Research Article Computerized Neuropsychological Assessment in...

Research ArticleComputerized Neuropsychological Assessment in AgingTesting Efficacy and Clinical Ecology of Different Interfaces

Matteo Canini1 Petronilla Battista1 Pasquale Anthony Della Rosa1 Eleonora Catricalagrave2

Christian Salvatore1 Maria Carla Gilardi1 and Isabella Castiglioni1

1 Institute of Molecular Bioimaging and Physiology National Research Council (IBFM-CNR) Segrate 20090 Milan Italy2 Laboratory of Neuropsychology The Foundation of the Carlo Besta Neurological Institute IRCSS 20133 Milan Italy

Correspondence should be addressed to Pasquale Anthony Della Rosa pasqualedellarosaibfmcnrit

Received 18 April 2014 Revised 2 July 2014 Accepted 2 July 2014 Published 24 July 2014

Academic Editor Fabio Babiloni

Copyright copy 2014 Matteo Canini et al This is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Digital technologies have opened new opportunities for psychological testing allowing new computerized testing tools to bedeveloped andor paper and pencil testing tools to be translated to new computerized devices The question that rises is whetherthese implementations may introduce some technology-specific effects to be considered in neuropsychological evaluations Twocore aspects have been investigated in this work the efficacy of tests and the clinical ecology of their administration (theability to measure real-world test performance) specifically (1) the testing efficacy of a computerized test when response tostimuli is measured using a touch-screen compared to a conventional mouse-control response device (2) the testing efficacy ofa computerized test with respect to different input modalities (visual versus verbal) and (3) the ecology of two computerizedassessment modalities (touch-screen and mouse-control) including preference measurements of participants Our results suggestthat (1) touch-screen devices are suitable for administering experimental tasks requiring precise timings for detection (2) intrinsicnature of neuropsychological tests should always be respected in terms of stimuli presentation when translated to new digitalizedenvironment and (3) touch-screen devices result in ecological instruments being proposed for the computerized administrationof neuropsychological tests with a high level of preference from elderly people

1 Introduction

Computerized neuropsychological tests have been used inresearch for almost fifty years [1] Although many differenttest batteries have been developed and new batteries areintroduced every year for clinical screening not sufficientnormative data and standardized psychometric measuresare yet available [2] Conversely paper and pencil tests arewidely approved and are still regarded as keynote tools forneuropsychological assessment due to their high validity andreliability [3] Paper and pencil neuropsychological tests arebased on the presence of a neuropsychologist essential for theassessment of cognitive abilities especially for the evaluationof a person with brain injury or cognitive impairment andfor the selection administration and interpretation of testsAlthough suffering from some levels of subjectivity variabil-ity and long times (due to the fact that it is often necessaryto do a screening and also a diagnostic deepening) paper

and pencil tests have been validated for the administrationof reliable tests able to pinpoint a potential deficit involvinga specific cognitive ability or to discriminate among impair-ments in different cognitive domains [4]However neuropsy-chological evaluation can also provide information concern-ing normal brain functioning and allows monitoring thecognitive status of an individual especially throughout olderage Therefore its results are extremely important to trace acontinuum of normal functioning in the aging populationnot only in presence of pathologies Documenting changes incognition is indeed an important issue in neuropsychologi-cal assessment as the clinicianresearcher is often called uponto determine if and when cognitive functioning has changedAnother important advantage of the conventional paperand pencil neuropsychological assessment is their ecologicalvalidity In the context of neuropsychological testing ecolog-ical validity refers to the degree to which test performancecorresponds to real-world performance Validity does not

Hindawi Publishing CorporationComputational and Mathematical Methods in MedicineVolume 2014 Article ID 804723 13 pageshttpdxdoiorg1011552014804723

2 Computational and Mathematical Methods in Medicine

apply to the test itself but to the inferences that are drawnfrom the test [5 6] Most importantly clinical sensitivityallowing neuropsychologist to capture potential shades ina specific domain and to trace a specific cognitive profilemay results difficult to be translated in a computer-basedassessment Despite these numerous advantages traditionalpaper and pencil tools show some limitations particularlywhen assessing cognitive changes in a relatively short follow-up period The most commonly administered tests usuallydo not provide alternative forms of administration [7] thusprecluding to repeat testing over short intervals (eg lt6months) [8] Other specific limitations concern the intrinsicnature of the tools and include assisting the setting andmanually computing scoring by the experimenter [9] longduration of the assessment potential bias related to differentexaminers [10] impossibility to provide a precise time controlon stimuli presentation andor the lack of an accuratemeasurement of motor response accuracy [11] and greatercosts Although there are specific tests for the assessmentof attentional and executive functions which can evaluateindividual components and can diagnose specific deficits [12]some executiveattention abilities could take advantage fromthe use of a computerized tool in particular in the assessmentof response inhibition resistance to distraction planningproblem solving working memory operations and mentalset shifting divided attention

All of these limitations could be overcome by a comput-erized assessment on condition that efficacies and ecolog-ical measurements are carried out About these issues theAmerican Psychological Association (APA) has recognizedthe importance of computerized psychological testing andhas suggested how to implement and interpret computerizedtest results in its guidelines [13] Furthermore computerizedassessment of cognitive functions can be self-administeredand can have a shorter duration (eg by reducing ldquodeadrdquotimes in stimuli presentation) They may have great valid-ity and reliability due to their great objectivity precisionand standardization Computerized performance can alsominimize the so called ldquofloor and ceiling effectsrdquo occurringwhen differences among participant performance are notfully captured thus they can provide more standardizedmeasures of subject performance crucial for example for anaccurate and early detection of specific pathological disease(eg dementia) [14]

It appears clearly that computerized testing will representan essential part of the clinical setting in the nearest futureabove all in screening procedures on condition that thesenew instruments and their results are governed by experts

Feasible efficacious and ecological computerized test-ing could allow clear pictures of normal cognition to bemeasured and monitored also at home pinpointing specificdeficits in each cognitive domain in aging people Validatedcomputerized tools could also provide stronger groundingto overcome the lack of a consensus regarding the feasi-bility and testing efficacy related to the different types oftechnological solutions and settings andor response layoutschosen for the assessment For instance a more ecologicaltechnological solution can overcome one important obstacleto the wide use of computerized assessment attributable to

the familiarity with technological devices [11] while a moreecological setting in the nearest future could partially allowthe administration of the tests without the support of aspecialized clinician or the tests could be potentially self-administered or assessed by a caregiver at home

The aims of this study were to evaluate the following

(1) the testing efficacy of a computerized test (in arepresentative case of executive function assessment)when response to stimuli is measured using a touchscreen compared to a conventional mouse-controlresponse device

(2) the testing efficacy (the efficacy of the neuropsycho-logical test) of a computerized test (in a representativecase of memory function assessment) with respect totwo different input modalities a visual presentationmodality of the test replicating the most diffuseddigital versions of the test and a verbal presentationmodality replicating the classical clinical administra-tion of the test

(3) the ecology of both computerized assessment modal-ities (touch-screen tablet and mouse-control PC)including preference measurements of participants

2 Related Works

Computerized cognitive batteries are already available andused for the screening and the assessment of dementiaSome of these instruments were appropriately created for theassessment of cognitive decline in dementia other ones wereadapted to fulfill this role in aging Some tools are designedfor research use and others tools have been designed mainlyfor clinical use some of them already implemented in clinicalguidelines

An interesting review [2] reported seventeen comput-erized test batteries used in the measurement of cognitiveabilities of adults Some of these tools are able to run only ona PClaptop and others are available only on web sites

Among these tools

(1) CNS vital signs [15] battery is developed as a routineclinical screening instrument It includes seven testsverbal and visual memory finger tapping symboldigit coding Stroop test test of shifting attention andcontinuous performance test

(2) CogState [16] battery was developed as a dementiascreening instrument and it is implemented as a cardgame formThe participant plays different games thatare adapted accordingly to performance CogStaterequires an active internet connection to generatea report The participantrsquos data are uploaded andanalyzed Then a report is generated and e-mailedback to the provider

