding from

experiments 1 and 3 was that the actual eye position and saccadic parameters were not

affected by the JM, which suggests that the manipulation had no effect on the alpha

motoneuron activity and did not result in change of muscle tension. Instead, the JM

affected the participants’ pointing and perceptual responses. Overall, these results imply

that participants made judgments based on the altered registered eye position signal from

EOM proprioceptors and not on the actual eye position signal which was sent to the eye

muscles.

In conclusion, our results suggest that registered eye position is altered by the JM

while the actual eye position is not affected. We propose that this effect may be mediated

via the activity of non-twitch motorneurons. These results may have important clinical

implications for the treatment of strabismus, which is an ocular disorder involving

deviation of one or both eyes due to extraocular muscle (EOM) imbalance. Surgical

intervention, which involves cutting the EOM at the musculotendinous junction, is a

common treatment for strabismus, but often does not result in regaining optimal function.

Many children have to undergo multiple surgeries and yet they do not develop normal

binocular function (stereoscopic vision and vergence eye movements). It is possible that

the lack of success is partly due to the damage sustained at the myotendinous region of

the muscles which contains the putative proprioceptors of the eye muscles (Steinbach,

1987).

67

PAPER 2

LOCALIZATION IN THE FRONTAL PLANE IS NOT SUSCEPTIBLE

TO MANIPULATION OF AFFERENT FEEDBACK

VIA THE JENDRASSIK MANEUVER

1E. Niechwiej-Szwedo, E.G. González, M.C. Verrier, A. M. Wong, M.J. Steinbach

Vision Research (2008); 48:724-732

1 See Appendix 1 for contributions of each author

68

ABSTRACT

We have previously shown that registered vergence eye position is altered while participants

perform the Jendrassik Maneuver (JM). We proposed that the altered eye position signal

registration is due to the effect of the JM which changes the gain of the sensory feedback from

the eye muscles, possibly via the activity of non-twitch motoneurons. We conducted two studies

to further extend and clarify one of our previous findings by examining whether the JM also

affects registered eye position during localization in the frontal plane. Since the non-twitch

motoneurons do not receive premotor input from areas involved in the programming of saccades,

we hypothesized that localization responses associated with the saccadic system should not be

affected by the JM. The data confirmed our prediction. We propose that the non-twitch

motoneurons are involved in parametric adjustment of the proprioceptive feedback loops of

vergence but not version eye movements.

69

1.0 Introduction

Good eye-hand coordination is essential for accurate performance of daily activities.

For example, reaching to pick up a cup of coffee is a simple movement and yet it requires a

complex sensorimotor transformation of visual and somatosensory afference into a coordinated

pattern of muscle activations. To perform this simple motor act, the central nervous system

(CNS) has to process and integrate information from several receptors: the retinal location of the

cup, the position of the eyes in the orbits, the position of the head, the arm and the hand.

In the case of skeletal muscles, it has been unequivocally recognized that muscle spindles

provide the CNS with information concerning limb position and velocity (Matthews, 1981). In

addition, a large body of research has addressed the structural properties, anatomical pathway,

and central control of muscle spindles (for a review see Hulliger, 1984). In contrast, eye muscle

proprioceptors have not received similar attention. Although there is still controversy regarding

the afferent pathway, a recent, elegant study by Wang and colleagues (Wang, Zhang, Cohen &

Goldberg, 2007) provided evidence that eye position is represented in the somatosensory area 3a

in rhesus monkeys.

The CNS can obtain eye position information from two non-retinal sources: outflow

(copy of the motor command) and inflow (signals from the eye muscles). Although, the

contribution of inflow to eye position sense has been debated for years (for a review see

Donaldson, 2000), studies have shown that proprioceptive signals from the eye muscles in

neurologically intact individuals play a significant role in the programming of eye movements

(Knox et al., 2000; Weir & Knox, 2001), during egocentric localization tasks (Bridgeman &

Stark, 1991; Gauthier et al., 1990a,b; Roll et al., 1991; Velay et al., 1994) and adaptation of

smooth pursuit (vonDonkelaar et al., 1997).

70

Extraocular muscles (EOM) contain at least two receptors which could provide eye

position information: muscle spindles and palisade endings (PE). Muscle spindles are found in

the orbital layer of the human EOM; however, their function has been questioned due to their

unusual morphological characteristics (for a review see Ruskell, 1989). In addition, muscle

spindles have not been found in the EOM of some species, such as cats or rhesus monkeys. In

contrast, PE have been found in the EOMs of all the species tested to date including humans,

cats, rats, sheep, and rehsus monkeys (Alvarado-Mallart & Pinçon-Raymond, 1979; Blumer et

al., 1998; Buttner-Ennever et al., 2001; Eberhorn, Horn, Eberhorn et al., 2005; Richmond et al.,

1984). PE are uniqe to the EOM and they are associated with the myotendious region of the

global multiply innervated fibers (MIF). Several studies have considered that they might be the

EOM proprioceptors based on their morphological characteristics (Alvarado-Mallart & Pinçon-

Raymond, 1979) and retrograde tracing studies (Billig et al., 1997). However, recent

histochemical examination of the musculotendinous junction in the cat and monkey has revealed

that the region containing the PE is immounoreactive to markers for cholinergic nerve fibers and

nerve terminals, which have been traditionally associated with motoneurons (Konakci, Streicher,

Hoetzenecker, Blumer et al., 2005; Konakci, Streicher, Hoetzenecker, Haberl et al. 2005).

Although the question of whether PE have a sensory or a motor function has yet to be

resolved, several authors have proposed the possibility that PE, along with the MIF, might have a

proprioceptive role in the control of eye movements. Robinson (1991) was the first to use the

term ‘inverted muscle spindle’ to suggest that the non-twitch MIF and the PE might be

comparable to the gamma ()-spindle system found in the skeletal muscles. This hypothesis has

been further extended by Buttner-Ennever and colleagues (Buttner-Ennever et al., 2002) based

71

on their neuroanatomical tracing studies which demonstrated that the MIF receive innervation

from separate groups of ocular motoneurons (Buttner-Ennever et al., 2001).

The goal of the present studies was to test the above hypothesis using a psychophysical

approach and the Jendrassik Maneuver (JM). The JM is a forceful voluntary muscle contraction

of any muscle group. Previous studies have shown that the JM alters the excitability of tendon

reflexes. Specifically, the amplitude of the reflex is enhanced and the facilitation is dependent on

the strength of the reinforcing maneuver (Delwaide & Toulouse, 1981). One hypothesis that has

been proposed to explain the reinforcing effect is that the JM increases the excitability of the

gamma system (Murthy, 1978). Since the gamma system regulates the baseline activity of

spindles, the JM increases the baseline activity of spindles which become more sensitive to the

upcoming stimulus resulting in a larger response when the muscle is stretched.

JM not only affects the excitability of reflexes but also the perceived position of the arms

and eyes. In a recent study, Yasuda and colleagues (Yasuda, Izumizaki, Ishihar, Sekihara,

Atsumi & Homma, 2006) examined the upper limb position sense while participants performed a

reinforcing maneuver with their quadriceps muscles. Data showed that the arm was perceived in

a more extended position when the JM was performed and the error increased with the intensity

of the quadriceps contraction.

The effect of JM on registered vergence eye position was shown in our previous study

(Niechwiej-Szwedo, González, Bega, Wong, Verrier & Steinbach, 2006). In short, while the JM

was performed, targets were perceived as farther while the actual eye position was not affected.

In the current investigation we conducted two studies to examine whether the JM affects

localization responses associated with saccadic eye movements. Manual pointing responses were

examined in the first study and perceptual localization was investigated in the second study.

72

Although version and vergence share a final common pathway, different pre-motor areas

are involved in programming of these eye movements (for a review see Buttner-Ennever et al,

2005). In addition, conjugate and disconjugate eye movements are differentially susceptible to

manipulations of afferent feedback. For example, saccades do not seem to be affected by

sectioning of the ophthalmic branch of the trigeminal nerve (deafferentation) whereas vergence

is disrupted by the same procedure (Guthrie et al., 1982). Additionally, the accuracy of pointing

to targets arranged along the horizontal axis was not affected by deafferentation (Lewis,

Gaymard & Tamargo, 1998). In contrast, binocular depth discrimination was impaired in cats

following the same procedure (Fiorentini, Mafei, Cenni & Tacchi, 1985). Given the differences

between the saccadic and vergence systems, our studies were designed to further explore the

hypothesis that version and vergence are differentially susceptible to manipulations of afferent

feedback using the JM.

The current studies also help to extend and clarify the findings from one of our previous

experiments (Niechwiej-Szwedo et al., 2006). First, we showed that the JM affected pointing

responses to targets in depth, but it is possible that the pointing error was due to the effect of the

JM on the upper limb muscles instead of the eye muscles. Thus, in the first experiment of our

current examination, participants pointed to targets presented along the frontal plane. It was

hypothesized that if the JM affects the activity of non-twitch motoneurons which do not receive

premotor monosynaptic input from areas involved in the programming of saccades (Wasicky et

al., 2004), the pointing responses should not be affected by the JM. Alternatively, pointing

responses might be affected if the result that we have previously observed was due to the effect

of the JM on the limb muscles.

73

Secondly, we showed that the JM does not affect the actual vergence or saccadic eye

movements; however, the perceptual localization of eccentric targets was not explicitly

examined. Therefore, in the second experiment we asked participants to localize briefly flashed

targets after they made a saccadic eye movement. We hypothesized that if the JM acts through

the activity of non-twitch motoneurons then perceptual localization associated with saccadic eye

movements should not be affected. Alternatively, if the effect of the JM occurs via a different

neural mechanism then we might see increased localization errors during saccades similar to the

overshoot errors found in the case of the vergence system. Eye movements were recorded in

experiment 1 in order to replicate our previous findings and verify that the JM does not affect

activity of the twitch motoneurons.

2.0 Experiment 1

2.1 Methodology

2.1.1 Participants

Participants in both studies had normal or corrected-to-normal visual acuity of 20/20 or

better and stereopsis of at least 40 seconds of arc as measured with the Titmus test (Titmus

Optical Co., Inc., Petersburg, Virginia 23805). All experimental protocols were approved by the

Ethics Review Boards at the University of Toronto and the University Health Network. The

research adhered to the tenets of the Declaration of Helsinki, and all participants gave their

written informed consent prior to participating. Ten healthy adults (8 females) with no history of

ocular disorders and a mean age of 34±16 years, participated in the first experiment. Five of the

participants also participated in the experiment that involved pointing in depth which was

reported in our previous paper (Niechwiej-Szwedo et al, 2006).

74

2.1.2 Stimuli

The stimuli were 3 red light-emitting diodes (LEDs) placed on a custom-made black

board and controlled by the experimenter via a trigger box. The fixation stimulus was aligned

with the participant's midline and the other two LEDs were located 8 to the left and right of

fixation. All LED’s were presented slightly below eye level. The two eccentric targets were

located 51 cm from the participants so everyone could point to the target comfortably. The board

was positioned so that participants could not see their arms and they had no feedback about the

accuracy of their pointing. The only difference between the stimuli presented in this experiment

and those used in the previous study (Niechwiej-Szwedo et al., 2006) was the fact that the two

LEDs were presented eccentrically in the frontal plane and not in depth.

2.1.3 Apparatus

Horizontal and vertical position of both eyes was monitored and recorded using an infra-

red eye-tracker system (El-Mar series 2020, Toronto, Ontario, Canada). Horizontal and vertical

eye positions were obtained from the relative positions of two corneal reflections and the center

of the pupil. Prior to data collection, the eye tracker was calibrated. The system accuracy is 0.5°

with a linear visual range of ±40° horizontally and ±30° vertically, the resolution is 0.1°, and it is

free from drift. Eye position data were sampled at 120 Hz and stored on a computer for further

analysis. Arm movement data were recorded at 60 Hz with a resolution of 0.5 mm using an

electromagnetic device (Flock of Birds, Ascension Technology Co, Burlington, Vermont, USA).

The receiver was placed on the thumb of the participant's dominant hand, which was used for

pointing.

JM involved an isometric, voluntary muscle contraction which was performed with the

75

abductor muscles of the legs against resistance. The device used for resistance was a Thigh

Master™. Participants were asked to perform each contraction at 75% level of their maximal

voluntary contraction, which was determined prior to the initiation of the experiment. To ensure

that the isometric contraction was performed at a consistent level throughout the experiment, a

string tied around the Thigh Master™ was pulled taut when the muscle contraction was

executed. Participants were instructed to hold the string taut while performing the JM.

2.1.4 Procedure

Participants were seated in total darkness while performing an open-loop pointing task.

During the experimental procedure participants were instructed to look and point by raising the

thumb to be exactly underneath the target (red LED) as accurately as possible when cued by the

experimenter. Participants initiated the pointing movements from the same starting position

which was identified by a tactile cue placed on a table at their midline. All extraneous visual cues

were removed to ensure that participants had to use a non-visual source of information to

localize the target. There were three experimental conditions randomized as to order: (a) control:

look and point to the target; (b) task 1: look and point to the target while performing the JM; (c)

task 2: look at the target while performing the JM and point 2–3 sec after the contraction had

been released (see Figure 8 for illustration of the protocol). Participants completed 15 trials in

each condition.

76

Figure 8: Schematic illustration of the experimental protocol used in experiment 1: (a) control:

no JM; (b) task 1: JM performed during saccade and pointing; (c) task 2: JM performed during

saccade, but not during pointing.

77

2.1.5 Data analysis

Data analysis was conducted using custom software and focused on the end-point

accuracy of the hand and saccadic eye movements. The pointing data for each trial were

examined visually. The end of the pointing trajectory was established from the position and

velocity traces. The end point was the position of the hand when it came to rest and it was

calculated as the mean of 50 ms when the hand velocity was 0/sec. The calibration of the hand

position performed at the end of the experiment for each subject could not be used to calculate

the errors due to a noisy signal. We have, however, relative error measures that allowed us to

compare the two experimental conditions.

Pointing data, saccade amplitude and saccade peak velocity were submitted to repeated

measures, two-way analyses of variance (ANOVA) with condition (control, task 1, task 2) and

target position (left, right) as the independent variables.

2.2 Results

The data from individual participants for each condition were plotted and visually

inspected for trends to determine if participants were more likely to overshoot or undershoot the

target in the experimental condition in comparison to the control condition. No trends were

evident: half of the participants overshot the target regardless of the experimental condition. The

lack of a reliable effect was confirmed by the statistical analysis. Results from the ANOVA for

the pointing response showed no significant interaction effect between the task and pointing to

the left or right targets (F(2,18)=0.66, p > .05). The mean pointing responses to the right sided

targets were 8.95±2.66 cm for the control condition, 9.57±2.79 cm for task 1 and 9.10±2.56 cm

for task 2. The mean pointing responses to the left sided targets were 9.742.50 cm for the

78

control condition, 9.79±2.85 cm for task 1 and 9.56±2.90 cm for task 2. There was no difference

in pointing accuracy between the conditions (F(2,18)=1.33, p > .05). The distribution of pointing

responses to both targets for each task is shown in the boxplots in Figure 9.

