Full Terms & Conditions of access and use can be found athttps://www.tandfonline.com/action/journalInformation?journalCode=ncen20

Journal of Clinical and Experimental Neuropsychology

ISSN: 1380-3395 (Print) 1744-411X (Online) Journal homepage: https://www.tandfonline.com/loi/ncen20

Anxious, but not depressive, symptoms areassociated with poorer prospective memoryperformance in healthy college students:Preliminary evidence using the tripartite model ofanxiety and depression

Marissa Ann Bowman, Tony J. Cunningham, Holly F. Levin-Aspenson, AndreaE. O’Rear, Joseph R. Pauszek, Stephanie Ellickson-Larew, Brandy S. Martinez& Jessica D. Payne

To cite this article: Marissa Ann Bowman, Tony J. Cunningham, Holly F. Levin-Aspenson, AndreaE. O’Rear, Joseph R. Pauszek, Stephanie Ellickson-Larew, Brandy S. Martinez & Jessica D. Payne(2019) Anxious, but not depressive, symptoms are associated with poorer prospective memoryperformance in healthy college students: Preliminary evidence using the tripartite model of anxietyand depression, Journal of Clinical and Experimental Neuropsychology, 41:7, 694-703, DOI:10.1080/13803395.2019.1611741

To link to this article: https://doi.org/10.1080/13803395.2019.1611741

Published online: 14 May 2019. Submit your article to this journal

Article views: 396 View related articles

View Crossmark data Citing articles: 1 View citing articles

Anxious, but not depressive, symptoms are associated with poorer prospectivememory performance in healthy college students: Preliminary evidence usingthe tripartite model of anxiety and depressionMarissa Ann Bowmana, Tony J. Cunninghamb,c, Holly F. Levin-Aspensonc, Andrea E. O’Rearc, Joseph R. Pauszekc,Stephanie Ellickson-Larewc,d, Brandy S. Martinezc and Jessica D. Paynec

aDepartment of Psychology, University of Pittsburgh, Pittsburgh, PA, USA; bDivision of Sleep Medicine, Harvard Medical School, Cambridge,MA, USA; cDepartment of Psychology, University of Notre Dame, Notre Dame, IN, USA; dMassachusetts Veterans Epidemiology Research andInformation Center, Veterans Affairs Boston Healthcare System, Boston, MA, USA

ABSTRACTProspective memory (PM) – or memory for tasks to be completed in the future – is essential fordaily functioning. Although depression and anxiety have been shown to impair PM performance,few studies have explored the relative contributions of different symptom domains. Here, weexamined the relation between anxiety, depression, negative mood, and PM performance usingthe tripartite model. The tripartite model attributes the substantial overlap between anxiety anddepression to general distress/negative affect. Twenty-seven non-diagnosed undergraduate par-ticipants first completed self-report measures of depression (Beck Depression Inventory-II), anxi-ety (Beck Anxiety Inventory [BAI], State Trait Anxiety Inventory [STAI]), and affect (Positive andNegative Affect Schedule). They were then given an event-based PM instruction to be completedduring three ongoing cognitive tasks. Depressive symptoms and positive affect were unrelated toPM performance. Higher anxiety symptoms (BAI, r = −0.62; STAI, r = −0.41) and negative affect(r = −0.45) were associated with poorer PM performance, with anxiety doubling the varianceexplained over-and-above negative affect (ΔR2 = 0.20). These preliminary results suggest thatanxiety symptoms may be uniquely related to impairments in PM function, and highlight theneed for future studies to consider the individual contributions of symptoms to understandchanges in cognition and behavior.

ARTICLE HISTORYReceived 12 October 2018Accepted 19 April 2019

KEYWORDSProspective memory;anxiety; depression;tripartite model

Introduction

The ability to remember to do something at the appropriatetime in the future, or prospective memory (PM), is criticalfor daily living (McDaniel & Einstein, 2007). Difficulty inPMcan have consequences for employment (e.g.,missing adeadline) and health (e.g., failing to take medication).Individuals with anxiety and depression often experiencethese problems of daily living, with meta-analytic evidencedemonstrating increased odds of absenteeism (Bhui,Dinos, Stansfeld, & White, 2012), and medication non-adherence (DiMatteo, Lepper, & Croghan, 2000). A small,but converging, literature has shown that individuals withanxiety and depression have impairments in PM, whichmaybe one possible pathway for these results. Developing abetter understanding of the relationship between anxiety,depression, and PM is thus important for improving bothcognitive and functional outcomes.