(3) NeuroTrax [17] includes custom software on thelocal testing computer and serves as a platform forinteractive cognitive tests that provide precise accu-racy and reaction time data The level of difficultyis graded The NeuroTrax tests different cognitive

Computational and Mathematical Methods in Medicine 3

Table 1 Features of computerized cognitive batteries cited in Section 2 Hardware input modality and context of use information areprovided

Battery name Hardware used Input modality UseCNS vital signs PClaptop Keyboard Research use

CogState PClaptoptablet Keyboard Research use(Web based)

Neurotrax PClaptop (web based) Keyboardmouse Clinical use (AACN consensus)IntegNeuro PClaptop Keyboard Research useTouch panel dementiaassessment scale PClaptop Touch screen Research use

CADi iPad Touch screen Research useCANTAB mobile iPad Touch screen Research use

domains memory (verbal and nonverbal) executivefunction visual spatial skills verbal fluency attentioninformation processing and motor skills The testswere designed for usewith the older adults Responsesare collected with the mouse or with the number padon the keyboard NeuroTrax also generates a reportcontaining raw and standardized scores immediatelyfollowing testing Administrative features are webbased

(4) IntegNeuro [18] in brief investigates the followingdomains of cognitive function sensorimotor ver-bal and language memory executive planning andattention Scoring of responses is obtained by usingan automated software program Trained researchassistants conducted the hand scoring of some testsand oversight is implemented to monitor accuracy

(5) Touch panel dementia assessment scale [19] hardwarecomprises a 14-inch touchpanel display and computerdevices built into one case The TDAS runs on Win-dows OS and was bundled with a custom programmadewith reference to theADAS-cogwith the elderlyunder the control of a physician

More recently screening tools for assessment of cognitiveimpairment able to run on an iPad tablet device have madetheir appearance as well They can be downloaded and areself-administrable (see Table 1)

CADi [20] and CANTAB mobile [21] are the first twotools which exploited this technology The CADi consists of10 very brief tests the purpose is to provide a mass screeningin the Japanese population in a relatively short time andwith overall cost substantially lower than paper and pencilbased examination The most important limit of this tool isthe cultural background underlying the validation as it isavailable only in Japanese language Notwithstanding thereare 18 different language versions of the CANTAB mobiletool it comprises focused screening tests able to investigateonly episodic memory and learning abilities (PAL task)

3 Materials and Methods

31 Participants A group of 38 healthy participants wasrecruited among the Italian elderly population (20 males age

range = 53 87 age mean = 64474 standard deviation [SD] =plusmn8462 education range = 5 19 education mean = 11263[SD] = plusmn4131)

Participants with any history of neurological illnesses andaMinimental state examination (MMSE) score lower than 26were excluded from the study (MMSE range = 26 30 MMSEmean = 28711 [SD] = plusmn1183) (see Table 2) Informed consentwas obtained from all participants

32 Hardware Device In order to compare the participantresponses from neuropsychological assessment deliveredthrough mouse-control PC and touch-screen tablet we usedan Asus T100T notebook PC (CPU quad core Intel Atomprocessor RAM 2GB Screen 10110158401015840HD and screen reso-lution 1366lowast768 with multitouch) This device consists of a101 inches tablet running a Windows 81 OS which can beused as a standalone touch-screen device or combined witha mobile dock as a ldquostandardrdquo PC with conventional inputperipherals (mouse and keyboard)

33 Software and Scripting The experimental tasks wereimplemented and administered using the presentation soft-ware (httpwwwneurobscom) This is an object-orientedprogramming language allowing a sharp control on stimulipresentation (eg objects presentation timing and random-ization of trials) and on response tracking Experimentalparadigms are set up through scripts consisting of two logicalcomponents (i) a Scenario Definition Language (SDL) wherethe objects of the task (ie stimuli organizedwith hierarchicallevels of complexity) are defined along with their specifictiming and responsematching and a (ii) PresentationControlLanguage (PCL) where objects presentation parameters arecontrolled (eg possibility to present stimuli through loopsconditionals and subroutines for trials randomization) Anadditional third component the SDL header is used tospecify general parameters that is those values which willbe used as default for the presentation of all stimuli unlessa different specification is given in the stimulus definition(eg default font size background color stimuli duration andtiming logfiles specifications and so forth)

Presentation tracks and stores information about stimuliadministration timing and participant response behaviors(ie the response given by the subject) in terms of response


Table 2 Age education and MMSE scores are provided for the whole (119873 = 38) group (left column) Independent sample Studentrsquos 119905-testsvalues for these variables are shown for the AVLT (verbal versus visual) inputmodality groups (center column) and for the ANT (touch versusmouse) response modality groups

Overall group119873 = 38

20M18 F

AVLT groups119873 = 38

visual group (11M8 F)verbal group (9M10 F)Studentrsquos 119905-test (119875 values)

ANT groups119873 = 38

touch group (10M9 F)mouse group (9M10 F)Studentrsquos 119905-test (119875 values)

Agemean 64474

0970 0831st dev 8462range 53 87

Educationmean 11263

0487 0970st dev 4131range 5 19

MMSEmean 2891

0344 0272st dev 067range 28 30

classification and reaction times Information is subsequentlyanalyzed using ad hoc data processing procedures in order toassess participant performance

Parameters to be measured for response behaviors ofparticipants are defined by the experimenter and specifiedin terms of (i) input devices used by the participant and (ii)response buttons in GUIrsquos dedicated response setting panelassociated to the selected device to be recalled in the SDLcomponent Contextually to our experimental setting twoldquoresponse layoutsrdquo (ie mouse-control or touch-screen) wereconfigured in the following manner

(i) A ldquomouse response layoutrdquo where the docking-station built-in trackpad mouse was specified asinput device and pressing one of the two mousebuttons was specified as response behavior twodistinct classes of responses were coded each oneassociated to a specific button press (ie pressing theleft and right mouse buttons) Each response classwas recorded and directly matched with the stimulustarget response in order to assess the responsesaccuracy

(ii) A ldquotouch response layoutrdquo where the capacitive screensurface was specified as input device and the screenpress was specified as response behavior in thiscase this was the only available class of responseThe two (ie left and right) response classes wereobtained through data after processing 119883 and 119884coordinates were registered each time the subject gavea response to a target stimulus and the 119883 coordinatewas used to estimate whether the response occurredon the left or right portion of the screen positive119883 corresponded to right touch screen press andnegative119883 corresponded to left touch screen press

34 The Neuropsychological Tests

341 The Attentional Networks Test (ANT) The attentionalnetwork test (ANT) [22] is designed to test three differentattentional networks namely the executive control and thealerting and the orienting components of attention

Each trial starts with a fixation point and ends with thetarget stimulus consisting of an array of five contiguousarrows the participant is asked to state as fast as possible thedirection (ie left or right) of the central arrow (see Figure 1)

The array can be defined according to a congruencyfactor (depending on the direction of the arrows flanking thecentral target stimulus) and a spatial factor depending uponwhether the array can occur either above or below the fixationpoint or at the center of the screen further the presentation ofthe array can be cued or not by an asterisk (see Figure 1 for adetailed description of all experimental conditions) Differentcombinations of these experimental factors are selectivelyused to assess the efficiency of the three different networks(ie executive control and alerting and orienting component)through accuracy and reaction times (RT) analysis as follows

(i) Conflict effect (CE) is assessed by subtracting RTsbelonging to congruent trials (flanking arrows point-ing in the same direction of the central target arrow)from RTs belonging to incongruent trials (flankingarrows pointing in the opposite direction of thecentral target arrow)

(ii) Alerting effect (AE) is calculated by subtractingmeanRT of the double cued trials (target trial presentationis preceded by two spatial cues occurring contem-porarily above and below the fixation point) fromthe mean RT of the no cue trials (target trial is notpreceded by any cue)

(iii) Orienting effect (OE) is calculated by subtracting themean RT of the central cue condition (target trial


Neutral Congruent Incongruent

No cue Center cue Double cue Spatial cue

+ +

+

+

lowast

lowast

lowast

lowast

lowast

(a)