Figure 9: Distribution of pointing responses obtained in experiment 1. The boxplot contains the

middle 50% of the data (the upper edge is the 75th

percentile and the lower edge is the 25th

percentile), the cross in the box represents the median. The lines extending from the boxplot

(whiskers) indicate the 1st and 99

th percentile.

The data showed no significant differences between the conditions for the amplitude of

saccadic eye movements (F(2,18)=0.67, p > .05). The mean amplitude of the first saccade to the

79

right sided target was 6.73±1.4 for the control condition, 6.62±1.8 for task 1 and 6.65±1.1 for

task 2. The mean amplitude of the first saccade to the left sided target was 6.081.3 for the

control condition, 6.15±1.3 for task 1 and 5.97±2.4 for task 2. Statistical analysis also showed

no significant differences between conditions for the velocity data (F(2,18)=1.19, n.s). The mean

velocity of the first saccade to the right sided target was 234±1.49/sec for the control condition,

228±67/sec for task 1 and 231±44/sec for task 2. The mean velocity of the first saccade to the

left sided target was 22067 for the control condition, 226±52/sec for task 1 and 215±84/sec

for task 2.

2.4 Discussion

Results from this study show that the accuracy of pointing responses associated with

saccades are not affected by the JM. These data help to clarify our previous findings, which show

that the JM affects pointing responses to targets presented in depth (Niechwiej-Szwedo et al.,

2006). Since the JM has a general effect on the gamma system, presumably affecting all muscles

(Delwaide & Toulouse, 1981), the pointing error obtained in the previous study could have been

due to the effect of the JM on the limb muscles instead of the EOM. Given that the non-twitch

motoneurons do not receive monosynaptic input from premotor areas involved in the

programming of saccades (Wasicky et al, 2004), the results from the current study support the

hypothesis that the pointing error obtained in the previous study was due to the effect of the JM

on the EOM muscles and not on the limb muscles.

As expected, we found that the JM did not affect the actual eye movements as shown by

the lack of significant differences in saccadic parameters between the control and experimental

conditions. Again, the negative findings from this experiment provide additional support to our

80

hypothesis that the JM acts via the activity of non-twitch motoneurons and has no effect on the

twitch motoneurons.

3.0 Experiment 2

3.1 Methodology

3.1.1 Participants

Eleven healthy adults (8 females) with no history of ocular disorders and a mean age

34±16 years participated in the second experiment.

3.1.2 Stimuli

The stimuli were 0.25 white dots displayed on a black background on a flat CRT

monitor (refresh rate 160 Hz). The display was programmed using VPixx (VPixx Technologies,

Inc., Montreal, QC), a graphics generation and psychophysics testing software, controlled by a

Macintosh G4 computer. The fixation stimulus was presented in the participants’ midline and the

other two targets were presented randomly at a 10 eccentricity to the left or right of fixation.

The comparison stimulus was also a 0.25 white dot and it was presented on the same side as the

target in one of five locations: 8, 9, 10, 11 and 12 away from the fixation stimulus.

3.1.3 Apparatus

The JM manipulation procedure was the same as described in the methods section of

experiment 1.

3.1.4 Procedure

Participants were seated in total darkness and performed a two-alternative forced choice

81

task (2-AFC) using the method of constant stimuli. At the beginning of each trial participants

were instructed to look at the fixation dot, which was shown for a variable time ranging from 1.5

sec to 2 sec. The target was flashed briefly for 50 ms to the left or the right of fixation which was

determined randomly by the computer. Participants were instructed to move their eyes as quickly

as possible to the location were they saw the target appear and to keep fixating on that location.

The comparison target was shown after 2.5 sec at 1 of 5 possible locations and participants had

to report whether the comparison target was to the left or to the right of their current fixation.

There were three conditions: (a) control: participants executed the eye movement and made the

judgement without the JM; (b) task 1: participants performed the JM while the target was shown

and during the eye movement, but not during the perceptual judgment; (c) task 2: participants

performed the JM during the perceptual judgements but not during the eye movement (see Figure

10 for illustration of the protocol).

3.1.5 Data analysis

For each participant, the proportion of ‘left’ responses was calculated for each, target

presentation side and task at the five locations where the comparison target was shown and a

psychometric function was fitted. All psychometric curves were visually inspected for trends.

Subsequently, an overall psychometric function based on the mean of all participants was fitted

for each task.

82

Figure 10: Schematic illustration of the experimental protocol used in experiment 2: (a) control:

no JM; (b) task 1: JM performed during the presentation of the standard target and saccade; (c)

task 2: JM performed during perceptual localization.

83

Data for each participant, target side and task were fitted using a logistic regression (SAS,

ver 8.1). The goodness of fit of the model was tested using the Hosmer-Lemeshow statistic and

a non-significant result was used to verify that the logistic model was appropriate. The point of

subjective equality (PSE) was calculated using the estimated parameters (slope and intercept)

from the logistic model. The PSE is the point at which the logistic function yields a probability

of 0.5 (i.e. the comparison target is perceived to the left of the standard target 50% of the time).

The PSE, intercept and slope were submitted to repeated measures ANOVA each with task

(control, task 1, task 2) and target side as the independent variables.

3.2 Results

Preliminary inspection of the individual psychometric curves did not reveal any

consistent trends in differences between the conditions. The logistic model fitted the majority of

the psychometric curves well (62 out of 66). Although in four cases (two in the control condition,

one in task 1 and one in task 2) the Hosmer-Lemeshow test was statistically significant (i.e., the

data did not fit the model), the logistic model was still used to fit the data.

The mean psychometric curve is shown in Figure 11. Results from the statistical analyses

showed no significant interaction effect between the experimental conditions and target side:

PSE (F(2,19)=3.28, p > .05), slope (F(2,19)=0.19, p > .05), intercept (F(2,19)=0.54, p > .05). The

main effect of condition was also not significant: PSE (F(2,20)=1.01, p > .05), slope

(F(2,20)=0.26, p > .05), intercept F(2,20)=0.13, p > .05).

84

Figure 11: Mean proportion of ‘left’ responses for the five comparison targets (at ±10 both the

standard and comparison target were presented at the same location). (a) standard target

presented in the left hemifield; (b) standard target presented in the right hemifield. Error bars

show ±1 standard error of the mean.

4.0 Discussion

We conducted two studies to examine whether the JM affects the saccadic system

similarly to the vergence system, which would help us to elucidate the potential neural

mechanism involved in mediating the effect of the JM on the vergence system. We have

previously proposed that the JM acts through the activity of non-twitch motoneurons (Niechwiej-

85

Szwedo et al., 2006) which receive monosynaptic input from premotor areas involved in the

programming of vergence eye movements but not from areas involved in the programming of

saccades (Wasicky et al., 2004). Thus, we hypothesized that if the JM acts through the non-

twitch motoneurons then the localization associated with saccades should not have been affected

by the manipulation. In contrast, if we had found that responses associated with saccades were

affected by the JM, then it would have been more likely that a different neural mechanism was

involved in mediating the effect. Overall, the results from both experiments showed that the JM

did not affect pointing or perceptual localization of targets presented in the frontal plane. Our

data support the first hypothesis and our earlier proposal that the JM affects the gain of the

proprioceptive feedback from EOM via the non-twitch motoneurons.

One of the limitations of our studies is the fact that our hypotheses are based on

neuroanatomical tracing studies which were conducted in sub-human primates. At the present

time it is unknown whether the human EOM fibers also receive dual innervation from ocular

motor nuclei (Buttner-Ennever et al., 2001). Nonetheless, human EOM do contain similar fiber

types to those found in primates and other mammals (Wasicky, Ziya-Ghazvini, Blumer, Lukas &

Mayr, 2000) and the PE are found in the global the multiply innervated fibers (MIF) in humans

(Richmond et al., 1984) and monkeys (Ruskell, 1978). Overall, the human and primate EOM are

remarkably similar in their organization, histochemical properties and repertoire of eye

movements, thus, we believe that it is likely that the dual innervation hypothesis can be extended

to humans as well.

A longstanding question in oculomotor physiology concerns the functional significance

of the MIF. It has been proposed that the MIF might participate in fine foveation eye movements

or might be part of a proprioceptive feedback loop (Spencer & Porter, 2005; Buttner-Ennever,

86

Konakci & Blumer, 2005). These two possibilities are not mutually exclusive as it is certainly

possible that very fine eye adjustments rely on proprioceptive feedback. Moreover, the fact that

the global MIF are associated with the PE, the putative EOM proprioceptors, makes the

proprioceptive hypothesis viable. Two questions that remain are: 1) why is proprioceptive

feedback subject to gamma modulation when the localization response is associated with the

vergence but not with the saccadic system?, and more generally 2) what is the role of the gamma

system in oculomotor control?

The fact that the saccadic system was not affected by the JM perturbation in our study is

analogous to the findings of Guthrie and colleagues (1982) who reported that cutting the

monkeys’ ophthalmic branch of the trigeminal nerve (i.e., deafferentation) altered their vergence

responses but had no effect on their conjugate eye movements. Fusional vergence involves

disjunctive eye movements which are driven by disparity (i.e., the eyes move in opposite

direction when the stimulus falls on non-corresponding retinal points in order to avoid double

vision); thus, vergence eye movements require precise adjustment of both eyes to foveate the

target and maintain single vision. The CNS might monitor proprioceptive feedback from the

EOM for optimal performance in this task. In brief, our results reinforce the previous findings

and emphasize the importance of the EOM proprioceptive feedback loop for binocular function.

The modulation of proprioceptive feedback by the gamma system has been studied

extensively in the case of the skeletal system; in contrast, only one study, to our knowledge,

examined the gamma system in the EOM. Whitteridge (1959) demonstrated that proprioceptive

feedback from the EOM in the ungulate is modulated by the gamma system. Direct experimental

evidence of the gamma system in other species has been precluded by the lack of information

about the afferent pathway and location of the cell body. Nonetheless, we have used an indirect

87

method to change the excitability of the gamma system by using the JM manipulation. Using this

proxy method we have provided evidence to support the notion that proprioceptive feedback

from the human EOM is also subject to gamma modulation.

It has been proposed that the gamma system is important for parametric adjustment of the

proprioceptive feedback loops to match the demands of different tasks (Prochazka, 1989). For

example, in the case of the skeletal system an increase in the sensorimotor gain is associated with

the performance of difficult or novel tasks in contexts that evoke generalized arousal and

vigilance. In other words, the gamma system allows state-dependent adjustment of

proprioceptive feedback which can be adjusted to address the specific sensorimotor requirements

predicted for the upcoming movement. An example of the parametric feedback adjustment for

the jaw muscles has been provided by Taylor and Gottlieb (1985). They suggested that the gain

of proprioceptive feedback might depend on the phase of the jaw movement. For instance, the

control of the velocity and displacement of the jaw are critical until the moment of tooth contact

whereupon the control of the force becomes critical. Moreover, they suggested that the CNS can

use proprioceptive feedback to determine the nature of the controlled variable (i.e., velocity or

force).

The requirement for gamma modulation of feedback from the EOM might be different

for saccades and vergence eye movements. Saccades are fast, ballistic eye movements ranging in

amplitude from 3 min arc to 90 and lasting between 15 to 100 ms. Programming of a saccade

involves a pulse and a step, which are related to the velocity and the amplitude of the eye

movement. In contrast, fusional vergence responses are slow (up to 1 sec) and generally small.

Vergence angle changes about 14 when gaze is moved from infinity to approximately 25 cm

(Howard, 2002). Even though horizontal saccades and vergence are both driven by the medial

88

and lateral recti muscles, the differences in neural control for version and vergence eye

movements are reflected in the premotor input and the activity of the motoneurons. For example,

for conjugate adduction the premotor excitatory input comes from the abducens internuclear

neurons. In contrast, premotor input for convergence eye movements comes from supraocular

motor area (Mays, 1984). In addition, Mays and Porter (1984) reported that the relationship

between eye position and motoneuron firing rate is dependent on whether the eye movements are

conjugate or disconjugate. In their study, recordings were made from the abducens nucleus

during conjugate adduction and during convergence. Data showed that for a given eye position

the firing rate was greater for convergence compared to conjugate adduction suggesting that

there would be greater co-contraction in convergence.

Miller et al. (2002) tested this hypothesis by measuring the oculorotary forces in the

medial and lateral recti muscles during both types of eye movements. In contrast to the

hypothesis, they found decreased forces in both muscles. These results showed that the

innervation of the EOM is much more complex than previously acknowledged and that the motor

commands sent to the eye muscles differ during convergence and adduction. Given our results,

we propose that that the CNS can also set the gain of the proprioceptive feedback differently for

vergence and saccades via the gamma system.

In conclusion, we have examined whether registered eye position during saccadic eye

movements is affected by the JM manipulation which alters the excitability of the gamma

system. We have shown that the JM does not affect manual or perceptual localization of targets

presented in the frontal plane. Overall, data from the present study help to clarify findings from

our previous study which examined pointing in depth and strengthen our hypothesis that the JM

affects the activity of the non-twitch motoneurons. We propose that the non-twitch motoneurons

89

might be involved in the parametric adjustment of the proprioceptive feedback loops to match

the demands of different types of eye movements.

90

PAPER 3

MANIPULATION OF EXTRAOCULAR MUSCLE AFFERENCE

HAS NO EFFECT ON HIGHER ORDER PERCEPTUAL JUDGMENTS

1E. Niechwiej-Szwedo, E.G. González, B. Bahl, M.C. Verrier, A. M. Wong, M.J. Steinbach

Vision Research (2007); 47(26):3315-3323

1 See Appendix 1 for contributions of each author

91

ABSTRACT

Observers perceive targets as farther while performing the Jendrassik Maneuver (JM)

suggesting that eye position is registered as more divergent. We examined the effects of the

JM perturbation in three studies of perceptual judgment that rely on accurate registration of

absolute distance: size constancy, stereoscopic depth, and the magnitude of the Pulfrich

illusion. The data showed no significant differences between the JM and control conditions.

The lack of an effect may be due to the fact that vergence is not a perfect cue to distance.

Furthermore, the relative contribution of extraocular muscle afference to registered eye

position may be less significant for higher order perceptual judgements.

92

1.0 Introduction

Accurate estimation of egocenteric distance is not only critical for the performance of

reaching and grasping movements, but perceptual constancies, such as size and depth also

rely on accurate registration of absolute distance. The central nervous system (CNS) can

obtain depth information from static and dynamic monocular and binocular cues (Howard &

Rogers, 2002). Ocular vergence is an extraretinal binocular cue which has been shown to

provide crude but reliable distance information in a visually impoverished environment (for

reviews see Collewijn & Erkelens, 1990, Foley, 1980;).

Information about vergence eye position can come from two sources: the efference

copy (outflow) and afferent feedback (inflow) from the eye muscles (Steinbach, 1987). There

are at least two receptors in the extraocular muscles (EOM) that could provide proprioceptive

information about eye position: muscle spindles and palisade endings (PE) (for a review see

Donaldson, 2000). Muscle spindles have been unequivocally shown to provide

proprioceptive information from skeletal muscles; however, their role in EOM is not as clear.