400ms

100ms

1700msuntilresponse

+

+

+

lowast

(b)

Figure 1 (a) Attentional network test experimental conditions are shown each trial presented during the task is a combination of congruency(3 levels on top) and cue (4 levels below) conditions (adapted from [22]) (b) An example of spatially cued incongruent trial is presentedstimuli timings and interstimulus interval of this customized version of the task are provided below each stimulus (adapted from [23])

is preceded by a cue occurring at the center of thescreen) from the mean RT belonging to the spatialcue condition (target trial is preceded by a spatial cueoccurring either above or below the fixation pointdepending on the position of the incoming targettrial)

342 The Auditory Verbal Learning Test (AVLT) The ReyAuditory Verbal Learning Test [24] assesses the long-termverbal memory or the ability to learn and store in long-term memory unstructured verbal material It consists of 5consecutive repetitions for learning the material and then along delay free recall 15 minutes later The test consists of 15unrelated words presented orally First the participant hasto memorize this list of 15 words and immediately try torecall them after each presentation (immediate recall IR)Second after 15 minutes in the meanwhile the participantis performing nonverbal tests the participant is asked torecall the same list (delayed recall DR) Finally the subjectis presented with a longer list of distracter words and askedto recall the previously learned words (recognition task) TheItalian test has three alternatives but completely equivalentlists [25]

An interference task not involving verbal or memorycognition is administered between the third (ie last) IRsession and the DR session

35 Testing Efficacy I Executive Function Assessment (ANT)Given our purpose to compare the testing efficacy of acomputerized test administered by touch-screen tablet versusconventional mouse-control PC we created two versions ofthe ANT task the first one for the mouse layout responseand the second for the touch layout response (see Section 33Software and Scripting)

Both versions were programmed using the presentationsoftware and shared the exact same picture stimuli stimulitimings and randomization criteria they were thus virtuallyequal differing only for the input devices layout configura-tion

351 Mouse-Control PC Implementation Each experimentalcondition of the ANT task was represented by 8 trialsresulting in a total of 96 [(3 congruency times 4 cue) times 8] trialsfor each experimental condition

To this extent an array of 96 trials was specified eachtrial consisting of a variable number of component stimulidepending on the experimental condition

Each trial was thus specified in terms of (i) stimuli objects(ii) time intervals occurring between successive stimuli and(iii) target response (see Figure 1)

The trial presentation order was specified using a fullyrandomized design A set of three different interstimulusintervals (ISI) (min value 1873ms max value 4964ms) wasalso specified this means that a different time interval couldoccur between successive trials in order to avoid habituationeffects Two response buttons were assigned and the par-ticipant was required to press as fast as possible the left orright mouse button (depending on the target trial type aspreviously described)

Responses accuracy was coded in a logfile by comparingparticipantrsquos response to a given target stimulus to theexpected response for that stimulus coded in the stimulusparameters in the SDL

Stimuli presentation and response timing and accuracywere also recorded

352 Touch-Screen Tablet Implementation A touch versionof the ANT task was implemented using the same numberof trials experimental design and randomization order ofthe ANT task in the mouse-control PC implementationThe only difference between the touch layout version andthe mouse layout version resides in the response deviceassigned to the participant While in the mouse versionthe participant was asked to respond by pressing the twomouse buttons in this touch version of the task heshe wasrequired to press the leftmost or right most part of the screenwith the thumbs handling the touch device with both thehands Touch location was coded as previously described


(eg negative119883means a touch occurred in the left portion ofthe screen and corresponds to the left mouse button press)

36 Testing Efficacy II Memory Function Assessment (AVLT)Given our purpose to assess the efficacy of different modali-ties (ie verbal versus visual implementation) of a computer-ized test administered by a touch-screen tablet two versionsof the AVLT were created (1) one verbal version (VerbalAVLTTask) consisting in the presentation of auditory stimuliand requiring the subjects to verbally recall answers (whichwere both recorded and coded by the tester) thus replicatingthe conventional administration of AVLT used in clinicalcontext and (2) one visual version (Visual AVLT Task)consisting in the presentation of visual list of words on thescreen and requiring the subjects to recognize target stimuli bytouching the screen when a presented target word appearedWe have chosen not to calibrate the level of difficulty forthe recognition test stimuli to be equivalent to that of therecall stimuli in order to detect any differences between thetwo stimuli presentation modalities including the level ofdifficulty

361 AVLT Equivalent Lists Creation Since both lexical andpsycholinguistic variables can influence behavioral perfor-mance the characteristics of the original 15 Rey list wordswere extracted and analyzed in order to create 5 alternativesbut completely equivalent lists Words frequency values forthe original 15 words were determined from the Italianlexicon [26] while for the familiarity (FAM) concreteness(CNC) age of acquisition (AoA) and imageability (IMG)values were extracted from the MoA database [27] Foreach variable the mean of the distribution of values forthe fifteen words and the interquartile range (75ndash25) wascomputed Only the words falling in this range and withvalues for at least three variables overlapping with those ofeach word were selected Following this criteria 50 wordswere selected and matched to 10 original words included inthe original Rey listThe 50 words were subsequently dividedinto 5 lists each containing 10 words Lists were balancedand statistically matched to the original Rey list A one-way ANOVA analysis revealed no significant difference for(i) word length (119865(5 54) = 0095 119875 = 0993) (ii) FAM(119865(5 54) = 0856 119875 = 0517) (iii) CNC (119865(5 54) = 1146119875 = 0348) (iv) AoA (119865(5 54) = 1 416 119875 = 0233) and(v) IMG (119865(5 54) = 1096 119875 = 0373) Bonferroni post hoccorrections for pairs of lists revealed no significant differencebetween lists for none of the above-mentioned variables (all119875values lt1) Semantic or phonemic similarities between wordswithin each list were also excluded

Among these five lists two were selected as target stimulilists while words belonging to the remaining three lists wereused as filler words for the recognition part of the visualversion of AVLT (see Section 363) Each of the two-targetlists was pseudorandomized according to three differentwordorders to control for possible word list sequence effectsThusone out of the resulting 6 lists (ie two words sets eachrandomized three times) was selected as target list for eachparticipant in both versions of the task

362 Verbal AVLT Task by the Touch-Screen Tablet A list of10 target words was presented verbally to the participantsSpecifically the words were administered as auditory signaland the sound was generated using a text-to-speech software(Audacity the free cross-platform sound editor softwarehttpwwwaudacitysourceforgenet) and recorded refinedand normalized in order to equalize all words in terms ofvoice quality and volume

After each presentation the participant was asked toverbally recall as much words as possible

The participant was tested three times consecutivelyas in [19] for the immediate memory recall and once forthe delayed memory component Vocal recordings of targetwords were used as stimuli in this task version and wereadministered through the integrated speakers of the touchscreen tablet Stimuli were presented consecutively with aninterstimulus interval (ISI) of 3000msoccurring between onestimulus and the next one Recall sessions were tested usinga voice recording script in order to store an audio logfileof participant responses Parallel to the sound recordingthe 10 target words were constantly displayed on the screenonly for the experimenter view allowing making an onlinemonitoring of given responses by the experimenter (thismade the task not purely computerized)

363 Visual AVLT Task by the Touch-Screen Tablet In thevisual version of the task a total number of 10 target wordswere presented to the participant Immediately after viewingthese 10 target stimuli heshe was presented with the same 10words randomized together with another set of 10 fillers (iewords not present in the targets lists) and asked to touch thescreen each time heshe recognized an item belonging to thepreviously presented list The same setting was used for thedelayed part of the task with the only exception that heshewas presented with the same 10 words randomized togetherwith another set of 20 fillers including the 10 fillers presentedbefore We chose 20 fillers for the delayed task because wetripled the number of the target words (10)This approach hasbeen adopted in the original paper and pencil Rey auditoryverbal test [24]

Prior to task administration the experimenter handedthe tablet to the participant who was therefore activelyrequired to use the testing apparatus while the experimenterwould only passively control upon subjectrsquos performance

Stimuli were presented as white words (with a font sizeof 36 points) at the center of a black screen and lasted for3000ms an ISI of 2000ms occurred between one stimulusand the next one

Accuracies were estimated by comparing each stimuluscode (ie target or filler) to subject response and stored in alogfile

37 Experimental Design for Testing Efficacy A combinationof one AVLT version (ie visual or verbal) and one ANT (ietouch or mouse) layout was administered to each subjectresulting in a factorial design in which (i) 19 subjects (10males) performed the touch response layout of ANT and theother 19 subjects (9 males) performed the mouse responselayout and (ii) 19 out of 38 subjects (11 males) performed


the visual version of AVLT while the remaining 18 subjects(9 males) performed the verbal version of AVLT

Groups were statistically matched for age educationand MMSE and no significant difference emerged whencomparing the (i) mouse versus touch group (age 119875 = 0831education 119875 = 0970 and MMSE 119875 = 0272) and the (ii)verbal versus visual group (age 119875 = 0970 education 119875 =0487 and MMSE 119875 = 0344)

AVLT lists were randomly and evenly distributed acrosssubjects for both verbal and visual versions of the task Eachsubject was first asked to complete the immediate memoryrecallrecognition of AVLT and was tested 15 minutes laterfor the delayed long-term retrieval component [24]

Two experimental runs of ANT (7 minutes each) wereadministered in between the two AVLT components asinterference task that is a task critically involving neither theverbal nor the learning cognitive resources recruited duringAVLT

38 Ecology Satisfaction Survey After the administrationof the tests independently from the test (AVLT or ANT)and response layout (touch or mouse) participants wererequired to complete a satisfaction survey The surveyis a 16-item self-report questionnaire that uses a 5-pointLikert scale (for the complete list of items of the ques-tionnaire see Supplementary Materials available online athttpdxdoiorg1011552014804723)

The survey measured the following the participant fre-quency of use of touch-screen tablet and of mouse-controlledPC (items 1 2 and 3) the participant qualitative perceptionof the familiarity with the touch-screen tablet (item 4)the participant qualitative perception of the comfortablenesswith the touch-screen tablet (items 5 6 and 7) the participantqualitative perception of the testing environment (items 8and 9) the participant fatigue of using sensory functionswhile interacting with the touch-screen tablet (items 1011 12 13 and 14) the participant fatigue of maintainingthe concentration while interacting with the touch-screentablet (item 15) and the participant time perception of theneuropsychological tests (item 16)

Items administration was customized depending on theexperimental (ie different combinations of touch or non-touch versions of AVLT andANT) setting of each participant

4 Data Analysis

41 Testing Efficacy I Executive Function Assessment Inorder to assess possible differences between the two settings(touch-screen tablet versus mouse-control PC) we evaluatedthe recorded hardware uncertainties to control for specificsetting effects (touch versusmouse) on stimulus presentationHardware uncertainties are provided by the presentation soft-ware giving information on how the hardware is managingstimuli presentation and if some hardware-based source ofvariability is altering the script management by the softwareSpecifically for each Presentation event in the logfile (exceptfor pause resume and quit events) presentation provides atime of occurrence119879 (ms) and an uncertainty 119889119879 (ms)Thesetwo numbers provide bounds on the time of occurrence of

a presentation event If Hardware uncertainties 119889119879 remainless than 06ms the response script is running as expected(httpwwwneurobscom Presentation help) Metrics mea-sured by hardware uncertainties refer to time responses ofthe device to the stimuli software presentation independentlyfrom the subject response to stimuli presentation

411 Data Screening Concerning the ANT we consideredthe distributions of reaction times (RT) and accuraciesSpecifically with respect to reaction times (RT) any trial withrecorded RT that fell two SDs above or below the calculatedRT mean (RT ge mean + 2SD) (RT le mean plusmn 2SD) wasrejected

With respect to accuracy each experimental conditionwith measured accuracy that fell under 80 was rejected

412 Statistical Analysis Based on the selected datasets weperformed a 3 times 2 repeated measures ANCOVA with effecttype (ie conflict effect alerting effect and orienting effect)as within-subjects factor with three levels

A between-subjects factor was considered namely theparticipant group with two levels (mouse versus touch lay-out) Although groups were sampled with comparable valuesof MMSE age and education (see Section 37 ExperimentalDesign for Testing Efficacy for groups comparisons statistics)these variables were included as covariates in order toaccount for their potential influence on task performanceWeassessed the main effect of group of participants covariatesand exclusively interactions with the between-subjects factor(ie group of participants)

42 Testing Efficacy II Memory Function Assessment Con-cerning the verbal AVLT vocal recordings were listened toand classified as correct or incorrect while for the visualAVLT responses to target stimuli were stored in a logfile andsuccessively coded as correct or incorrect

Accuracies for the three immediate recall sessions (ieIR-1 IR-2 and IR-3) and the delayed recognition session (DR)were coded as percentages Analyses of equality betweenverbal and visual AVLT were performed with a Mann-Whitney 119880 test (nonparametric) for independent samples(ie according to AVLT verbal or visual condition) on thearcsine-transformed percentages of accuracy (for each recall)and the delayed session independently Correlations betweenthe MMSE scores and the arcsine-transformed percentagesof correct responses for IR-1 IR-2 IR-3 and DR were thenevaluated using Kendallrsquos tau correlation coefficient split byAVLT verbal and AVLT visual conditions

43 Ecology Firstly we performed one sample Wilcoxonsigned rank test for each of the 16 items measured by thesurvey administered to the participants versus the middlelevel of perceived scale quality (3 with respect of a scale max-imum of 5) Furthermore we investigated the relationshipbetween the level of preference of participants for the testingenvironment (ie item 7 using the touch-screen tablet withrespect to an external device such as a mouse or a keyboard)


and the perceived degree of easiness when using a touch-screen tablet (ie item 5) or the perceived degree of easinesswhen touching the screen (ie item 6)

5 Results

51 Testing Efficacy I Executive Function Assessment Noanomaly linked to hardware management of stimuli presen-tation was detected when screening both mouse and touchlogfiles

511 Data Screening All subjects had an RT outlier percent-age lt30 on each experimental condition and accuracieswere gt80 for all participants no subject was thereforeexcluded from the analyses

512 Statistical Analysis Results Mean RTs for trial condi-tions for the calculation of each of the three effects are sum-marized in Table 3 The main effect of group of participantswas found not to be significant (119865(1 33) = 0008 119875 =0929) revealing that performances are comparable in termsof overall RTs for the ANT with mouse or touch responselayout No significant main effects of covariates were found(MMSE 119865(1 33) = 0049 119875 = 0826 age 119865(1 33) = 1221119875 = 0277 and education 119865(1 33) = 0359 119875 = 0553)

The interaction between group of participants and effectsshowed a trend toward significance (119865(2 66) = 2357 119875 =0116 Greenhouse-Geisser correction) and a plot of this 2-way interaction showed a reduced CE and an incremental AEand OE effect for the touch group (see Figure 2) pairwisepost hoc comparisons revealed a trend towards significancewhen considering CE differences between groups (meandifference Touch minus Mouse = minus37280 119875 = 0170) in termsof a reduced CE for the touch while the two groups differedto a lesser degree for the OE (Mean Difference Touch minusMouse = 18218 119875 = 0232) and AE (Mean DifferenceTouch minusMouse = 16418 119875 = 0261)

52 Testing Efficacy II Memory Function Assessment No sig-nificant differences were observed between verbal and visualrecallrecognition performances for IR-1 (Mann-Whitneytest 119875 = 0365) IR-2 (Mann-Whitney test 119875 = 0293) IR-3 (Mann-Whitney test 119875 = 0694) and DR (Mann-Whitneytest 119875 = 0988)

For AVLT verbal MMSE scores correlated with perfor-mance on IR-1 IR-2 immediate and with the delayed recallsession a trend towards correlation was found for the thirdimmediate recall session

For AVLT visual MMSE scores showed a correlationtrend with IR-1 and a significant correlation with IR-2 IR-3and DR