First, muscle spindles are only found in the orbital layer of the EOM and they are

morphologically different from the spindles found in skeletal muscle (for a review see

Ruskell, 1989). Second, several species, such as cat, rabbit, horse, and mouse do not have

muscles spindles in their EOM (Maier et al., 1974). In contrast, PE have been found in the

EOMs of all the species tested to date, such as cat, rhesus monkey, sheep, rat, and human

(Alvarado-Mallart & Pinçon-Raymond, 1979; Blumer et al., 1998; Buttner-Ennever et al.,

2001; Eberhorn, Horn, Eberhorn et al., 2005; Richmond et al., 1984). PE are also referred to

as innervated myotendious cylinders (Ruskell, 1978) and they are uniquely associated with

the multiply innervated non-twitch fibers (MIF) of the global layer of the eye muscles.

93

Neuroanatomical tracing studies by Buttner-Ennever and colleagues (2001) have

shown that the MIF receive innervation from a distinct set of non-twitch motoneurons found

in the periphery of the twitch motoneurons that control eye movements. The authors

hypothesized that these non-twitch motoneurons could be involved in modulating the gain of

sensory feedback from the extraocular muscles, analogous to the gamma (fibers which

control the sensitivity of muscle spindles in the skeletal muscles.

Our previous study (Niechwiej-Szwedo et al., 2006) tested the above hypothesis using

a behavioural approach and a manipulation called the Jendrassik maneuver (JM). The JM

refers to a voluntary, forceful contraction of any muscle group. While the JM is performed,

the amplitude of skeletal reflexes is facilitated (Delwaide & Toulouse, 1981; Murthy, 1978).

One of the mechanisms proposed to explain the reflex reinforcement effect is that the muscle

contraction has a general effect that results in up-regulation of the motoneuron activity

which increases the baseline excitability of muscle spindles and, consequently, results in a

larger response when the muscle is stretched. We hypothesized that the JM would also alter

the gain of the afferent feedback from eye muscles which would result in misregistration of

eye position and localization errors.

Altering the feedback from the eye muscles during vergence eye movements via the

JM resulted in misregistration of eye position. In particular, when the JM was performed, eye

position was registered as more divergent while the actual eye position did not change

(Niechwiej-Szwedo et al., 2006). Based on these results we hypothesized that the JM would

also alter higher order perceptual judgements that rely on accurate registration of absolute

distance. This hypothesis was tested in three experiments.

94

In the first experiment, size constancy was examined while feedback from extraocular

muscles was perturbed by the JM. Since the vergence angle of the eyes is an important

source of extraretinal information contributing to size constancy, we hypothesized that

participants would perceive the size of a constant retinal stimulus as larger when feedback

from the eye muscles was altered via the JM.

Stereoscopic depth constancy was examined in the second experiment. Horizontal

disparities must be scaled by viewing distance in order for depth constancy to be preserved

and the vergence angle of the eyes can be used to calibrate horizontal disparities for different

viewing distances. We hypothesized that for the same disparity, the perceived depth would

be greater when the JM is performed compared to the control condition without JM.

In the third experiment we examined whether perceived depth during the Pulfrich

illusion was affected by the JM. In the Pulfrich effect a pendulum objectively swinging in the

frontal plane appears to move in an elliptical orbit in depth. The effect results from the

cortical time delay, interpreted as a disparity induced when one eye views it through a neutral

density filter. It has been shown that the perceived depth (i.e., the short axis of the ellipse) is

dependent on the viewing distance, so we hypothesized that the perceived depth would be

greater while participants perform the JM.

All participants in the three studies had normal or corrected-to-normal visual acuity of

20/20 and stereopsis of at least 40 seconds of arc as measured with the Titmus test (Titmus

Optical Co., Inc., Petersburg, Virginia 23805). All experimental protocols were approved by

the Ethics Review Boards at the University of Toronto and the University Health Network

and participants gave their informed consent prior to participating.

95

2.0 Experiment 1: The effect of JM on perceived size

Even though the size of the image projected on the retina changes substantially over a

range of viewing distances, observers perceive the size of an object as relatively constant.

This is referred to as size constancy. There are three laws which describe the relationship

between the object size, image size and distance: 1) for a constant object size, the image size

varies inversely with distance, 2) for a constant image size, object size is proportional to

distance, and 3) image size is proportional to object size for an object presented at a fixed

distance (Howard & Rogers, 2002).

Observers are remarkably good at judging object size accurately in natural

environments where the CNS can use multiple cues to distance. As depth cues are reduced so

is the degree of size constancy and observers tend to rely more on the retinal image size to

make judgements (Ono, 1966). Ocular vergence and accommodation are the only cues that

the CNS can use in an unstructured visual environment to judge the size of unfamiliar

objects; however, vergence is only reliable when the distance to the stimulus is less than 2 m

(Harvey & Leibowitz, 1967; Leibowitz & Moore, 1966; Wallach & Floor, 1971).

The contribution of oculomotor cues to size perception within near visual space was

shown by Wallach & Zuckerman (1963). Participants were asked to judge the size of a wire-

form pyramid while vergence and accommodation were altered using mirrors and lenses. The

size estimates obtained experimentally varied accordingly with the changes in oculomotor

cues, thus confirming that the perception of size relies on these cues.

More recently, Mon-Williams and colleagues (Mon-Williams, Tresilian, Plooy, Wann

& Broerse, 1997) examined the role of vergence in explaining the illusory size change of an

afterimage (i.e., Emmert’s law, which states that the perceived size of an afterimage is

96

dependent on the perceived distance to the surface). Participants judged the vertical size of an

afterimage in two conditions. In the control condition, the card on which the afterimage was

created was moved by the participant but fixation was maintained on a stationary light

emitting diode (LED). In the experimental condition participants made converging or

diverging eye movements when judging the size of the afterimage. The experimental results

clearly supported the hypothesis that vergence is necessary and sufficient to explain the

illusory change in the size of the afterimage. Specifically, converging eye movements were

associated with a smaller perceived size of the afterimage whereas diverging eye movements

led to reports of a larger afterimage.

Experiment 1 was designed to examine whether the perceptual phenomenon of size

constancy was affected by the JM which has been shown to affect the registered vergence

eye position. We employed a two-alternative forced choice paradigm (2 AFC) and the

method of constant stimuli. The perturbation (JM) occurred when either the standard or the

comparison stimuli were shown. We hypothesized that the order of the JM would affect the

size judgement. Two specific predictions were made for the case when both the standard and

comparison stimuli had the same retinal size: 1) if the JM were performed while participants

viewed the standard stimulus, the comparison would be perceived as smaller, and 2) if the

JM were performed while viewing the comparison stimulus, participants would report it as

larger. Eye movements were not recorded in this study because our previous work had shown

that the actual vergence eye position was not affected by the JM (Niechwiej-Szwedo et al.,

2006). In addition, a pilot study using the current methodology measured the eye movements

of three participants and found no differences between the JM and control conditions.

97

2.1 Methodology

2.1.1 Participants

Twenty healthy adults (9 males) with no history of ocular disorders and a mean age of

28±13 years participated.

2.1.2 Stimuli

The initial fixation stimulus was a green LED controlled by a custom-made trigger

box. The LED was placed at a distance of 25 cm in front of the participant in the midline.

The height of the LED placement was adjusted for each participant individually to prevent

obstruction of the stimulus.

The standard stimulus for the psychophysical procedure was a grey square (4.7

visual angle) presented on a black background and displayed on a flat CRT monitor (refresh

rate 85 Hz). The viewing distance was 100 cm. There were five comparison stimuli: 4.5,

4.6, 4.7, 4.8 and 4.9. The display was programmed using VPixx (VPixx Technologies,

Inc., Montreal, QC), a graphics generation and psychophysics testing software, controlled by

an Macintosh iBook computer.

2.1.3 Apparatus

The JM consisted of an isometric voluntary contraction against resistance with the

abductor muscles of the legs. The device used for resistance was a Thigh Master™.

Participants were asked to perform each contraction at a 75% level of their maximal

voluntary contraction, which was determined prior to the initiation of the experiment. To

ensure that the isometric contraction was performed at a consistent level throughout the

98

experiment, a string was tied around the Thigh Master™ which pulled taut when the muscle

contraction was executed. Participants were instructed to hold the string taut while

performing the JM.

2.1.4 Procedure

Participants were seated in complete darkness. At the beginning of each trial they

fixated on the LED for 2.5 sec in order to standardize the initial vergence eye position. As

soon as the LED was switched off, the standard stimulus was presented for 2.5 sec and

participants were asked to fixate on it and to remember its size when it disappeared. The

comparison stimulus was shown at the same location 1.5 sec after the standard had

disappeared. Five sizes of the comparison stimulus were tested and their presentation order

was determined randomly by the computer. On each trial participants were asked to report

whether the comparison stimulus was ‘smaller’ or ‘larger’ than the standard. The comparison

stimulus disappeared after participants made the judgment. There were three experimental

conditions: Control: the judgment task was performed without the JM; Task 1: participants

performed the JM while viewing the LED and the standard stimulus and relaxed the

contraction before the comparison stimulus was presented (verbal judgment was made

without the JM); Task 2: participants fixated the LED and the standard stimulus without the

JM and started the JM when the standard stimulus disappeared (verbal judgement was made

with JM). The protocol is illustrated in Figure 12. The five comparison stimuli were tested 10

times in each experimental condition for a total of 150 trials. All experimental conditions

were completely randomized.

99

Figure 12: Schematic illustration of the experimental protocol used in experiment 1.

2.1.5 Data Analysis

The proportion of ‘smaller’ responses was calculated and plotted for each participant

and condition for the five sizes of the comparison stimuli. Data was visually inspected for

trends and then fitted with a psychometric function using a logistic regression (SAS, Ver.

8.1). The goodness of fit of the model was tested using the Hosmer-Lemeshow statistic and a

non-significant result was used to verify that the logistic model was appropriate.

Subsequently, an overall psychometric function based on the means of all participants was

fitted for each task.

For each participant, the point of subjective equality (PSE) and the just noticeable

difference (JND) were calculated using the estimated parameters (slope and intercept) from

100

the logistic model. The PSE is the point at which the logistic function yields a probability of

0.5 (i.e., the comparison stimulus is perceived as smaller than the standard stimulus 50% of

the time) and it reflects the accuracy of the judgment. The JND is the smallest possible

physical difference that can be detected reliably and it reflects the precision of the judgment.

The PSE, JND, intercept, and slope were submitted to a one-way repeated measures ANOVA

each with condition (control, task 1, task 2) as the independent variable.

The magnitude of the illusion (i.e., the proportion of ‘smaller’ responses) was

examined for the condition in which the comparison stimulus was the same size as the

standard. Data for each participant and condition were submitted to a one-way repeated

measures ANOVA with condition (control, task 1, task 2) as the independent variable.

2.2 Results

Preliminary inspection of the individual psychometric curves did not reveal consistent

differences between the conditions. The data were collapsed and the mean performance of all

participants in each condition is shown in Figure 13. The mean psychophysical curves clearly

show that participants were able to judge the size of the comparison square accurately in each

of the conditions.

The logistic model fitted the experimental data well for the majority of the psychometric

curves (59 out of 63), which was supported by the non-significant result from the Hosmer-

Lemeshow test. Although in four cases (one in Task 1 and three in Task 2) the test was

statistically significant, the logistic model was still used to fit the data. In contrast to the

hypothesis, no significant differences were found for any of the variables: PSE

(F(2,38)=2.61, n.s.), JND (F(2,38)=1.52, n.s.), slope (F(2,38)=1.53, n.s), y-intercept

101

(F(2,38)=1.21, n.s). The magnitude of the illusion was not significantly different at the point

where both the standard and comparison stimuli were physically the same (F(2,38)=1.20,

n.s.).

Figure 13: Mean proportion responding ‘smaller’ for each of the five sizes of the comparison

square (at 0 both the standard and comparison squares were the same physical size, negative

values indicate that the comparison square was smaller). Error bars show 1 standard error.

102

2.3 Discussion

It was expected that the order of JM would affect the size judgments; however,

neither the accuracy nor the precision of the perceptual judgment were affected by the

perturbation. A potential weakness of the experiment is that participants were aware that the

stimulus was always presented on a flat monitor screen at a fixed distance. Previous research

has shown that observers tend to make judgments of distal (object) or proximal (image) size

depending on the experimental conditions. For example, when no specific instructions were

given and with unrestricted viewing, observers tended to judge the distal size. In contrast,

when all visual cues were eliminated and viewing was monocular, observers judged the

proximal size (Ono, 1966). When only binocular cues were present, observers also tended to

judge the distal size and size constancy was preserved, at least up to 30 feet (Chalmers,

1952). Although the current experiment was conducted in the dark and no other visual cues

were available, participants could have relied on the oculomotor cues of vergence and

accommodation. Thus, it is likely that participants used distal size to make judgements in the

current study.

3.0 Experiment 2: The effect of JM on stereoscopic depth judgments

The perceptual phenomenon of depth constancy is conceptually similar to size

constancy and refers to the ability of the observer to judge the linear extent of a stimulus in

the saggital plane accurately despite changes in viewing distance (Ono & Comerford, 1977).

Depth constancy depends on the accurate registration of disparity and fixation distance.

When two images are separated in depth, they fall on non-corresponding or disparate retinal

points, which is the bases of stereopsis. By convention, points that are nearer to the observer

103

than the fixation point have crossed disparity, conversely, points farther than the fixation

point have uncrossed disparity. Stereopsis is one of the cues contributing to depth perception

but it can only provide relative information. Specifically, a given disparity can be associated

with different depth intervals depending on the fixation distance. For example, a disparity of

50 min arc will be interpreted as a depth interval of 5 cm when viewed at 50 cm or a depth

interval of 20 cm if viewed at 100 cm. Likewise, a 5 cm depth interval viewed at 100 cm will

have disparity of 12 min of arc and the disparity will be 4 times larger when viewed from 50

cm away. To sum up, for a constant physical depth, retinal disparity decreases in proportion

to the square of the absolute distance (Ono & Comerford, 1977). Therefore, in order for

depth constancy to occur, the CNS must take into account the fixation distance or, in other

words, horizontal disparity must be calibrated for different fixation distances.

Wallach and Zuckerman (1963) were among the first to empirically examine whether

changes in vergence and accommodation contribute to depth constancy. In their experiments

the oculomotor cues were altered by optical means and the obtained depth estimates

approximated those that were predicted by the inverse square law. A detailed examination of

vergence contribution to depth constancy was reported by Ritter (1977). All distance cues

except for convergence and accommodation were removed and participants viewed a

stereoscopically presented image of a pyramid at different viewing distances. Results showed

that in the case of vergence-accommodation conflict, the depth interval was perceived based

on the convergence distance.

Experiment 2 was designed to examine whether the perceptual phenomenon of depth

constancy could be modified by the JM. Participants were asked to judge the separation in

depth between two lines when the registered vergence eye position was perturbed by the JM.