ForAVLTverbal all immediate anddelayed recall sessionssignificantly correlated with age

For AVLT visual recognition sessions did not signifi-cantly correlate with age

For AVLT verbal education scores significantly corre-lated with all the immediate and the delayed recall sessionsFor AVLT visual education scores did not correlate withimmediate recall sessions while a significant correlation was

200000

150000

100000

50000

000

Mea

n RT

s (m

s)

Flanker groupEffectsConflict effectOrienting effect

Alerting effect

Mouse Touch

Figure 2 Plot of 2-way interaction group (mouse or touch)119883 effect(conflict effect orienting effect and alerting effect) Mean RTs (119910-axis) for conflict Effect (Blue Line) orienting Effect (Green Line)and alerting Effect (Brown Line) are shown for mouse (leftmostdots) and touch (rightmost dots) performances (119909-axis)

found with education for the delayed recall session Table 4summarizes correlations coefficients and statistical signifi-cance between each AVLT recall session and socio demo-graphics (age and education) and cognitive index (MMSE)variables tested

53 Ecology Wilcoxon one sample signed rank test indicatedthat themiddle percentage of quality ratings was significantlylower than 3 for item 1 (ldquohow much do you use the tabletin your daily liferdquo 119875 = 0003) for item 4 (did you feeluncomfortable using the tablet 119875 = 0000) for item 10 (didyou feel fatigued while handling the tablet 119875 = 0000) foritem 12 (did you feel fatigued while touching the screen119875 = 0000) for item 13 (did you feel fatigued while listeningto vocal recordings 119875 = 0005) for item 14 (did you feelfatigued producing a vocal response 119875 = 0001) and foritem 16 (did you feel the experiment had a too long duration119875 = 0000) which is indicative of very low levels of uneasiness(ie item 4) or fatigue of using sensory functions (ie items12 13 and 14) and a perceived very long duration of testadministration (ie item 16) notwithstanding a very lowuse of a touch-screen tablet in everyday life (ie item 1)The relationship between the preference for using the touchscreen (ie item 7) and the perceived degree of easinesswhen using a tablet (ie item 5 Kendallrsquos tau = 0484 119875 =0009) or the perceived degree of easiness when touchingthe screen (ie item 6 Kendallrsquos tau = 0397 119875 = 0025)were both significant (Kendallrsquos tau = 083 119875 = 002) (seeSupplementary Materials online for the complete items list)


Table 3 Mean RTs (along with SD) for each experimental condition used to calculate effects of ANT are shown in the table for both touch(upper panel) and mouse (lower panel) versions of the task Congruent trials are subtracted from incongruent trials in order to calculateconflict effect double cue trials are subtracted from no cue trials in order to calculate the alerting effect and spatial cue trials are subtractedfrom central cue trials in order to calculate the orienting effect

Trial type Mean SD Trial type Mean SD Effect Mean SDTouch layout

Incongruent 1016229 122973 Congruent 890436 125071 Conflict effect 125793 61692No cue 93642 122244 Double cue 914954 127472 Alerting effect 21466 496Central cue 932806 117579 Spatial cue 908124 129387 Orienting effect 24682 34646

Mouse layoutIncongruent 941746 181163 Congruent 786627 109758 Conflict effect 155118 89207No cue 83305 125772 Double cue 828475 135419 Alerting effect 457 34988Central cue 830022 129443 Spatial cue 820197 140739 Orienting effect 9825 54498

Table 4 Results show correlations coefficients (upper values in cells) and statistical significance (lower values in cells) between each AVLTrecall session (ie IR-1 IR-2 IR-3 andDR) and sociodemographics (age and education) and cognitive index (MMSE) variables testedThe leftmost panel of table shows statistics belonging to the group tested with verbal version of the AVLT while the rightmost panel shows statisticsbelonging to the group tested with the visual version of the task

Verbal VisualIR-1 (120591) IR-2 (120591) IR-3 (120591) DR (120591) IR-1 (120591) IR-2 (120591) IR-3 (120591) DR (120591)

MMSE 0633 0492 0373 0406 0246 0579 0451 0407(119875 = 0001lowastlowast) (119875 = 0013lowast) (119875 = 0040lowast) (119875 = 0069) (119875 = 0195) (119875 = 0003lowastlowast) (119875 = 0020lowast) (119875 = 0035lowast)

Age minus054 minus0479 minus0434 minus0353 minus0137 minus019 0007 minus0216(119875 = 0031lowastlowast) (119875 = 0008lowastlowast) (119875 = 0050lowast) (119875 = 0240) (119875 = 0448) (119875 = 0303) (119875 = 0970) (119875 = 0240)

Education 0448 0527 0408 0404 0183 025 0164 038(119875 = 0017lowast) (119875 = 0006lowastlowast) (119875 = 0039lowast) (119875 = 0034) (119875 = 0332) (119875 = 0197) (119875 = 0409) (119875 = 0049lowast)

(120591 = Kendallrsquos tau lowastCorrelation is significant at the 005 level 2-tailed)(lowastlowastCorrelation is significant at the 001 level 2-tailed)

6 Discussion

As general consideration for the purpose of our work wehave chosen two representative tests to assess two mainobjectives

We investigated the memory domain given that theearlier cognitive symptoms reported in Alzheimerrsquos diseasethe most common form of dementia involve memory [28]Therefore the majority of tools have focused mainly onthis cognitive ability and have implemented tests tailoredat investigating memory impairments of which the mostclinically validated and commonly employed in a clinicalsetting is the AVLT test Furthermore memory impairmentsin particular are a cardinal feature of themajority of dementiasyndromes

We investigated the executive function domain by the useof the ANT as a useful test for assessing differences betweenresponses given that it encompasses three different effectsrelying on three different cognitive mechanisms (ie conflicteffect alerting effect and orienting effect) [29]

The first aim of the present study was to evaluate the test-ing efficacy of a computerized neuropsychological assessmentwhen implemented on a touch-screen device

To this aim we created and tested two different experi-mental settings in terms of response layout specifically twoidentical versions of an experimentally validated attentional

task (ie ANT [22 23]) differing only in terms of responsemodality were implemented one version of the task requiredsubjects to give a response with a mouse device while theother one by using a touch screen This was made in orderto directly compare the testing efficacy of a psychologicaltest (in our case the evaluation of executive function) whenadministered by a touch-screen device with respect to a moreconventional mouse-control PC

Comparisons of reaction times between subjects usingtouch screen or mouse and of their test performance revealedno significant overall differences suggesting that touchscreen and mouse can be equally chosen as response devicessince they grant the same experimental outcome Thesefindings strengthen the results highlighted by Sears andShneiderman [30] who although under different experimen-tal conditions compared touch-screen response layout versusmouse response layout Their results suggested substantialcomparability between these two input devices Our findingconsistently with results from other authors is of particularinterest given that the touch-screen technology is currentlywidely spreading also among the elderly population Touch-screen tablets are innovative technological solutions whichare emerging also as devices for healthcare interventionHealthcare services are indeed progressively showing anincreasing interest in translating services into touch-screenbased environment [31 32] From this perspective it is an


important topic to test if such devices can guarantee thesame testing efficacy of more conventional and extensivelyvalidated devices as mouse-control PC which are still todayused in clinical environment to administer neuropsycholog-ical tests Our work despite being validated on a limitednumber of subjects suggests that this technological solutionis feasible for test administration

The second aim of our work was to evaluate the testingefficacy of a computerized test (in a representative case ofmemory function assessment) when administered by thetouch-screen tabletwith respect to twodifferent experimentalsettings in terms of stimuli presentation

To this aim we created and tested two different versionsof AVLT a widely standardized and validated neuropsycho-logical test (i) a visual version replicating the visual portingof this task which is currently used by a set of differentdigital neuropsychological training batteries [19 33] in thisversions stimuli were presented visually and subjects wereasked to recognizememorized stimuli amongother nontargetstimuli and (ii) a verbal version replicating the ldquoclassicalrdquoadministration of the task in the clinical context in thisversion stimuli were presented verbally and subjects wereasked to freely recall all the memorized target stimuli Thiswas done with the purpose to assess the effects of the stimulipresentation modality on the efficacy of test (in our case theevaluation of memory function)