104

Since during the JM eye position is registered as more divergent, we hypothesized that

participants would perceive the depth interval as larger while the JM was performed as

compared to the control condition. Vertical lines were used instead of random dot

stereograms to reduce the likelihood of participants using vertical disparities to judge depth.

3.1 Methodology

3.1.1 Participants

Five healthy adults with no history of ocular disorders, mean age 27.26.6 years,

participated.

3.1.2 Stimuli

The stimuli consisted of two vertical white lines subtending 1.5° by 0.1° of visual

angle, which were presented on a black background inside a white square outline subtending

5° x 5° of visual angle. The stimuli were viewed on a flat CRT monitor (Viewsonic, refresh

rate 60 Hz) at a viewing distance of 57 cm. The stereo images were displayed with 5 crossed

disparities: 1.17, 2.34, 3.51, 4.68, 5.85 min of arc. A rating scale consisting of 10 horizontal

lines ranging in length from 1 mm to 10 mm was displayed at the end of each trial and

numbers from 1 to 10 were displayed above the corresponding horizontal line.

3.1.3 Apparatus

The stimulus presentation was controlled by VPixx (VPixx Technologies, Montreal,

QC), a graphics generation and psychophysics testing software, controlled by Macintosh G4

computer. The stereo images were seen using liquid crystal glasses (CrystalEyes

105

Workstation, Stereographics, San Rafael, CA). The JM procedure was the same as in

Experiment 1.

3.1.4 Procedure

Participants saw the display with the room lights turned off. At the beginning of each

trial, they fixated on a standard vertical line presented for 1 sec. A second, or comparison,

vertical line appeared to the left of the first one with a variable crossed disparity and for a 2

second duration. Participants were instructed to remember the separation in depth between

the two vertical lines. After the comparison stimulus disappeared, participants were shown 10

horizontal lines and were asked to estimate the distance in depth between the two vertical

lines by choosing one of the horizontal lines. Participants made a verbal response indicating

the number (1-10) corresponding to the depth interval that they saw between the two vertical

lines. They were not informed that only five disparity stimuli were used. Participants

completed 10 trials for each stimulus disparity with and without the JM for a total of 100

stereoscopic depth judgments. For the JM trials participants started the isometric contraction

prior to seeing the stimulus with the disparity and held it while viewing it. Prior to data

collection, all participants completed 20 practice trials to become acquainted with the task.

3.1.5 Data Analysis

For each disparity value the mean perceived depth was calculated and plotted for all

participants. Data were fitted using a linear regression model. The slope and y-intercept

parameters obtained from the model were submitted to a paired Student’s t-test with

condition (control, JM) as the dependent variable.

106

3.2 Results

The mean responses of all participants ranged in values from 2 to 7; thus, participants

used the middle range of the scale and responses at the limits were not frequent. The

individual curves showed that participants could judge the depth difference reliably, which is

evident in the slope values and the measure of goodness of fit of the model (Table 2). A

paired samples t-test showed no significant differences between the control condition and the

JM condition for the slope or y-intercept values. The lack of difference is illustrated in Figure

14 which shows the mean data of all participants.

Table 2: Parameters obtained from the linear regression model for individual participants

Participant Model fit (R2) Slope Y- intercept

value

Con

trol

JM Con

trol

JM Con

trol

JM

1 0.92 0.90 1.11 0.99 0.59 1.47

2 0.87 0.75 0.78 0.58 2.46 3.36

3 0.97 0.97 0.67 1.04 2.21 0.94

4 0.85 0.71 1.02 0.94 1.30 1.76

5 0.88 0.55 1.36 0.86 0.60 0.76

107

Figure 14: Mean perceived depth for stereoscopically presented stimuli in experiment 2.

Error bars show 1 standard error.

3.3 Discussion

The experimental results did not confirm our hypothesis and showed that JM did not

affect judgments of stereoscopic depth. The CNS must use oculomotor cues or vertical

disparities in order for stereoscopic depth constancy to be preserved. The disparity stimulus

in the current study consisted of vertical lines presented at the midline so no vertical

disparities were present in the field of view and the CNS must have relied on the only

available cues, which were the vergence and accommodative state of the eyes.

We chose to examine the effect of JM on depth constancy by presenting the stimulus

using stereoscopic goggles, which allowed us to precisely control the disparity. However,

108

stereoscopic presentation has a disadvantage: the oculomotor cues of vergence and

accommodation are in conflict (Ono & Comerford, 1977). The disparity stimulus which

drives the vergence system creates an illusion of depth, but there is no retinal blur and the cue

of accommodation informs the system that both stimuli are at the same distance. Ritter

(1977), however, showed that in a case of vergence-accommodation conflict the judgments

of perceived distance are based on the cue of convergence, so it is unlikely that the mismatch

between these cues contributed to the present findings. The next study was designed to

further examine the effect of JM on depth constancy using the Pulfrich phenomenon for

which there was no conflict between oculomotor cues.

4.0 Experiment 3: The effect of JM on perceived depth during the Pulfrich illusion

In the Pulfrich illusion, a pendulum moving sinusoidally in the frontoparallel plane

appears to move along an elliptical path plane when viewed through a neutral density filter

placed over one eye. Placing the filter in front of one eye creates a luminance difference

between the two eyes which leads to a temporal delay in transmitting visual information to

the cortex. The cortical time delay is interpreted by the CNS as binocular disparity of the

moving object between the images seen by the two eyes (Howard & Rogers, 2002).

The effect of fixation distance on the magnitude of perceived depth during the

Pulfrich illusion was studied by Lit and Hyman (1951). They systematically investigated

whether variables such as differences in illumination, distance to target, and velocity of the

target influenced the magnitude of the stereoscopic depth effect. Their results clearly showed

that for any given illumination difference value, the pendulum’s motion depended on the

fixation distance with the largest depth interval observed for the largest fixation distance and

109

greater illumination difference between the two eyes. These results were later replicated by

Wallach and colleagues (Wallach, Gillam & Cardillo, 1979).

More recently, Nakamizo and Lei (2000) examined the magnitude of the illusion at

larger viewing distances raging from 1 m to 4 m for stimulus velocities of 0.2, 0.4, and 0.6

Hz. Two procedures for measuring depth were used: participants had to match the perceived

depth interval using a probe or to reproduce the depth interval using a tape measure.

Although there were no significant differences between the two response methods (matching

and reproduction), the matching response produced depth estimates that were closer to those

that would be expected if the Pulfrich effect increased in direct proportion to viewing

distance.

In summary, previous studies have shown that the perceived depth of the Pulfrich

effect depends on the viewing distance and vergence angle of the eyes; therefore, experiment

3 was designed to examine whether the perceived depth during the Pulfrich illusion is also

affected by the JM. Participants viewed the moving stimulus through a pair of different filters

placed in front of the two eyes: one of the filters was constant while the other was adjusted

by the participant. Participants were asked to null the apparent depth by adjusting one of the

filters. We hypothesized that during the JM the depth interval would be perceived as larger

and that in order to null the illusion participants would compensate by over-adjusting the

variable filter.

110

4.1 Methodology

4.1.1 Participants

Five healthy adults with no history of any ocular disorders, mean age 47.213.5

years, participated in experiment 3.

4.1.2 Stimuli

The fixation stimulus was a 0.25 black dot and the target stimulus was a black

vertical 1.5 bar both displayed on a white background on a flat CRT monitor (Viewsonic,

refresh rate 85 Hz). The target stimulus moved sinusoidally in the frontoparallel plane at a

peak velocity of 15/sec. The display was programmed using VPixx (VPixx Technologies,

Inc., Montreal, QC) and controlled by a Macintosh G4 computer.

4.1.3 Apparatus

Participants viewed the vertical bar through a custom-made apparatus that contained

two round apertures. Participants were seated behind the apparatus and the height of the chair

was adjusted individually. A variable neutral density (ND) filter (luminance values ranging

from 2.5 cd/m2 to 102.5 cd/m

2) mounted on a movable wheel, was placed over the right

aperture. The density of the variable filter could be adjusted by turning a knob. A 360

protractor was attached to the movable wheel so that the responses could be read out with an

accuracy of 1/10 of a degree. Three constant, non-adjustable ND filters (0.2, 0.7, and 1.0 log

units) were used during the experiment and placed over the left aperture.

The JM procedure was the same as in Experiments 1 and 2.

111

4.1.4 Procedure

At the beginning of each trial participants closed their eyes. The experimenter spun

the wheel with the variable ND filters to vary the initial value of the filter between trials. One

of the three ND filters, randomly determined prior to the experiment, was placed to cover the

left aperture. Participants opened their eyes when cued by the experimenter and viewed the

moving bar while fixating the dot. The moving target was viewed through different filters

over each eye which produced an illusion of elliptical movement (Pulfrich effect).

Participants were asked to null the illusion by adjusting the variable ND filter with the

movable knob. They were allowed as much time to make the adjustment as they needed to

make sure that the elliptical movement of the target stimulus disappeared. Once the

participant indicated that the illusion had disappeared, the experimenter recorded the

response which was the number indicated by the protractor. The task was performed while

participants performed the JM and without the JM and these two conditions were randomized

prior to the experiment. Participants completed five trials in each experimental condition for

a total of 30 trials.

4.1.5 Analysis

Data for individual participants were plotted for the three values of ND filters for the

two JM conditions and inspected visually for trends. Subsequently, data were submitted to 2-

way repeated measures ANOVA with two factors: condition (control, JM) and filter value

(0.2, 0.5, 1.0 log units).

112

4.2 Results

Preliminary inspection of the individual response curves did not reveal consistent

differences between the experimental and control conditions. The data were collapsed and

the mean performance of all participants in each condition is shown in Figure 15. As

expected, analysis of variance (ANOVA) yielded a significant main effect of filter

(F(2,140)=541.68, p<0.0001) showing that the perceived depth interval varied across the

Figure 15: Mean values obtained when participants were asked to null the Pulfrich illusion by

adjusting the value of the variable filter. Error bars show 1 standard error.

113

three filter values. Specifically, the largest depth interval was perceived with ND filter 1.0

which created the largest luminance difference between the two eyes. In contrast to our

hypothesis, there were no significant differences between the control and the JM condition

(F(1,140)=3.43, n.s.) and the interaction effect was also not significant (F(2,140)=0.36, n.s.).

4.3 Discussion

Although the Pulfrich effect has been shown to depend on viewing distance, the

current study showed that it was not affected by the JM perturbation. Overall, the results

from experiments 2 and 3 showed that the JM had no effect on the perceptual phenomenon of

depth constancy.

5.0 General Discussion

The purpose of our studies was to examine whether higher order

perceptual judgments which require accurate registration of absolute distance are

affected by a manipulation which we have shown alters the gain of the

proprioceptive feedback from the EOM. Contrary to our hypotheses, we found

that the JM manipulation did not significantly affect judgments of size, depth or

the Pulfrich illusion. These are important findings as they help to establish that

proprioceptive feedback plays a negligible role in maintaining the perceptual

phenomena of size and depth constancy.

Previous studies have shown that proprioceptive signals from the eye

muscles play a significant role in the programming of eye movements (see section

1.1.2.2, Knox et al., 2000; Weir & Knox, 2001), during egocentric localization

114

tasks (Bridgeman & Stark, 1991; Gauthier et al., 1990a,b; Lewis & Zee, 1993;

Roll et al., 1991; Steinbach, 1987; Velay et al., 1994) and adaptation of smooth

pursuit (vanDonkelaar al., 1997).

Two methods have been used to manipulate EOM afference in binocularly intact

observers: a vibration stimulus applied over the muscle and passive deviation of the eye

using a suction lens. Vibration provides a good stimulus for activating the Ia afferents which

in skeletal muscles activate the tonic vibration reflex (i.e., contraction of the vibrated

muscle). Roll and colleagues (1991) applied vibration over the inferior rectus muscle while

subjectes were fixating a single light in the dark. During the vibration trials subjects reported

that the target moved up and they also pointed above the target. Similar results were obtained

by Velay et al. (1994) where vibration of the right lateral rectus muscle resulted in an illusory

movement of the target to the left. Overall, results from these studies suggest that vibration of

an EOM muscle leads to the perception that the muscle is lengthening, and participants report

that the target is moving in a direction opposite to that of the vibrated muscle.

The afferent signals from eye muscles can be also altered by passively moving the

eye using a suction lens. This method, introduced by Gauthier and colleagues (1990a), offers

a way of distinguishing the contributions of inflow and outflow to registered eye position.

The paradigm involves subjects fixating a target with one eye while the other eye is

occluded. In the first experimental condition the fixating eye is deviated, thus, the amount of

innervation sent to the muscles must be increased in order to maintain fixation. Since both

eyes receive the same amount of innervation during conjugate eye movements (Hering’s

law), the occluded eye should deviate by an amount corresponding to the efferent signal sent

to the fixating eye. In this condition, the efferent signal to the eye muscles must be increased

115

to compensate for the perturbation, but the afferent feedback from the fixating eye is not

changing because the eye is not changing position. Therefore, this task allows one to examine

the effect of efference on registered eye position. The second experimental condition

involves passive deviation of the occluded eye. In this case, the amount of innervation does

not change, but the afferent feedback from the deviatated eye does change. Therefore, the

second condition allows one to examine the contribution of afference to regitstered eye

position.

The current study is the first examination of the role of afference in higher order

perceptual phenomena using the JM manipulation to alter the feedback from the eye muscles.

The JM has been used extensively to alter the excitability of spinal reflexes (Dowman &

Wolpaw, 1988; Gregory et al., 2001; Murthy, 1978; Zehr & Stein, 1999) and limb position

information (Yasuda et al., 2006). Our previous studies (Niechwiej-Szwedo & Steinbach,

2007) were the first to show that the JM can also be used to alter proprioceptive feedback

from the EOM. Specifically, we showed that participants made consistent perceptual errors

when localizing targets in depth during the JM while the actual eye position was not affected.

Thus, we expected that perceptual judgments that require accurate registration of depth

would be also affected by the JM. This was not confirmed by the results from the three

experiments.

The perceptual phenomena of size and depth constency depend on the preceived

distance, which is an internally generated estimate of the viewing distance. In the real world,

the neural estimate of viewing distance is based on multiple visual and oculomotor cues. In

the present experiments most visual cues were removed and oculomotor cues provided the

only input for distance estimation; nontheless, the perturbed vergence signal was not taken

116

into account by the CNS. This result could be explained by considering the relative

contribution of efference and afference to registered eye position examined by Gauthier and

colleagues (1990a) and Bridgman and Stark (1991). Both studies used the passive eye

deflection paradigm (described above) and found significant localization errors with open-

loop pointing responses, which were always correlated with the direction of the deviated eye.

However, the localization errors were only found when the occluded eye was deviated by a

large amount (>10º). Based on the localization errors and using a regression analysis, these

authors calculated that the contribution of proprioception from the eye muscles to the

registered eye position was approximately 30%, under their experimental conditions. It is

possible that the afferent contribution from the EOM to judgments involving size and depth

constancy or the Pulfrich illusion is even less significant and is not actually used for higher

order perceptual judgments by the CNS. Alternatively, it is also possible that the perturbation

that we are using, i.e., the JM, is not large enough to disrupt perceptual constancies2. In short,

the lack of significant effect was most likely due to a combination of factors, such as the fact

that vergence is not a perfect cue to distance and the JM manipulation is not a strong

petrubation of the feedback from the extraocular muscles. In addition, perceptual constancies

rely on multiple cues and are not easily perturbed, therefore, a strong manipulation might be

necessary to definitively determine whether the afferent feedback from the eye muscles plays

a role in maintaining perceptual constancies.

In summary, results from the present study showed that altering feedback from the

EOM via the JM did not affect perceptual judgments of size or depth. The lack of a

significant effect might not be surprising given that the JM manipulation affects the

registered vergence eye postions, but vergence itself is not a perfect cue to distance. Overall,

2 We thank an anonymous reviewer for this suggestion.

117

the role of afference in oculomotor control and visuomotor behavior is not well understood

yet; however, the current study is the first to report that perturbations of afferent input from

the extraocular muscles do not affect higher order perceptual judgements.

118

PAPER 4

LOCALIZATION IN DEPTH IS NOT AFFECTED BY THE JENDRSSIK

MANEUVER IN PATIENTS OPERATED FOR STRABISMUS: CASE STUDIES

119

ABSTRACT

Strabismus is a disorder involving misalignment of the eye(s). Previous studies have reported

that surgery, which is a common treatment to correct the misalignment of the eyes, can

damage proprioceptive receptors in the eye muscles and patients might be left deafferented.

The present study tested whether surgically treated patients would be affected by the

Jendrassik Maneuver (JM) which was shown to alter registered eye position when

binocularly healthy observers localize targets in depth. It was hypothesized that the patients’

responses would not be affected by the JM perturbation because the surgeries most likely

compromised the extraocular muscle feedback loops. Data from patients with congenital

strabismus confirmed our hypothesis. A larger study should be conducted to determine

whether the lack of gain modulation was due to damage sustained from the surgical

intervention or whether patients had an abnormal proprioceptive apparatus prior to surgery.

120

Introduction

Strabismus is an ocular disorder which involves misalignment of the eye(s). It is a

heterogeneous disorder which can be divided into several subtypes based on the deviation of

the eyes. For example, esotropia occurs when the eyes deviate medially, and exotropia refers

to laterally deviated eyes. Strabismus may be constant (the amount of deviation is constant

over time), intermittent (the eye deviates under some conditions, such as stress or anxiety). It

may be comitant (amount of deviation does not vary with gaze direction) or incomitant

(amount of deviation varies with gaze direction) as in paralytic strabismus (i.e., due to EOM

paralysis) (Von Noorden & Campos, 2002). The prevalence of strabismus in the general

population has been estimated to be between 3 and 6% (Stidwill, 1997; Graham, 1974).

The major consequences of a strabismus that has an early onset are deficits in

binocular function and amblyopia. Binocular function requires sensory and motor fusion:

sensory fusion involves the ability to fuse the retinal images from both eyes into a single

percept, and motor fusion refers to the ability to align the eyes using vergence eye

movements to maintain sensory fusion (Von Noorden & Campos, 2002). In healthy

individuals, when both eyes fixate an object, the lines of sight from both eyes intersect on it

and the images fall on corresponding retinal points thereby giving rise to a single percept. In

strabismus, the eyes are misaligned and the two images of the object fall on non-

corresponding retinal points. In young children and over time, the brain suppresses the visual

input from one eye, which can result in reduced acuity for that eye (amblyopia), and lack of

binocular function (stereopsis and vergence). Deficits in binocular function in patients with

strabismus can lead to difficulties in the performance of tasks that involve eye-hand

coordination (Fronius & Sireteanu, 1994; Grant, Melmoth, Morgan & Finlay, 2007), spatial

121

localization (Weir, Cleary, Parks & Dutton, 2000) and saccadic conjugacy (Bucci, Kapoula,

Yang, Roussat & Bremond-Gignac, 2002; Kapoula, Bucci, Eggert, & Garraud, 1997).

Surgical intervention is often employed to correct strabismus. For example, for

esotropia, the surgical procedures are aimed at weakening the MR and/or strengthening the

LR (von Noorden & Campos, 2002). Recession of the MR muscle is performed to weaken

the muscle by changing the tangential force, thereby decreasing its rotational force on the

globe. It involves cutting the muscle tendon close to insertion and reinserting it on the sclera

posterior to the original insertion. Another procedure that is very effective in decreasing the

strength of the muscle is marginal myotomy, which is sometimes considered in cases of a

large deviation that cannot be corrected with maximal recession. Myotomy involves reducing

the number of contractile elements and it is rarely performed due to its irreversibility.

Resection of the LR muscle is performed to strengthen the muscle. It involves shortening the

length of the muscle by excising a portion of the tendon and reinserting it to the original

insertion.

A major goal of corrective surgery for strabismus is to obtain good alignment of the

eyes. The predominant view being that the potential for binocular function is greatest when

both eyes are aligned properly. The standard for the functional outcome of surgery is

stereopsis and fusional vergence eye movements, as well as symmetrical pursuit and OKN.

Unfortunately, these functional outcomes are frequently not obtained and as many as 50% of

children have to undergo multiple surgeries (Helveston, Neely, Stidham, Wallace et al.,

1999). It is possible that one of the factors contributing to the suboptimal outcome is the fact

that EOM afferent feedback loops are compromised during the surgery (Steinbach, 1987) .

122

The development of optimal surgical techniques is partly hindered by the lack of

understanding of how the CNS processes and uses proprioceptive information from the

EOM. There is ample experimental evidence which shows that feedback from EOM

contributes to registered eye position during visuomotor tasks (for a review see Donaldson,

2000). In addition, we demonstrated that healthy observers are susceptible to a manipulation

of afferent feedback via the Jendrassik Maneuver (JM) while localizing targets in depth

(Niechwiej-Szwedo et al., 2006). The JM changes the excitability of spinal and brainstem

reflexes, presumably via the activity of the gamma system. Thus, we proposed that the CNS

might modulate the gain of the afferent signal from EOM possibly via the activity of the non-

twitch motoneurons identified by Buttner-Ennever and colleagues (2001). It has been

hypothesized that the non-twitch motoneurons could be involved in modulating the gain of

sensory feedback from EOM, analogous to the gamma-efferent fibers which control the

sensitivity of muscle spindles in the skeletal muscles (Buttner-Ennever et al., 2002; Porter et

al., 1995; Robinson, 1991).

The present study examined the effect of JM on target localization in patients with

strabismus who have had surgeries that potentially compromised the EOM feedback loops.

We have previously reported that the JM altered registered vergence eye position in

binocularly healthy observers (Niechwiej-Szwedo et al., 2006), thus, we used the same

protocol to examine whether patients’ responses were also affected by the perturbation. It

was hypothesized that patients’ responses would not be affected by the JM perturbation

because the gain of sensory signal could not be altered via the non-twitch motoneurons due

to compromised EOM proprioceptive receptors and/or afferent pathway.

123

Methodology

Patients

Five patients (aged 16-55 years) who underwent surgical treatment for strabismus

were tested in the experiment. Due to the heterogeneity of the disorder, each patient is

described in detail in Table 1. All patients had a self reported binocular corrected visual

acuity of 20/20 or better. Patients were included in the study if they had binocular

stereoacuity of at least 140 sec arc, as determined by the Titmus test. The fact that patients

were required to have residual stereoactuity to participate in the experiment restricted the

number of eligible patients who were able to complete the study. The experimental protocol

was approved by the Ethics Review Boards at the University of Toronto and the University

Health Network and participants gave their informed consent prior to participating.

Experimental Procedure

The experimental procedure used in this study was exactly the same as that of

Experiment 2, in Paper 1 (described on pages 51 to 53 of this thesis). Briefly, patients were

tested using a two-alternative forced choice procedure while seated in the dark. Two targets

(visual angle 0.25) were presented sequentially in depth and patients had to report whether

the second target was nearer or farther than the first one. The standard target was presented at

a viewing distance of 69.7 cm throughout the experiment. The comparison target was shown

randomly at one of 5 locations and the viewing distance ranged between 67.6 cm to 71.7 cm.

The vergence angle required to converge the 5 targets ranged between 5 to 5.5 degrees.

124

Table 3: Patients’ clinical characteristics and surgical procedures

Patient Gender Age Age at

surgery

Stereoacuity

(arc sec-1

)

Deficit prior to surgery Surgical procedure

1 M 16 First: 2

Second: 7

140 Congenital left esotropia,

bilateral IO overactivation

Left exotropia developed two

years after first surgery

First: left MR recession (4.5 mm),

bilateral IO myotomy

Second: left LR recession (7 mm)

2 F 45 <2 40 Congenital left esotropia Could not be obtained

3 F 17 16 40 Congenital right 4th

nerve

palsy

Left IR recession (2.5 mm)

Right IO myectomy (5 mm)

4 F 55 54 60 Left 6th

nerve palsy due to

previous surgery to remove a

tumour

Transposition of left SR and IR to

strengthen the LR

5 M 43 43 40 Acquired right 4th

nerve palsy

due to head injury

Right IO recession (10 mm)

*MR = medial rectus; LR = lateral rectus; IR = inferior rectus; SR= superior rectus; IO = inferior oblique.

125

One of the comparison targets was shown at the same location as the standard target and the

other four were shown closer or farther than the standard. There were four experimental

conditions: (a) Control: standard and comparison targets were shown with no JM; (b) Task 1:

standard target appeared during the JM, and comparison target appeared after the JM was

released; (c) Task 2: standard target appeared when the JM was not performed, and the

comparison target appeared during the JM; (d) Task 3: standard and comparison targets

appeared while the JM was performed (see Figure 4 for an illustration of the protocol). The

experimental conditions were completely randomized. In each experimental condition the

comparison target was shown 10 times at each of the 5 locations for a total of 200 trials per

participant.

Results

Data for each patient were plotted individually and are presented in Figure 16. Visual

inspection of the data from three of the patients (patients 1 to 3) clearly show that the JM did

not affect the perceptual localization judgement. In contrast, patient 4 showed a similar

pattern of responses as the binocularly healthy observers (described in Paper 1, experiment

2), which is apparent in Figure 16. This patient reported that the location of the comparison

target was ‘farther’ while performing the JM in comparison to the condition when the JM

was performed during the presentation of the standard target. The psychometric functions of

patient 5 are more variable. There is a difference between the conditions when the JM was

performed during the presentation of the standard and comparison targets. Similarly to

patient 4 and the binocularly healthy observers, patient 5 reported that the comparison target

126

-2 -1 0 +1 +2

1.0

.8

.6

.4

.2

0

Location of comparison target (cm)

patient 1 patient 2

patient 3 patient 4

patient 5

control condition (no JM)

JM during standardJM during comparison

JM during standardno JM during comparison

no JM during standardJM during comparison

Location of comparison target (cm)

-2 -1 0 +1 +2

1.0

.8

.6

.4

.2

0

-2 -1 0 +1 +2

1.0

.8

.6

.4

.2

0-2 -1 0 +1 +2

1.0

.8

.6

.4

.2

0

-2 -1 0 +1 +2

1.0

.8

.6

.4

.2

0

Figure 16: Psychometric functions of the patients tested. The location of the comparison

target with respect to the standard is shown in cm with positive values indicating nearer and

negative farther positions from the observer. Proportion of ‘near’ responses for each

comparison target location (at 0 both targets were presented at the same location).

127

was ‘farther’ when the JM was performed during the presentation of the standard target for 3

(out of 5) possible locations of the comparison. However, this was not the case when the

comparison target was presented nearer than the standard target (Figure 16, patient 5).

Discussion

The purpose of this study was to determine whether patients who had undergone

strabismus surgery were susceptible to a manipulation that has been shown to alter registered

vergence eye position in binocularly healthy observers. It was hypothesized that patients who

had surgeries on their eye muscles would not be affected by the JM because the afferent

feedback loops would have been compromised by the surgeries. Specifically, since the

experimental task requires vergence eye movements, it was expected that patients who have

had surgeries on their horizontal eye muscles would have been least affected by the JM

manipulation. Although patients tested in the study had different deficits prior to surgery and

a variety of surgeries, data from three patients who had congenital strabismus supported the

hypothesis. Two of these patients had infantile esotropia and the surgery was performed on

the horizontal muscles in one of the patients (patient 1). The medical record for patient 2

could not be obtained, but the surgery was most likely performed on the horizontal muscles.

In contrast, patient 3 who was also not affected by the JM manipulation had surgery on the

inferior rectus and inferior oblique muscles. Data from the two patients who had acquired

strabismus do not support the hypothesis. One of these patients (patient 4) had a transposition

surgery which consisted of a partial transposition of the muscle fibers from the superior and

inferior recti to strengthen the lateral rectus. Since this surgery did not involve a complete

tenotomy, it is possible that the proprioceptive feedback loop was not affected to the same

128

extent compared to the other surgeries; thus the patient was able to perform the task similarly

to observers who did not have surgeries. Data from patient 5 is more variable which might be

partly explained by the etiology. The patient acquired superior oblique palsy due to a head

injury. The fact that a diffuse head injury might have affected his ability to concentrate and

perform the task consistently on all trials could not be ruled out. Alternatively, it is also

possible that the surgery disrupted the proprioceptive feedback loops to some extent which

resulted in increased variability of the responses.

The fact that surgical intervention can compromise the EOM afferent signals has been

previously reported by Steinbach and Smith (1981). Specifically, open-loop pointing

responses were examined in two groups of patients: one group that had had a single surgery

and a second group that had more than one surgery on the same eye muscles. The data

showed that patients in the first group were able to point accurately to targets immediately

after the surgery when the eye was uncovered and before visual experience was allowed.

Furthermore, the shift in the pointing responses corresponded to approximately one-quarter

of the amount that the eye was rotated during the surgery. These data suggest that patients

were able to use proprioceptive signals from the EOM to inform the CNS about the

surgically altered eye position. In contrast, the localization responses of patients in the

second group who had had multiple surgeries did not show a significant difference after the

surgery, which means that the altered eye position was not taken into account by the CNS.

The difference in performance between the patients that had one surgery in comparison to

patients who had multiple surgeries was most likely due to the progressive damage to the

EOM proprioceptors sustained during successive surgeries. Even a single surgery can

compromise the EOM afferent signal if there is significant damage to the myotendinous

129

junction. For example, Steinbach, Kirshner and Arstikaitis (1987) compared the effects of

recession and marginal myotomy and found that following myotomy, patients did not use the

proprioceptive signal to compensate for the changed eye position.