Our results for the verbal version of AVLT showedsignificant correlations between MMSE scores and perfor-mance on IR-1 IR-2 immediate and with the delayed recallsession A trend towards significance was found for the thirdimmediate recall session (IR-3) Similarly results for thevisual version of AVLT showed a correlation trend betweenMMSE scores and performance with IR-1 and a significantcorrelationwith IR-2 IR-3 andwith the delayed recall sessionFor AVLT verbal all immediate and delayed recall sessionsshowed an inverse and significant correlation with age (ielower scores of age correspond to higher values of recallperformance) education scores significantly and positivelycorrelated with all immediate and the delayed recall sessions(ie higher values of education correspond to higher valuesof recall performance) For AVLT visual a significant anddirect correlation with recognition performance scores wasfound only with education for the delayed recall sessionand no significant correlations were found between age andperformance on both immediate and delayed recognitions(see Table 4 for detailed results)

Overall these results suggest that both implementations ofthe test (ie visual and verbal) are affordable measures of thegeneral cognitive status directly correlated with a measureof general cognitive status assessment (ie MMSE) fromthis point of view they can be both considered affordabletools for a broad cognitive assessment In spite of this onlythe verbal version of the task showed a correlation withthe sociodemographical data of our sample (ie an inversecorrelation with age and direct correlation with education)

It should be in fact recognized that we have not comparedverbal with ldquopurerdquo visual stimuli presentation since fillerwords were presented during the task under this light thetwo tasks share some common features but differ for others in

terms of both experimental setting and underlying cognitiveprocesses

In fact the verbal version requires an active retrievalfrom memory of the presented stimuli while the visualversion requires the recognition and discrimination of thememorized target stimuli among other nontarget stimuliFrom this point of view they require the subject to usedifferent strategies to be solved and thusmay involve differentbrain networks (eg [34]) For example from the memorypoint of view the visual modality could be easier since thesubject is provided with a cue (namely the target stimulusis directly presented to the subject) but at the same timeit requires the inhibition of distracters and may be moredifficult in terms of executive functioning cognitive load(namely the subject has to discriminate the target stimulusamong other nontarget words)

Another crucial aspect concerning the introduction ofnew technological solutions in the everyday life is the degreeof ecology and the level of preference of the computer-ized assessment modalities regarding the administration ofthe neuropsychological tests Although some authors foundsignificant draw backs to touch screens in the elderly [35]others (eg [31]) reported that touch-screen devices are idealinstruments for assessing populations with low technologicalfamiliarity such as elders and patients Our results on elderlyand healthy participants confirm this finding consideringthat our subjects felt comfortable using the touch-screendevice and did not experience unease or fatigue feelings whileperforming the tests Crucially all subjects possessed lowfamiliarity with such devices and in some cases it was theirfirst experience of physical interaction with a touch-screentablet

Given the performance comparability between responsesusing mouse and touch it is important to introduce someconsiderations for future evaluation regarding whether (andunder which circumstances) it is preferable to choose one orthe other response layout

While no main effect of group was highlighted ouranalysis revealed that subjects performing theANT task usingthe touch response layout showed a tendency towards anadvantage for all three effects accounted by the task andnamely a trend towards a significant reduction of conflicteffect and slightly larger alerting and orienting effects (seeFigure 2 and Table 3 for details)

However it must be acknowledged that RTs from whichthe effects are derived showed the same pattern for both themouse and the touch layout namely for the conflict effectRTs are the longer with respect to congruent ones while foralerting effect RTs measured following the double cue wereshorter with respect to when no cue was presented and fororienting effect responses after a spatial cue were faster withrespect to a cue presented centrally This means that there isno difference in terms of performance for all ANT conditionsHowever when assessing specific cognitive processes mea-sured through the difference between RTs it appears that thetouch device may provide some benefits mainly on cognitivecontrol enhancing performance on themore demanding trialtype Namely conflict effect is calculated by subtracting RTsbelonging to congruent trials (flanking arrows pointing in


the same direction of the central target arrow) from RTsbelonging to incongruent trials (flanking arrows pointingin the opposite direction of the central target arrow) Tothis extent lower values of conflict effect indicate a higherperformance on cognitive conflict resolution

This finding becomes of particular interest when consid-ering the kind of trial associated to this reduction in the effectsize Incongruent trials require the resolution of a cognitiveconflict to be solved and they are known to be the slowestANT experimental conditions in terms of response speed[22]

Forlines et al [36] showed that tasks requiring the use oftwo hands (bimanual tasks as the ANT) are better performedwhen using touch-screen devices with respect to mouse-control PC To a similar extent Rogers et al [37] found thattouch-screen devices are particularly suitable for responsecollection when compared to another indirect responsedevice (namely a rotary encoder)

Given our results it is important to acknowledge thisproposed dichotomy between direct (the touch-screen inour case) and indirect response devices When using adirect input device the distance between the subject (hisherfingers) and the causal effect heshe carries on the envi-ronment modification (touching stimuli on the screen asrequired by the task) is reduced Touch-screen devices inthis framework lead a virtual environment to amore tangibleand ecological dimension One possible consequence of suchphenomenon could be an increase in self-commitment orin self-perceived efficacy towards the task and this couldlead to an enhancement by establishing a direct link betweenthe subject and the task reality In other words a differentperception of the self-commitment could be associated withresponses given with direct input devices shifting the taskenvironment perception into amore concrete entity onwhichthe subject acts as a physical agentThus critically the subjectinvolvement into the task could have been enhanced

Under this light one would expect to observe a greatereffect for those trials requiring a greater cognitive demand(ie incongruent trials) A greater involvement could trans-late into greater resources dedicated to task solution To thisextent cognitively simpler trials (ie congruent trials) wouldbenefit less since they do need less work to be solved on thecontrary trials requiring greater cognitive effort to be solvedsuch as incongruent trials would greatly benefit from suchresource availability

Although this scenario is suggestive some dedicatedexperimental investigation is needed to shed light on thecognitive basis of this behaviorally observed phenomenon

These evidences taken together with results on theease of their use highlighted by the survey indicate touch-screen devices as an ecological and suitable tool for thecomputerized administration of neuropsychological testsFurthermore other authors [35] showed that alternativeresponse input devices such as a light pen or touch screenare highly intuitive and have the advantage of bypassing thekeyboardThey demonstrate how these devices allow subjectsto focus their attention directly on the video display terminaland not have to shift their attention from the monitor to thekeyboard to locate a response key Nevertheless light pens

and touch screens also have their disadvantagesThey requirethe subject to hold his or her arm in an ldquouprdquo position andmove it along the screen This can produce fatigue and somevariation in reaction time

It should be noted that the computerized assessmentdoes not represent an alternative to the clinical settingHowever it can contribute in a significant manner to thetraditional evaluation Nevertheless there is a need to furtherdetail some aspects of our investigation (i) in order toincrease the inferential power and experimental validity ofour findings the tests will need to be administered to a largernumber of participants (ii) among these participants specificcognitively-impaired populations and physically-impairedpopulations (eg those subjects with motor function deficitfrom a brain injury that could affect the test performance)will need to be tested in order to assess if these instrumentscan be a valid and accessible tools in the clinical contextand (iii) most importantly a dedicated version of cognitivedomain-specific tests will need to be implemented and case-wise tested in order to detail whether and to which extentthey can be a valid alternative to more conventional pc-basedandor pencil and paper testing approaches

7 Conclusions and Future Perspectives

This work provides new data on the experimental feasibilityand clinical ecology of computerized neuropsychologicalassessment by addressing the impact of the implementationof different user interfaces and different stimuli presentationmodality

In order to set up an innovative computerized testingenvironment while keeping it feasible and ecological it isfundamental to detail how this conversion process impactsthe experimental and clinical neuropsychological settingsAlthough limited on approximately 40 healthy subjects andexperimented only on representative not exhaustive neu-ropsychological tests (on memory and attention functions)our evidences suggest that touch-screen devices can beconsidered for the computerized administration of neuropsy-chological tests