In the present study we tested patients who had different types of surgeries and only

one patient had two surgical interventions. Despite this, data from three patients clearly

showed that the gain of EOM feedback was not altered by the JM perturbation. There could

be several reasons to explain the lack of effect. First, it is possible that the surgery

compromised the receptors at the myotendinous junction, particularly in the patient who

experienced more than one surgery. Second, the gain of the signal could have also been

compromised if the afferent or the efferent nerves which supply the PE were affected by the

surgery. At present, the course of the afferent fibers has not been fully traced, but several

studies have shown that the afferent nerve travels with the ocular motor nerves and then

transfers to the ophthalmic branch of the trigeminal nerve near the apex of the orbit or in the

region of the cavernous sinus (for review, see Donaldson, 2000). It is also possible that the

surgery affected the axon that runs along the muscle fiber to the tendon and then loops back

to enter the capsule as it bifurcates to supply the PE.

One of the limitations of this study is the fact that patients were only tested after

surgery. Thus, it can not be determined whether the lack of gain modulation was due to

damage sustained due to the surgical intervention or whether these patients had an abnormal

proprioceptive apparatus prior to surgery. The three patients who were not affected by the JM

manipulation all had congenital deficits. Two studies that previously examined the EOM

myotendinous junction of patients who had congenital strabismus reported structural

alterations in the receptors and their innervation (Corsi, Sodi, Salvi & Faussone-Pellegrini,

130

1990; Domenici-Lombardo, Corsi, Mencucci, Scrivanti, Faussone-Pellegrini & Salvi, 1992);

thus, it is possible that gain modulation was impaired in these three patients prior to surgery.

The current study cannot distinguish between the two possibilities. Future studies should

examine a larger sample of patients both before and after the surgery.

In conclusion, the results from this study provide critical support for the hypothesis

that the effect of the JM on the localization responses in depth found in binocularly healthy

observers is specific to the eye muscles. The fact that the responses of patients with

congenital strabismus were not susceptible to the manipulation suggests that the effect most

likely depends on an intact EOM proprioceptive apparatus.

131

CHAPTER IV: GENERAL DISCUSSION

4.1 Summary of findings

The overall objective of this thesis was to test the hypothesis that the non-twitch

motoneurons associated with multiply innervated fibers (MIF) of the global layer of eye

muscles control the gain of proprioceptive feedback from the palisade endings (PE) in a way

analogous to the gamma-spindle system found in skeletal muscles, as proposed by Buttner-

Ennever and her colleagues (2002). The hypothesis was tested in a series of behavioural and

psychophysical studies which examined whether the afferent signals from the extra-ocular

muscles (EOM), and consequently registered eye position, was altered by a manipulation

which is known to affect the afferent feedback in skeletal muscles, presumably via the

gamma system.

Our behavioral studies with healthy human observers showed that the Jendrassik

maneuver (JM) affects registered vergence eye position. Specifically, participants reported

that targets were farther away during this perturbation which suggests that eye position was

registered as more divergent. The misregistration was evident with both manual and

perceptual responses. In contrast, neither the manual nor the perceptual localization of targets

in the frontal plane were affected by the JM. Since the non-twitch motoneurons receive direct

premotor input from areas involved in the programming of vergence but not saccadic eye

movements, our data support the hypothesis that the JM alters proprioceptive feedback from

EOM possibly via the activity of non-twitch motoneurons.

A critical finding from our studies was the fact that the kinematics of neither vergence

nor saccadic eye movements were affected by the JM. Since the non-twitch motoneurons do

not add to the force that is used to move the eyes (Fuchs & Luschei, 1971), the eye

132

movement data from our study yield further support for our hypothesis that the JM affects the

activity of the non-twitch motoneurons and not the twitch motoneurons.

In contrast to our hypotheses, higher order perceptual judgments that require accurate

registration of absolute depth were not affected by the perturbation. The lack of significant

effect might not be surprising given that the JM manipulation affects registered vergence eye

postion, but vergence itself is not a perfect cue to distance. Consequently, the perturbed

vergence signal was not taken into account by the central nervous system.

The last paper examined whether JM affects registered vergence eye position in

patients with strabismus who have had surgeries that most likely compromised the EOM

afferent feedback loops. As hypothesized, data showed that responses of patients with

congenital strabismus were not affected by the JM perturbation which might be due to the

fact that proprioceptive feedback could not be altered via the non-twitch motoneurons. These

results provide critical support for the fact that the effect of the JM depends critically on the

intact proprioceptive EOM feedback loop.

In summary, using a proxy method that alters the EOM afferent signal, possibly via

the activity of the non-twitch motoneurons, we provided behavioural evidence to support the

hypothesis proposed by Buttner-Ennever and colleagues (2002), and Robinson‟s original

claim that the PE and MIF might be part of an „inverted muscle spindle‟ (Robinson, 1991).

4.2 Significance

This project provides a novel insight into the mechanism involved in the use of

sensory feedback from the extraocular muscles. Data showed that the gain of proprioceptive

feedback was altered by the JM during vergence but not during saccadic eye movements. The

133

fact that the saccadic system was not affected is analogous to the findings of Guthrie and

colleagues (1982) who reported that cutting the monkeys‟ ophthalmic branch of the

trigeminal nerve (i.e., deafferentation) altered their vergence responses but had no effect on

their conjugate eye movements. Our results reinforce the importance of the EOM

proprioceptive feedback loop for binocular function.

An important question that remains is: What is the role of the gamma () system and

proprioception in general in oculomotor control and visuomotor behaviour? The unique

structure of the EOM and the fact that the cell body and the afferent pathway of the putative

proprioceptors, the PE, have not been definitively identified, makes it difficult to study the

question. The role of EOM proprioception in the control and execution of different types of

eye movements has been reviewed extensively by Donaldson (2000), but the possibility that

the system might modulate the gain of sensory feedback was only briefly mentioned.

It has been suggested that in the skeletal system “the fusimotor system allows state

dependent parametric adjustment of proprioceptive feedback” (Prochazka, 1989). The

implication of this hypothesis is that the loop is important for parametric adjustment of the

feedback loops to match the demands of different tasks, which might also be relevant for the

oculomotor system. For example, many studies have shown that the relationship between eye

position and the firing frequency of the ocular motoneurons is highly correlated (Carpenter,

1988). However, a study by Mays and Porter (1984) reported that the relationship between

eye position and firing rate is also dependent on the type of eye movement. In their study,

recordings were made from the abducens nucleus during conjugate adduction and during

convergence. Data showed that for a given eye position there was an increase in the firing

rate during convergence compared to conjugate adduction. Extending on these results, Miller

134

and colleagues (2002) measured the oculorotary forces in the horizontal recti muscles to test

whether the force developed in the lateral rectus is in fact higher in the converged state.

Paradoxically and in contrast to the hypothesis, they found decreased forces in both the

lateral and medial recti muscles during convergence. These results clearly showed that the

innervation of the EOM is much more complex than previously acknowledged, and it is

possible that the motor commands to the eye muscles differ during convergence and

conjugate adduction. In light of our results, it should also be acknowledged that the gain of

the proprioceptive system might be set differently for different types of eye movements.

The results of this project may have important clinical implications for the treatment

of strabismus. Surgical intervention is a common treatment for strabismus, but often does not

result in regaining of stereoscopic vision and vergence eye movements. In fact, as many as

50% of children have to undergo multiple surgeries (Helveston, Neely, Stidham, Wallace et

al., 1999). It is possible that the lack of success is partly due to the damage sustained at the

myotendinous region which contains the putative proprioceptors of the eye muscles.

However, it has also been suggested that sparing of the receptors which send abnormal

signals might contribute to sensory and motor deficits in binocular vision in patients with

strabismus (Steinbach, 1986). On the other hand, it is also possible that congenital

abnormalities of the sensory receptors and pathway contribute to the development of

strabismus and whether the surgery affects the receptors has no bearing on the outcome of

the surgery.

135

4.3 Limitations

Although our research provides a novel insight into the mechanism of EOM

feedback, it also has limitations. Our manipulation, the JM, has been used extensively to alter

the excitability of spinal reflexes and limb position information. There is no dispute that JM

alters the afferent signals and has a general effect that affects all spinal and brainstem

reflexes but no studies to our knowledge have been able to unequivocally explain the

neurophysiological mechanism involved. Despite the debate, we assumed that one of the

mechanisms through which the JM operates is the gamma system and we used it to perturb

the afferent signals from EOM. Since it has been shown that the gamma system might be

predominately involved in modulating the reflex response during the steady phase of the

isometric contraction (Phase 3) (Delwaide & Toulouse, 1980), we ensured that the

localization responses that participants were asked to make in all the experiments also

occurred in that phase. Although we used an indirect method, our data clearly show that JM

affected localization responses, but not the actual eye position, associated with vergence but

not versional eye movements. Overall, these results provided support for the proposal that the

JM altered the afferent signals from EOM via the activity of non-twitch motoneurons.

Results from our studies highlight the importance of reporting negative findings. The

significance of publishing negative results in oculomotor research was emphasized by the

study by Keller and Robinson (1971), which reported the absence of stretch reflexes in EOM.

In the present thesis, we reported negative results that supported our hypotheses in Paper 2

and in Paper 3 our predictions were not supported by the data. Findings from both papers are

valuable and it is important to disseminate these results as they help to understand the

mechanisms involved in EOM proprioceptive feedback control.

136

The use of a proxy method in a behavioural study to infer neurophysiological

mechanisms is a major limitation; however, it is currently the only method of investigating

the question whether the gain of afferent feedback from EOM can be altered via the gamma

system in humans. The only method that would provide unequivocal evidence for a gain

control regulation of proprioceptive feedback from the eye muscles would be recording from

the cell body of the EOM sensory neuron while stimulating the non-twitch motoneurons.

This type of experiment may never be possible in the humans and a proxy method offers the

only way of investigating the question. Even in sub-human primates no one has yet recorded

from the somata of the palisade endings.

Another limitation of the present study is the fact that our hypotheses are based on

neuroanatomical tracing studies conducted in sub-human primates. At present, it is unknown

whether the human EOM fibers also receive dual innervation from ocular motor nuclei.

Nonetheless, human EOM do contain similar fiber types that are found in the primates and

other mammals (Wasicky et al., 2000) and the PE are also associated with the global multiply

innervated fibers (Richmond et al., 1984). Overall, the human and primate EOM are

remarkably similar in their organization, histochemical properties and repertoire of eye

movements so it is likely that the dual innervation hypothesis can be extended to humans as

well.

4.4 Future direction

The studies presented in this thesis were the first to examine whether sensory

feedback from EOM can be altered by a manipulation that presumably changes the

excitability of the gamma system. Data provide strong, but preliminary, support for the

137

hypothesis that the gain of EOM proprioceptive feedback loops might be set differently to

match the demands of different types of eye movements. Due to the limited scope of the

thesis, only two types of eye movements were examined: vergence and saccades. Future

studies should examine whether the JM also affects other oculomotor systems. For instance,

since the non-twitch motoneurons receive direct premotor input from the area involved in

gaze-holding mechanisms and pursuit (Wasicky et al., 2004; Ugolini et al., 2006), these

should be examined next. In particular, a recent retrograde tracing study by Billig and Strick

(2007) showed that the cortical inputs to the MIF come from the frontal eye field area that is

predominately involved in smooth pursuit. Our pilot studies have shown that pursuit of

targets in the frontal plane was not affected by the JM; however, this does not negate the fact

that pursuit of targets moving in depth might be affected, which should be examined next.

The proxy method we used to alter the activity of the gamma system is not the only

means by which the gamma system can be activated indirectly. Several studies have shown

that attentionally demanding tasks, such as mental calculation can also increase spindle

activity without concomitant activation of the alpha motoneurons (Ribot, Roll & Vedel,

1986; Ribot-Ciscar et al., 2000; Rossi-Durand, 2002). These studies used microneurography

and inferred gamma activity from the recordings from Ia afferents. One of the only studies

that directly recorded from gamma motoneurons confirmed that their activity was selectively

enhanced by mental computation, pinna twisting, startling the subject by hand clapping

behind their back or changes in the environmental conditions. Overall, these studies suggest

that gamma activity can be enhanced indirectly by any task that increases arousal, thus, any

of these tasks could be used instead of the JM to confirm the results of the current studies.

138

Results of this project may have important clinical implications for the treatment of

strabismus but only a few patients were tested and future studies should replicate the

experiment with a larger population. Additionally, patients who participated in the our study

(paper 4) had different deficits and surgeries which could affect the proprioceptive feedback

loops differently. We were unable to ascertain whether their performance characteristics were

due to damage sustained by the myotendinous junction involving the palisade endings or to a

disruption of the pathway that modulates the gain of proprioceptive feedback from extra-

ocular muscles. To shed more light on this issue, future studies should examine patients pre

and post surgical intervention.

In conclusion, after years of neglect, extra-ocular muscle proprioception has received its

due attention in the past few years. It is now indisputable that both afferent and efferent

signals play a role in oculomotor control and visuomotor behaviour and must be taken into

account when developing models of oculomotor control. Furthermore, the efferent signals

that have to be considered must include the alpha and gamma systems.

139

References

Alvarado-Mallart, R. M., & Pinçon-Raymond, M. (1979). The palisade endings of cat

extraocular muscles: a light and electron microscope study. Tissue Cell, 11(3), 567–

584.

Appelberg, B. (1981). Selective central control of dynamic gamma motoneurons utilized for

the functional classification of gamma cells. In A. Taylor & A. Prochazka (Eds.),

Muscle receptors and movement. (pp. 97–108). New York: Oxford University Press.

Appelberg, B., & Jeneskog, T. (1972). Mesencephalic fusimotor control. Experimental Brain

Research,15, 97–112.

Appenteng, K., Morimoto, T., & Taylor, A. (1980). Fusimotor activity in masseter nerve of

the cat during reflex jaw movements. Journal of Physiology, 305, 415–431.

Bach-y-Rita, P. (1971). The control of eye movements. In P. Bach-y-Rita & C.C. Collins

(Eds). The control of eye movements. (pp. 3–45). New York: Academic Press.

Billig, I., Buisseret Delmas, C., & Buisseret, P. (1997). Identification of nerve endings in cat

extraocular muscles. Anatomical Record, 248(4), 566–575.

Billig, I., & Strick, P.L. (2007). Evidence that the two types of abducens motoneurons are

controlled by distinct regions of the frontal eye field. Society for Neuroscience, 398.1.

Blumer, R., Lukas, J. R., Aigner, M., Bittner, R., Baumgartner, I., & Mayr, R. (1999). Fine

structural analysis of extraocular muscle spindles of a two-year-old human infant.

Investigative Ophthalmology and Visual Science, 40(1), 55–64.

Blumer, R., Lukas, J. R., Wasicky, R., & Mayr, R. (1998). Presence and structure of

innervated myotendinous cylinders in sheep extraocular muscle. Neuroscience

Letters, 248(1), 49–52.

140

Blumer, R., Wasicky, R., Hoetzenecker, W., & Lukas, J.R. (2001). Presence and structure of

innervated myotendinous cylinders in rabbit extraocular muscle. Experimental Eye

Research, 73(6), 787–796.

Bridgeman, B., & Stark, L. (1991). Ocular proprioception and efference copy in registering

visual direction. Vision Research, 31(11), 1903–1913.

Brindley, G. S., Goodwin, G. M., Kulikowski, J. J., & Leighton, D. (1976). Proceedings:

Stability of vision with a paralysed eye. Journal of Physiology, 258(2), 65P–66.

Bruckner, G., Brauer, K., Hartig, W., et al. (1993). Perineuronal nets provide a

polyanionic, glia-associated form of microenvironment and around certain neuron in

many parts of the rat brain. Glia, 8(3), 183–200.

Bruckner, G., Schutz, A., Hartig, W., Brauer, K., Paulke, B.R., & Bigl, V. (1994).

Projections of non-cholinergic basal forebrain neurons ensheathed with perineuronal

nets to rat mesocortex. Journal of Chemical Neuroanatomy, 8(1), 11–18.

Bruenech, R., & Ruskell, G. L. (2000). Myotendinous nerve endings in human infant and

adult extraocular muscles. Anatomical Record, 260(2), 132–140.

Bucci, M. P., Kapoula, Z., Yang, Q., Roussat, B., & Bremond-Gignac, D. (2002). Binocular

coordination of saccades in children with strabismus before and after surgery.

Investigative Ophthalmology and Visual Science, 43(4), 1040–1047.

Buisseret, P. (1995). Influence of extraocular muscle proprioception on vision. Physiology

Reviews, 75(2), 323–338.

Buller, A. J., & Dornhorst, A. C. (1957). The reinforcement of tendon-reflexes. Lancet,

273(7008), 1260–1262.

Burg, D., Szumski, A. J., Struppler, A., & Velho, F. (1973). Afferent and efferent activation

141

of human muscle receptors involved in reflex and voluntary contraction.

Experimental Neurology, 41(3), 754–768.

Burke, D., McKeon, B., & Skuse, N. F. (1981). The irrelevance of fusimotor activity to the

Achilles tendon jerk of relaxed humans. Annals of Neurology, 10(6), 547–550.

Burke, D., McKeon, B., Skuse, N. F., & Westerman, R. A. (1980). Anticipation and

fusimotor activity in preparation for a voluntary contraction. Journal of Physiology,

306, 337–348.

Bussel, B., Morin, C., & Pierrot-Deseilligny, E. (1978). Mechanism of monosynaptic reflex

reinforcement during Jendrassik manoeuvre in man. Journal Neurology Neurosurgery

and Psychiatry, 41(1), 40–44.

Buttner-Ennever, J. A. (2005). The extraocular motor nuclei: organization and functional

neuroanatomy. Progress in Brain Research, 151, 95–125.

Buttner-Ennever, J. A., & Horn, A. K. (2002). The neuroanatomical basis of oculomotor

disorders: the dual motor control of extraocular muscles and its possible role in

proprioception. Current Opinion in Neurology, 15(1), 35–43.

Buttner-Ennever, J. A., Horn, A. K., Graf, W., & Ugolini, G. (2002). Modern concepts of

brainstem anatomy: from extraocular motoneurons to proprioceptive pathways.

Annals of the New York Academy of Science, 956, 75–84.

Buttner-Ennever, J. A., Horn, A. K., Scherberger, H., & D'Ascanio, P. (2001). Motoneurons

of twitch and nontwitch extraocular muscle fibers in the abducens, trochlear, and

oculomotor nuclei of monkeys. Journal of Comparative Neurology, 438(3), 318–335.

Buttner-Ennever, J.A., Konakci, K.Z. & Blumer, R. (2005). Sensory control of extraocular

muscles. Progress in Brain Research, 151, 81–93.

142

Chalmers, L. (1952). Monocular and binocular cues in the perception of size and distance.

American Journal of Psychology, 65(3), 415–423.

Carpenter, R. H. S. (1988). Movements of the eyes. (2nd ed. ed.). London: Pion Limited.

Clare, M.H., & Landau, W.M., (1964). Fusimotor function. V. Reflex reinforcement under

fusimotor black in normal subjects. Archives of Neurology, 10, 123–127.

Clark, R. A., Miller, J. M., & Demer, J. L. (1997). Location and stability of rectus muscle

pulleys. Muscle paths as a function of gaze. Investigative Ophthalmology and Visual

Science, 38(1), 227–240.

Clark, R. A., Miller, J. M., & Demer, J. L. (2000). Three-dimensional location of human

rectus pulleys by path inflections in secondary gaze positions. Investigative

Ophthalmology and Visual Science,, 41(12), 3787–3797.

Collewijn, H. & Erkelens, C.J. (1990). Binocular eye movements and the perception of

depth. In E. Kowler (ed.), Eye movements and their role in visual and cognitive

processes. (pp. 213–261). New York: Elsevier Science.

Corsi, M., Sodi, A., Salvi, G., & Faussone-Pellegrini, M. S. (1990). Morphological study of

extraocular muscle proprioceptor alterations in congenital strabismus.

Ophthalmologica, 200(3), 154–163.

Dancause, N., Taylor, M.D., Plautz, E.J., et al. (2007). A stretch reflex in extraocular muscles

of species purportedly lacking muscle spindles. Experimental Brain Research,180(1),

15–21.

Delwaide, P. J., & Toulouse, P. (1980). Jendreassik Maneuver vs controlled contractions

conditioning the excitability of soleus monosynaptic reflexes. Archives of Physical

Medicine and Rehabilitation, 61, 505-510.

143

Delwaide, P. J., & Toulouse, P. (1981). Facilitation of monosynaptic reflexes by voluntary

contraction of muscle in remote parts of the body. Mechanisms involved in the

Jendrassik Manoeuvre. Brain, 104(Pt 4), 701–709.

Demer, J. L. (2002). The orbital pulley system: a revolution in concepts of orbital anatomy.

Annals of the New York Academy of Science, 956, 17–32.

Demer, J. L., Oh, S. Y., & Poukens, V. (2000). Evidence for active control of rectus

extraocular muscle pulleys. Investigative Ophthalmology and Visual Science, 41(6),

1280–1290.

Domenici-Lombardo, L., Corsi, M., Mencucci, R., Scrivanti, M., Faussone-Pellegrini, M.S.,

Salvi, G. (1992). Extraocular muscles in congenital strabismus: muscle fiber an nerve

ending ultrastructure according to different regions. Ophthalmologica, 205(1), 29–39.

Donaldson, I.M. (2000). The function of the proprioceptors of the eye muscles. Philosophical

Transactions of the Royal Society of London. Series B. Biological Sciences,

355(1404), 1685–1754.

Dowman, R., & Wolpaw, J. R. (1988). Jendrassik maneuver facilitates soleus H-reflex

without change in average soleus motoneuron pool membrane potential. Experimental

Neurology, 101(2), 288–302.

Eakins, K.E., & Katz, K.E. (1972). The pharmacology of extraocular muscles. In P. Bach-y-

Rita & C.C. Collins (Eds.). The control of eye movements. (pp. 237–258). New York:

Academic Press

Eberhorn, A.C., Ardeleanu, P., Buttner-Ennever, J.A., & Horn, A.K. (2005). Histochemical

difference between motoneurons supplying multiply and singly innervated

extraocular muscle fibers. Journal of Comparative Neurology, 491(4), 352–366.

144

Eberhorn, A.C. Horn, A.K., Eberhorn, N., Fischer, P., Boergen, K.P., & Buttner-Ennever,

J.A. (2005). Palisade endings in extraocualr eye muscles revealed by SNAP-25

immunoreactivity. Journal of Anatomy, 206, 307–315.

Eberhorn, A.C., Horn, A.K., Fischer, P., & Buttner-Ennever, J.A. (2005). Proprioception and

palisade endings in extraocular eye muscles. Annals of the New York Academy of

Science, 1039, 1–8.

Fiorentini, L., & Maffei, L. (1977). Instability of the eye in the dark and proprioception.

Nature 269(5626), 330–331.

Fiorentini, A., Mafei, L., Cenni, M.C. & Tacchi, A. (1985). Deafferentation of oculomotor

proprioception affects depth discrimination in adult cats. Experimental Brain

Research, 59(1), 296–301.

Foley, J. M. (1980). Binocular distance perception. Psychological Review, 87(5), 411–434.

Fronius, M., & Sireteanu, R. (1994). Pointing errors in strabismics: complex patterns of

distorted visuomotor coordination. Vision Research, 34(5), 689–707.

Fuchs, A. F., Luschei, E.S. (1971). Development of isometric tension in simian

extraocular muscle. Journal Physiology, 219(1), 155–166.

Gauthier, G. M., Nommay, D., & Vercher, J. L. (1990a). The role of ocular muscle

proprioception in visual localization of targets. Science, 249(4964), 58–61.

Gauthier, G. M., Nommay, D., & Vercher, J. L. (1990b). Ocular muscle proprioception and

visual localization of targets in man. Brain, 113(Pt 6), 1857–1871.

Gordon, J. & Ghez, C. (1991). Muscle receptors and spinal reflexes: the stretch reflex. In

E.R. Kandel, J.H. Schwartz, & T.M. Jessell (Eds.), Principals of Neural Science, 3rd

ed. pp.564–580. Norwalk: Appleton& Lange.

145

Gottlieb, S., & Taylor, A. (1983). Interpretation of fusimotor activity in cat masseter nerve

during reflex jaw movements. Journal of Physiology, 345, 423–438.

Graham, P. A. (1974). Epidemiology of strabismus. British Journal of Ophthalmology,

58(3), 224–231.

Granit, R., & Kaada, B.R., (1952). Influence of stimulation of central nervous structures on

muscle spindles in cat. Acta Physiologica Scandinavica, 27, 130–160.

Grant, S., Melmoth, D.R., Morgan, M.J., & Finlay, A.L. (2007). Prehension deficits in

amblyopia. Investigative Ophthalmology and Visual Science, 48(3), 1139–1148.

Graves, A. L., Trotter, Y., & Fregnac, Y. (1987). Role of extraocular muscle proprioception

in the development of depth perception in cats. Journal of Neurophysiology, 58(4),

816–831.

Gregory, J. E., Wood, S. A., & Proske, U. (2001). An investigation into mechanisms of reflex

reinforcement by the Jendrassik manoeuvre. Experimental Brain Research, 138(3),

366–374.

Grusser, O.J. (1995). On the history of the ideas of efference copy and reafference. Clio

Medica. 33, 35–55.

Guthrie, B.L., Porter, J.D. & Sparks, D.L. (1982). Role of extraocular muscle proprioception

in eye movement studies by chronic deafferentation of intra-orbital structures. Society

for Neuroscience, 8, 156.

Guthrie, B.L., Porter, J.D., & Sparks, D.L. (1983). Corollary discharge provides accurate eye

position information to the oculomotor system. Science, 221, 1193–1195.

146

Hagbarth, K. E., Wallin, G., Burke, D., & Lofstedt, L. (1975). Effects of the Jendrassik

manoeuvre on muscle spindle activity in man. Journal of Neurology Neurosurgery

and Psychiatry, 38(12), 1143–1153.

Hallet, P.E. (1986). Eye Movements. In K.E. Boff, L. Kaufman, & J.P. Thomas (Eds).

Handbook of perception and human performance. Volume I. Sensory Processes and

Perception. Pg. 10.1–10.10.112. New York: John Wiley and Sons.

Harvey, L.O. & Leibowitz, H.W. (1967). Effects of exposure duration, cue reduction, and

temporary monocualrity on size matching at short distances. Journal of the Optical

Society of America, 57, 279–283.

Hayman, M.R., & Donaldson, I.M. (1995). Deafferentation of pigeon extraocular muscles

disrupts eye movements. Proceedings of the Royal Society of London, B254, 115–

122.

Helveston, E. M., Neely, D. F., Stidham, D. B., Wallace, D. K., Plager, D. A., & Sprunger,

D. T. (1999). Results of early alignment of congenital esotropia. Ophthalmology,

106(9), 1716–1726.

Hertle, R.W., Chan, C.C., Galita, D.A., Maybodi, M., & Craford, M.A. (2002).

Neuroanatomy of the extraocular muscle tendon enthesis in macaque, normal human,

and patients with congenital nystagmus. Journal of the American Association of

Pediatric Ophthalmology and Strabismus;6, 319-327.

Howard, I.P. (2002). Seeing in depth. Volume 1: Basic Mechanisms. Toronto: I Porteous.

Howard, I.P. & Rogers, B.J. (2002). Seeing in depth. Volume 2: Depth Perception. Toronto: I

Porteous.

147

Hulliger, M. (1984). The mammalian muscle spindle and its central control. Reviews of

Physiology, Biochemistry and Pharmacology, 101, 1-110.

Hulliger, M. (1993). Fusimotor control of proprioceptive feedback during locomotion and

balancing: can simple lessons be learned for artificial control of gait? Progress in

Brain Research, 97, 173–180.

Jacoby, J., Chiarandini, D. J., & Stefani, E. (1989). Electrical properties and innervation of

fibers in the orbital layer of rat extraocular muscles. Journal of Neurophysiology,

61(1), 116–125.

Kapoula, Z., Bucci, M. P., Eggert, T., & Garraud, L. (1997). Impairment of the binocular

coordination of saccades in strabismus. Vision Research, 37(19), 2757–2766.

Keller, E.L., & Robinson, D.A. (1971). Absence of a stretch reflex in extraocular muscles of

the monkey. Journal of Neurophysiology, 34(5), 908–919.

Knox, P. C., Weir, C. R., & Murphy, P. J. (2000). Modification of visually guided saccades

by a nonvisual afferent feedback signal. Investigative Ophthalmology and Visual

Science, 41(9), 2561–2565.

Konakci KZ, Streicher J, Hoetzenecker W, Blumer, M.J., Lukas, J.R., & Blumer, R. (2005).

Molecular characteristics suggest an effector function of palisade endings in

extraocular muscles. Investigative Ophthalmology and Visual Science, 46(1), 155–

165.

Konakci KZ, Streicher J, Hoetzenecker W, Haberl, I., et al. (2005). Palisade endings in

extraocular muscles of the monkey are immunoreactive for choline acetyltransferase

and vesicular acetylcholine transporter. Investigative Ophthalmology and

VisualScience. 46, 4548–4554.

148

Leibowitz, H.W. & Moore, D. (1966). Role of accommodation and convergence in the

perception of size. Journal of the Optical Society of America, 56, 1120–1123.

Lewis, R.F., Gaymard, B.M., & Tamargo, R.J. (1998). Efference copy provides the eye

position information required for visually guided reaching. Journal of

Neurophysiology, 80(3), 1605–1608.

Lewis, R.F., Zee, D.S. (1993). Abnormal spatial localization with trigeminal-oculomotor

synkinesis. Brain, 116, 1105–1118.

Lewis, R.F., Zee, D.S., Gaymard, B.M., & Guthrie, B.L. (1994). Extraocular muscle

proprioception functions in the control of ocular alignment and eye movement

conjugacy. Journal of Neurophysiology, 72(2), 1028–1031.

Lewis, R. F., Zee, D. S., Goldstein, H. P., & Guthrie, B. L. (1999). Proprioceptive and retinal

afference modify postsaccadic ocular drift. Journal of Neurophysiology, 82(2), 551–

563.