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Authorsrsquo Contribution

Matteo Canini Petronilla Battista and Pasquale AnthonyDella Rosa contributed equally to the paper

Acknowledgment

This work was supported by the fund for research of the Ital-ian Ministry of University and Research within a frameworkagreement between Lombardy Region andNational ResearchCouncil of Italy (no 17125 27092012)


References

[1] R E Schlegel and K Gilliland ldquoDevelopment and qualityassurance of computer-based assessment batteriesrdquo Archives ofClinical Neuropsychology vol 22 no 1 pp 49ndash61 2007

[2] S Zygouris andM Tsolaki ldquoComputerized cognitive testing forolder adults A reviewrdquo American Journal of Alzheimerrsquos Diseaseamp Other Dementias 2014

[3] M D Lezak Neuropsychological Assessment Oxford UniversityPress Oxford UK 2004

[4] O Spreen A Compendium of Neuropsychological Tests Admin-istration Norms and Commentary Oxford University Press1998

[5] M D Franzen and P A Arnett ldquoThe validity of neuropsycho-logical assessment proceduresrdquo in Biological and Neuropsycho-logical Mechanisms Life-Span Developmental Psychology H WResse and M D Franzen Eds pp 51ndash69 Erlbaum MahwahNJ USA 1997

[6] R W Heinrichs ldquoCurrent and emergent applications of neu-ropsychological assessment problems with validity and utilityrdquoProfessional Psychology Research and Practice vol 21 no 3 pp171ndash176 1990

[7] H C Fichman R Nitrini P Caramelli and K SameshimaldquoA new brief computerized cognitive screening battery (Com-pCogs) for early diagnosis of Alzheimerrsquos diseaserdquo Dementia eNeuropsychologia vol 2 no 1 pp 13ndash19 2008

[8] C BartelsMWegrzyn AWiedl V Ackermann andH Ehren-reich ldquoPractice effects in healthy adults a longitudinal study onfrequent repetitive cognitive testingrdquo BMCNeuroscience vol 11no 1 article 118 2010

[9] J Fredrickson P Maruff M Woodward et al ldquoEvaluation ofthe usability of a brief computerized cognitive screening test inolder people for epidemiological studiesrdquo Neuroepidemiologyvol 34 no 2 pp 65ndash75 2010

[10] E Woo ldquoComputerized neuropsychological assessmentsrdquo CNSSpectrums vol 13 no 10 supplement 16 pp 14ndash17 2008

[11] J-A Witt W Alpherts and C Helmstaedter ldquoComputerizedneuropsychological testing in epilepsy overview of availabletoolsrdquo Seizure vol 22 no 6 pp 416ndash423 2013

[12] R C K Chan D Shum T Toulopoulou and E Y H ChenldquoAssessment of executive functions review of instruments andidentification of critical issuesrdquo Archives of Clinical Neuropsy-chology vol 23 no 2 pp 201ndash216 2008

[13] American Psychological Association Committee on Profes-sional Standards American Psychological Association Boardof Scientific Affairs and Committee on Psychological Tests andAssessment Guidelines for Computer-Based Tests and Interpre-tations The Association 1986

[14] K Wild D Howieson F Webbe A Seelye and J Kaye ldquoStatusof computerized cognitive testing in aging a systematic reviewrdquoAlzheimerrsquos and Dementia vol 4 no 6 pp 428ndash437 2008

[15] C T Gualtieri and L G Johnson ldquoReliability and validity ofa computerized neurocognitive test battery CNS Vital SignsrdquoArchives of Clinical Neuropsychology vol 21 no 7 pp 623ndash6432006

[16] P Maruff E Thomas L Cysique et al ldquoValidity of theCogState brief battery relationship to standardized tests andsensitivity to cognitive impairment in mild traumatic braininjury schizophrenia and AIDS dementia complexrdquo Archivesof Clinical Neuropsychology vol 24 no 2 pp 165ndash178 2009

[17] T Dwolatzky V Whitehead G M Doniger et al ldquoValidityof a novel computerized cognitive battery for mild cognitiveimpairmentrdquo BMC Geriatrics vol 3 article 1 2003

[18] R H Paul J Lawrence L M Williams C C Richard NCooper and E Gordon ldquoPreliminary validity of ldquointegneurordquoa new computerized battery of neurocognitive testsrdquo Interna-tional Journal of Neuroscience vol 115 no 11 pp 1549ndash15672005

[19] M Inoue D Jimbo M Taniguchi and K Urakami ldquoTouchpanel-type dementia assessment scale a new computer-basedrating scale for Alzheimerrsquos diseaserdquo Psychogeriatrics vol 11 no1 pp 28ndash33 2011

[20] K Onoda T Hamano Y Nabika et al ldquoValidation of anew mass screening tool for cognitive impairment cognitiveassessment for Dementia iPad versionrdquo Clinical Interventionsin Aging vol 8 pp 353ndash360 2013

[21] A Blackwell ldquoTag Archives CANTAB mobilerdquo 2011[22] J Fan B D McCandliss J Fossella J I Flombaum and M I

Posner ldquoThe activation of attentional networksrdquo NeuroImagevol 26 no 2 pp 471ndash479 2005

[23] P A Della Rosa G Videsott V M Borsa et al ldquoA neuralinteractive location for multilingual talentrdquo Cortex vol 49 no2 pp 605ndash608 2013

[24] A Rey LrsquoExamenclinique en Psychologie Presses Universitairesde France Paris France 1964

[25] C Caltagirone G Gainotti C Masullo and GMiceli ldquoValidityof some neuropsychological tests in the assessment of mentaldeteriorationrdquo Acta Psychiatrica Scandinavica vol 60 no 1 pp50ndash56 1979

[26] A Laudanna A M Thornton G Brown C Burani andL Marconi ldquoUn corpus dellrsquoitaliano scritto contemporaneodalla parte del riceventerdquo in III Giornate Internazionali diAnalisi Statistica dei Dati Testuali S Bolasco L Lebart and ASalem Eds vol 1 pp 103ndash109 Centro Informazione StampaUniversitaria Roma Italy 1995

[27] P A Della Rosa E Catricala G Vigliocco and S F CappaldquoBeyond the abstract-concrete dichotomy mode of acquisitionconcreteness imageability familiarity age of acquisition con-text availability and abstractness norms for a set of 417 ItalianwordsrdquoBehavior ResearchMethods vol 42 no 4 pp 1042ndash10482010

[28] A J Larner Neuropsychological Neurology The NeurocognitiveImpairments of Neurological Disorders Cambridge UniversityPress Cambridge UK 2013

[29] J WMacLeod M A Lawrence MMMcConnell G A EskesRMKlein andD I Shore ldquoAppraising theANT psychometricand theoretical considerations of the attention network testrdquoNeuropsychology vol 24 no 5 pp 637ndash651 2010

[30] A Sears and B Shneiderman ldquoHigh precision touchscreensdesign strategies and comparisons with a mouserdquo InternationalJournal ofMan-Machine Studies vol 34 no 4 pp 593ndash613 1991

[31] A Holzinger ldquoFinger instead of mouse touch screens asa means of enhancing universal accessrdquo in Universal AccessTheoretical Perspectives Practice and Experience pp 387ndash397Springer Berlin Germany 2003

[32] PThekkumpurath C VenkateswaranM Kumar A Newshamand M I Bennett ldquocreening for psychological distress inpalliative care performance of touch screen questionnairescompared with semistructured psychiatric interviewrdquo Journalof Pain and Symptom Management vol 38 no 4 pp 597ndash6052009


[33] H C Fichman R Nitrini P Caramelli and K SameshimaldquoA new Brief computerized cognitive screening battery (Com-pCogs) for early diagnosis of Alzheimers diseaserdquo Dementia ampNeuropsychologia vol 2 no 1 pp 13ndash19 2008

[34] G Radvansky Human Memory Pearson Education BostonMass USA 2006

[35] E Wood T Willoughby A Rushing L Bechtel and J GilbertldquoUse of computer input devices by older adultsrdquo Journal ofApplied Gerontology vol 24 pp 419ndash438 2005