Lewis, R.F., Zee, D.S., Hayman, M.R., & Tamargo, R.J. (2001). Oculomotor function in the

rhesus monkey after deafferentation of the extraocular muscles. Experimental Brain

Research, 1414(3), 349–358.

Lit, A. & Hyman, A. (1951). The magnitude of Pulfrich stereophenomeon as a function of

distance of observation. American Journal of Optometry and Archives of the

American Academy of Optometry, 28, 564–580.

Lukas, J. R., Aigner, M., Blumer, R., Heinzl, H., & Mayr, R. (1994). Number and

distribution of neuromuscular spindles in human extraocular muscles. Investigative

Ophthalmology and Visual Science, 35(13), 4317–4327.

149

Lukas, J. R., Blumer, R., Denk, M., Baumgartner, I., Neuhuber, W., & Mayr, R. (2000).

Innervated myotendinous cylinders in human extraocular muscles. Investigative

Ophthalmology and Visual Science, 41(9), 2422–2431.

Maier, A., DeSantis, M., & Eldred, E. (1974). The occurrence of muscle spindles in

extraocular muscles of various vertebrates. Journal of Morphology, 143(4), 397–408.

Malinov, I.V., Epelboim, J., Herst, A.N., & Steinman, R.M. (2000). Characteristics of

saccades and vergence in two kinds of sequential looking tasks. Vision Research,

40(16), 2083–2090.

Matin, L. (1976). A possible hybird mechanism for modification of visual direction

associated with eye movements-the paralyzed eye experiment reconsidered.

Perception, 5, 133–139.

Matthews, P.B.C. (1981). Evolving views on the internal operation and functional role of the

muscle spindle. Journal of Physiology, 320, 1–30.

Matthews, P. B.C. (1982). Where does Sherrington’s “muscular sense” originate? Muscles,

joints, coraollary discharges? Annual Reviews of Neuroscience, 5, 189–218.

Mays, L.E. (1984). Neural control of vergence eye movements: convergence and divergence

neurons in midbrain. Journal of Neurophysiology, 51, 1091–1108.

Mays, L. E., & Porter, J. D. (1984). Neural control of vergence eye movements: activity of

abducens and oculomotor neurons. Journal of Neurophysiology, 52(4), 743–761.

McClosky, D.I. (1981). Corollary dischrages: motor commands and perception. In Handbook

of physiology. II. Nervous sytem (ed. V.B. Brooks), pp. 1415-1447. Baltimore, MD:

Williams & Wilkins.

150

Miller, J. M., Bockisch, C. J., & Pavlovski, D. S. (2002). Missing lateral rectus force and

absence of medial rectus co-contraction in ocular convergence. Journal of

Neurophysiology, 87(5), 2421–33.

Mon-Williams, M., Tresilian, J.R., Plooy, A., Wann, J.P., & Broerse, J. (1997). Looking at

the task in hand: vergence eye movements and perceived size. Experimental Brain

Research, 133, 127–136.

Murthy, K. S. (1978). Vertebrate fusimotor neurons and their influences on motor behavior.

Progress in Neurobiology, 11(3–4), 249–307.

Nakamizo, S. & Lei, C. (2000). The Pulfrich effect and depth constancy. Japanese

Psychological Research, 42(4), 251–256.

Niechwiej-Szwedo, E. González, E., Bega, S., Verrier, M.C., Wong, A.M. & Steinbach, M.J.

(2006). Proprioceptive role for palisade ending in extraocular muscles: Evidence from

the Jendrassik Maneuver. Vision Research, 46, 2268–2279.

Niechwiej-Szwedo, E. González, E., Bahl, B., Verrier, M.C., Wong, A.M. & Steinbach, M.J.

(2007). Manipulation of extraocular muscle afference has no effect on higher order

perctpual judgments. Vision Research, 47, 3315–3323.

Niechwiej-Szwedo, E. González, E., Verrier, M.C., Wong, A.M. & Steinbach, M.J. (2008).

Localization in the frontal plane is not susceptible to manipulation of afferent

feedback via the Jendrassik Maneuver. Vision Research, 48, 724-732.

Niechwiej-Szwedo, E. & Steinbach M.J. (2007). Afferent and efferent contributions to

knowledge of eye position. In Advances in Understanding Mechanisms and

Treatment of Infantile Forms of Nystagmus (a tribute to Louis F. Dell'Osso). R. J.

Leigh & M. W. Devereaux (Eds.), New York: Oxford University Press

151

Ono, H. (1966). Distal and proximal size under reduced and non-reduced viewing conditions.

American Journal of Psychology, 79, 234–241.

Ono, H. & Comerford, J. (1977). Stereoscopic depth constancy. In W. Epstein (ed.), Stability

and constancy in visual perception. (pp. 91–128). Toronto: Wiley.

Oh, S. Y., Poukens, V., & Demer, J. L. (2001). Quantitative analysis of rectus extraocular

muscle layers in monkey and humans. Investigative Ophthalmology and Visual

Science, 42(1), 10–16.

Pasztor, E. (2004). Erno Jendrassik (1858–1921). Journal of Neurology, 251, 366–367

Pettorossi, V.E., Ferraresi, A., Draicchio, F., Errico, P., Santarelli, R., & Manni, E. (1995).

Acta oto-laryngologica. Supplement, 520(1), 77–81.

Porter, J.D., & Spencer, R.F. (1982). Localization of morphology of cat extraocular muscle

afferent neurons identified by retrograde transport of horseradish peroxide. Journal of

Comparative Neurology, 204(1), 56–64.

Porter, J.D., Baker, R.S., Ragusa, R.J. & Brueckner, J.K. (1995). Extraocular muscles:

basic and clinical aspects of structure and function. Survey of Ophthalmology,

39, 451–484.

Prochazka, A. (1989). Sensorimotor gain control: a basic strategy of motor systems?

Progress in Neurobiology, 33(4), 281–307.

Proske, U., Morgan, D.L., & Gregory, J.E. (1993). Thixotropy in skeletal muscle and in

muscle spindles : a review. Progress in Neurobiology, 41, 705–721.

Ribot, E., Roll, J.P., & Vedel, J.P. (1986). Efferent discharges recorded from single

skeletomotor and fusimotor fibres in man. Journal of Physiology, 375, 251–268.

152

Ribot-Ciscar, E., Rossi-Durand, C., & Roll, J. P. (2000). Increased muscle spindle sensitivity

to movement during reinforcement manoeuvres in relaxed human subjects. The

Journal of Physiology, 523 Pt 1, 271–282.

Richmond, F. J., Johnston, W. S., Baker, R. S., & Steinbach, M. J. (1984). Palisade endings

in human extraocular muscles. Investigative Ophthalmology and Visual Science,

25(4), 471–476.

Ritter, M. (1977). Effect of disparity and viewing distance on perceived depth. Perception &

Psychophysics, 22(4), 400–407.

Robinson, D. A. (1981). Control of eye movements. In V. B. Brooks (Ed.), The nervous

system, handbook of physiology. (pp. 1275–1320), Baltimore: Williams & Wilkins.

Robinson, D. A. (1991). Overview, in vision and vision dysfunction. In R. H. S. Carpenter

(Ed.), Eye movements (pp. 320–331). Boca Raton: CRC Press.

Roll, R., Velay, J. L., & Roll, J. P. (1991). Eye and neck proprioceptive messages contribute

to the spatial coding of retinal input in visually oriented activities. Experimental

Brain Research, 85(2), 423–431.

Rossi-Durand, C. (2002). The influence of increased muscle spindle sensitivity on Achilles

tendon jerk and H-reflex in relaxed human subjects. Somatosensensory and Motor

Research, 19(4), 286–295.

Ruskell, G.L. (1978). The fine structure of innervated myotendinous cylinders in extraocular

muscles of rhesus monkeys. Journal of Neurocytology, 7(6), 693–708

Ruskell, G. L. (1989). The fine structure of human extraocular muscle spindles and their

potential proprioceptive capacity. Journal of Anatomy, 167, 199–214.

153

Ruskell, G. L. (1999). Extraocular muscle proprioceptors and proprioception. Progress in

Retinal and Eye Research, 18(3), 269–291.

Shan, X., Tian, J. Ying, H.S., Quaia, C. Optican, L.M., Walker, M.F., Tamargo, R.J., & Zee,

D.S. (2007). Acute superior oblique palsy in monkeys: I. Changes in static eye

alignment. Investigative Ophthalmology and Visual Science, 48(6), 2602–2611.

Skavenski, A. A. (1972). Inflow as a source of extraretinal eye position information. Vision

Research, 12(2), 221–229.

Spencer, R. F., & Porter, J. D. (1988). Structural organization of the extraocular muscles.

Reviews of Oculomotor Research, 2, 33–79.

Spencer, R.F. & Porter, J.D. (2005). Biological organization of the extraocular muscles.

Progress in Brain Research, 151, 43–80.

Steinbach, M.J. (1986). Inflow as long-term calibrator of eye position in humans. Acta

Psychologica, 63, 297-306.

Steinbach, M. J. (1987). Proprioceptive knowledge of eye position. Vision Research,

27(10), 1737–1744.

Steinbach, M. J., Kirshner, E. L., & Arstikaitis, M. J. (1987). Recession vs marginal

myotomy surgery for strabismus: effects on spatial localization. Investigative

Ophthalmology Visual Science, 28(11), 1870–1872.

Steinbach, J.J., & Lerman, J. (1990). Gravity affects resting eye position in humans. Archives

of Ophthalmology and Vision Science (Suppl. 31), 553.

Steinbach, M. J., & Smith, D. R. (1981). Spatial localization after strabismus surgery:

evidence for inflow. Science, 213(4514), 1407–1409.

154

Stidwill, D. (1997). Epidemiology of strabismus. Ophthalmic and Physiological Optics,

17(6), 536–539.

Stevens, J.K., Emerson, R.C., Gerstein, G.L., et al. (1976). Paralysis of the awake human :

visual perceptions. Vision Research,16, 93–98.

Tanji, J. (1976). Selective activation of neurons in cortical area 3a associated with accurate

position maintenance. Brain Research, 115, 328–333.

Taylor, A., & Donga, R. (1989). Central mechanisms of selective fusimotor control. Progress

in Brain Research, 80, 27–35.

Taylor, A. & Gottlieb, S. (1985). Convergence of several senseory modalities in motor

control. In W.J.P. Barnes & M.H. Gladden (Eds.), Feedback and motor control in

invertebrates and vertebrates. (pp. 77–92). London: Croom Helm.

Tomlinson, R.D. & Schwarz, D.E. (1977). Response of oculomotor neurons to eye muscle

stretch . Canadian Journal of Physiology and Pharmacology, 55, 568–573.

Ugolini, B., Klam, F., Dans, M.D., Dubayle, D., Brandi, A.M., Buttner-Ennever, J., et al.

(2006). Horizontal eye movements networks in primates as revealed by retrograde

transneuronal transfer of rabies virus: differences in monosynaptic input to “slow”

and “fast” abducens motoneurons. Journal of Comparative Neurology, 498, 762-785.

van Donkelaar, P., Gauthier, G.M., Blouin, J., Vercher, J.L. (1994). The role of ocular

muscle proprioception during modifications in smooth pursuit output. Vision

Research, 37(6), 769–774.

Velay, J. L., Roll, R., Lennerstrand, G., & Roll, J. P. (1994). Eye proprioception and visual

localization in humans: influence of ocular dominance and visual context. Vision

Research, 34(16), 2169–2176.

155

Ventre-Dominey, J., Dominey, P. F., & Sindou, M. (1996). Extraocular proprioception is

required for spatial localization in man. Neuroreport, 7(9), 1531–1535.

Von Noorden, G.K., & Campos, E.C. (2002). Binocular vision and ocular motility: theory

and management of strabismus. 6th

ed. St. Louis: Mosby Inc.

Wallach, H. & Floor, L. (1971). The use of size matching to demonstrate the effectiveness of

accommodation and convergence as cues for distance. Perception & Psychophysics,

10, 423–428.

Wallach, H. & Zuckerman, C. (1963). The constancy of stereoscopic depth. American

Journal of Psychology, 76, 48–59.

Wallach, H., Gillam, B., Cardillo, L. (1979). Some consequences of stereoscopic depth

constancy. Perception & Psychophysics, 26(3), 235–240.

Wang, X., Zhang, M., Cohen, I.S. & Goldberg, M.E. (2007). The proprioceptive

representation of eye position in monkey primary somatosensory cortex. Nature

Neuroscience, 10, 640–646.

Wasicky, R., Ziya-Ghazvini, F., Blumer, R., Lukas, J.R., & Mayr, R. (2000). Muscle fiber

types of human extraocular muscles: a histochemical and immunohistochemical

study. Investigative Ophthalmology and Visual Science, 41(5), 980–990.

Wasicky, R., Horn, A.K., & Buttner-Ennever, J.A. (2004). Twitch and nontwitch

motoneurons subgroups in the oculomotor nucleus of monkeys receive different

afferent projections. Journal of Comparative Neurology, 479(2), 117–129.

Weir, C. R., Cleary, M., Parks, S., & Dutton, G. N. (2000). Spatial localization in esotropia:

does extraretinal eye position information change? Investigative Ophthalmology and

Visual Science, 41(12), 3782–3786.

156

Weir, C. R., & Knox, P. C. (2001). Modification of smooth pursuit initiation by a nonvisual,

afferent feedback signal. Investigative Ophthalmology and Visual Science, 42(10),

2297–2302.

Whitteridge, D. (1959). The effect of stimulation of intrfusal muscle fibres on sensitivity to

stretch of extraocular muscle spindles. Quarterly Journal of Experimental Physiology

and Cognitive Medical Science, 44, 385–393.

Yasuda, T., Izumizaki, M., Ishihar, Y., Sekihara, C., Atsumi, T. & Homma, I. (2006). Effect

of quadriceps contraction on upper limb position sense errors in humans. European

Journal of Physiology, 96(5), 511–516.

Zehr, E. P., & Stein, R. B. (1999). Interaction of the Jendrassik maneuver with segmental

presynaptic inhibition. Experimental Brain Research, 124(4), 474–480.

157

Appendix 1

Contribution of each author:

1. Ewa Niechwiej-Szwedo: design of the experiments, generation of hypotheses,

programming and generation of the stimulus, data collection, data analysis (including

statistical analysis), writing of the manuscripts.

2. Esther Gonzalez (advisory committee member): psychophysics and design of

experiments, generation of hypotheses, programming and generation of the stimulus,

comments on the manuscripts.

3. Agnes Wong (advisory committee member): discussion of experimental approach,

comments on the manuscripts.

4. Molly Verrier (supervisor): discussion of experimental approach, comments on the

manuscripts.

5. Martin J. Steinbach (supervisor): experimental approach to use the JM to manipulate

the gamma system, design of experiments, generation of hypotheses, comments on

the manuscripts.

6. Sivan Bega (summer student): data collection and analysis for experiment 3 in paper

1.

7. Bharat Bahl (summer student): data collection and analysis for experiment 1 in paper

3.