[36] C Forlines D Wigdor C Shen and R Balakrishnan ldquoDirect-touch vs mouse input for tabletop displaysrdquo in Proceedings ofthe 25th SIGCHI Conference on Human Factors in ComputingSystems (CHI rsquo07) pp 647ndash656 ACM May 2007

[37] W A Rogers A D Fisk A C McLaughlin and R Pak ldquoToucha screen or turn a knob choosing the best device for the jobrdquoThe Journal of the Human Factors and Ergonomics Society vol47 no 2 pp 271ndash288 2005

Submit your manuscripts athttpwwwhindawicom

Stem CellsInternational

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014


MEDIATORSINFLAMMATION

of


Behavioural Neurology

EndocrinologyInternational Journal of



Disease Markers


BioMed Research International

OncologyJournal of



Oxidative Medicine and Cellular Longevity


PPAR Research

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Immunology ResearchHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

ObesityJournal of



Computational and Mathematical Methods in Medicine

OphthalmologyJournal of


Diabetes ResearchJournal of



Research and TreatmentAIDS


Gastroenterology Research and Practice


Parkinsonrsquos Disease

Evidence-Based Complementary and Alternative Medicine

Volume 2014Hindawi Publishing Corporationhttpwwwhindawicom


apply to the test itself but to the inferences that are drawnfrom the test [5 6] Most importantly clinical sensitivityallowing neuropsychologist to capture potential shades ina specific domain and to trace a specific cognitive profilemay results difficult to be translated in a computer-basedassessment Despite these numerous advantages traditionalpaper and pencil tools show some limitations particularlywhen assessing cognitive changes in a relatively short follow-up period The most commonly administered tests usuallydo not provide alternative forms of administration [7] thusprecluding to repeat testing over short intervals (eg lt6months) [8] Other specific limitations concern the intrinsicnature of the tools and include assisting the setting andmanually computing scoring by the experimenter [9] longduration of the assessment potential bias related to differentexaminers [10] impossibility to provide a precise time controlon stimuli presentation andor the lack of an accuratemeasurement of motor response accuracy [11] and greatercosts Although there are specific tests for the assessmentof attentional and executive functions which can evaluateindividual components and can diagnose specific deficits [12]some executiveattention abilities could take advantage fromthe use of a computerized tool in particular in the assessmentof response inhibition resistance to distraction planningproblem solving working memory operations and mentalset shifting divided attention

All of these limitations could be overcome by a comput-erized assessment on condition that efficacies and ecolog-ical measurements are carried out About these issues theAmerican Psychological Association (APA) has recognizedthe importance of computerized psychological testing andhas suggested how to implement and interpret computerizedtest results in its guidelines [13] Furthermore computerizedassessment of cognitive functions can be self-administeredand can have a shorter duration (eg by reducing ldquodeadrdquotimes in stimuli presentation) They may have great valid-ity and reliability due to their great objectivity precisionand standardization Computerized performance can alsominimize the so called ldquofloor and ceiling effectsrdquo occurringwhen differences among participant performance are notfully captured thus they can provide more standardizedmeasures of subject performance crucial for example for anaccurate and early detection of specific pathological disease(eg dementia) [14]

It appears clearly that computerized testing will representan essential part of the clinical setting in the nearest futureabove all in screening procedures on condition that thesenew instruments and their results are governed by experts

Feasible efficacious and ecological computerized test-ing could allow clear pictures of normal cognition to bemeasured and monitored also at home pinpointing specificdeficits in each cognitive domain in aging people Validatedcomputerized tools could also provide stronger groundingto overcome the lack of a consensus regarding the feasi-bility and testing efficacy related to the different types oftechnological solutions and settings andor response layoutschosen for the assessment For instance a more ecologicaltechnological solution can overcome one important obstacleto the wide use of computerized assessment attributable to

the familiarity with technological devices [11] while a moreecological setting in the nearest future could partially allowthe administration of the tests without the support of aspecialized clinician or the tests could be potentially self-administered or assessed by a caregiver at home

The aims of this study were to evaluate the following

(1) the testing efficacy of a computerized test (in arepresentative case of executive function assessment)when response to stimuli is measured using a touchscreen compared to a conventional mouse-controlresponse device

(2) the testing efficacy (the efficacy of the neuropsycho-logical test) of a computerized test (in a representativecase of memory function assessment) with respect totwo different input modalities a visual presentationmodality of the test replicating the most diffuseddigital versions of the test and a verbal presentationmodality replicating the classical clinical administra-tion of the test

(3) the ecology of both computerized assessment modal-ities (touch-screen tablet and mouse-control PC)including preference measurements of participants

2 Related Works

Computerized cognitive batteries are already available andused for the screening and the assessment of dementiaSome of these instruments were appropriately created for theassessment of cognitive decline in dementia other ones wereadapted to fulfill this role in aging Some tools are designedfor research use and others tools have been designed mainlyfor clinical use some of them already implemented in clinicalguidelines

An interesting review [2] reported seventeen comput-erized test batteries used in the measurement of cognitiveabilities of adults Some of these tools are able to run only ona PClaptop and others are available only on web sites

Among these tools

(1) CNS vital signs [15] battery is developed as a routineclinical screening instrument It includes seven testsverbal and visual memory finger tapping symboldigit coding Stroop test test of shifting attention andcontinuous performance test

(2) CogState [16] battery was developed as a dementiascreening instrument and it is implemented as a cardgame formThe participant plays different games thatare adapted accordingly to performance CogStaterequires an active internet connection to generatea report The participantrsquos data are uploaded andanalyzed Then a report is generated and e-mailedback to the provider

(3) NeuroTrax [17] includes custom software on thelocal testing computer and serves as a platform forinteractive cognitive tests that provide precise accu-racy and reaction time data The level of difficultyis graded The NeuroTrax tests different cognitive






















20M18 F





Agemean 64474

0970 0831st dev 8462range 53 87

Educationmean 11263

0487 0970st dev 4131range 5 19

MMSEmean 2891

0344 0272st dev 067range 28 30















+ +

+

+

lowast

lowast

lowast

lowast

lowast

(a)

400ms

100ms

1700msuntilresponse

+

+

+

lowast

(b)



































4 Data Analysis












5 Results











200000

150000

100000

50000

000

Mea

n RT

s (m

s)


Alerting effect

Mouse Touch















6 Discussion






































Acknowledgment



References












































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of





































20M18 F





Agemean 64474

0970 0831st dev 8462range 53 87

Educationmean 11263

0487 0970st dev 4131range 5 19

MMSEmean 2891

0344 0272st dev 067range 28 30















+ +

+

+

lowast

lowast

lowast

lowast

lowast

(a)

400ms

100ms

1700msuntilresponse

+

+

+

lowast

(b)



































4 Data Analysis












5 Results











200000

150000

100000

50000

000

Mea

n RT

s (m

s)


Alerting effect

Mouse Touch















6 Discussion






































Acknowledgment



References












































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of



















+ +

+

+

lowast

lowast

lowast

lowast

lowast

(a)

400ms

100ms

1700msuntilresponse

+

+

+

lowast

(b)



































4 Data Analysis












5 Results











200000

150000

100000

50000

000

Mea

n RT

s (m

s)


Alerting effect

Mouse Touch















6 Discussion






































Acknowledgment



References












































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of





































4 Data Analysis












5 Results











200000

150000

100000

50000

000

Mea

n RT

s (m

s)


Alerting effect

Mouse Touch















6 Discussion






































Acknowledgment



References












































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of


















5 Results











200000

150000

100000

50000

000

Mea

n RT

s (m

s)


Alerting effect

Mouse Touch















6 Discussion






































Acknowledgment



References












































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of



























6 Discussion






































Acknowledgment



References












































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of














































Acknowledgment



References












































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of

















References












































of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of



























of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of
















Research Article Computerized Neuropsychological Assessment in...

Documents

Transcript of Research Article Computerized Neuropsychological Assessment in...