Anxiety and depression have been individually linkedto poorer PM performance in laboratory studies. For thereader’s convenience, we have summarized all extant

studies examining the association between depression,anxiety, and/or mood and PM performance in Table 1.These laboratory studies instruct participants to responddifferently at a certain occasion (event-based task) or at apredetermined time point (time-based task). Prospectivememory tasks may also vary as a function of the focalityof a cue – focal task cues are processed during theongoing task (e.g., the ongoing task is to count the num-ber of vowels in each word, and the focal task is torespond differently when three letter “E’s” appear),whereas non-focal task cues are not (e.g., in the sameongoing task, the non-focal task is to respond differentlywhen the word is a verb; Altgassen, Kliegel, & Martin,2009). One study showed that high state anxiety (howanxious a person feels now), but not trait anxiety (howanxious a person generally feels), was associated withpoorer performance on a focal PM task (Harris &Cumming, 2003). In another study, both state and traitanxiety were negatively associated with subjective reportsof PM ability (e.g., “Do you forget appointments if youare not prompted by someone else or by a reminder such

CONTACT Jessica D. Payne Jessica.Payne.70@nd.edu University of Notre Dame, Haggar Hall, Room 122-B, Notre Dame, IN 46556, USA

JOURNAL OF CLINICAL AND EXPERIMENTAL NEUROPSYCHOLOGY2019, VOL. 41, NO. 7, 694–703https://doi.org/10.1080/13803395.2019.1611741

© 2019 Informa UK Limited, trading as Taylor & Francis Group

Table1.

Prospectivemem

orystud

ieswith

measuresof

depression

,anxiety,and

/ormood.

Stud

yClinicalsymptom

(s)

Measure

ofsymptom

Ong

oing

task

Prospectivemem

orytask

Results

(separated

bytask

type)

Altgassen,

Kliegel,&

Martin

,2009

Depression;

adults

with

orwith

outdiagno

sed

depression

Diagn

osisof

major

depressive

disorder

from

Structured

Clinical

Interview

(SCID)

Coun

tthenu

mberof

vowelsin

word-pairs

todeterm

ine

which

wordhasmore

vowels

Event-based,

focaltask:Indicate

whenon

eof

thewords

hasthreee’s

Event-based,

non-focaltask:Indicate

when

oneof

thewords

isaverb

Focaltask:no

sign

ificant

difference

betweendepressedand

non-depressed

Non-focaltask:D

epressed

individu

alsperformed

worse

than

non-depressed

Altgassenet

al.,2011

Depression;

adults

with

orwith

outdiagno

sed

depression

Diagn

osisof

major

depressive

disorder

from

SCID;B

eck

DepressionInventory

(BDI)

Determinewhich

oftwowords

belong

edto

acatego

ry(sem

antic

catego

rization

task)

Event-based,

focaltask:Indicate

ifon

eof

the

words

was

onacued

listpresentedat

beginn

ingof

stud

y

Event-basedtask:D

epressed

individu

alsperformed

worse

than

non-depressed,

buton

lyforpo

sitively-valenced

stimuli;depressedindividu

alsneeded

morerepetitions

ofthecued

wordlistat

encoding

Arno

ldet

al.,

2015

Depression,

anxiety;no

n-clinicalun

dergradu

ates

BDI-II,HospitalA

nxiety

andDepressionScale

(HAD

S),State

Trait

AnxietyInventory(STA

I)

Identifywhether

acolored

wordmatches

thecolorof

previouslypresented

rectangles

Event-based,

non-focaltask:Indicate

ifon

eof

thecoloredwords

was

onacued

list

presentedat

thebeginn

ingof

thestud

y

Prospectivecomponent:Ind

ividualswith

high

stateanxiety

performed

worse

than

thosewith

low

stateanxiety;no

sign

ificant

difference

fordepressedcomparedto

non-

depressed

Retrospectivecomponent:n

osign

ificant

difference

for

anxietyanddepression

Cuttler&

Graf,2008

Depression,

anxiety,

checking

compu

lsions;

non-clinical

undergradu

ates

BDI,STAI

Determineifaletter

strin

gisa

wordor

non-word(lexical

decision

-makingtask)

Event-based,

non-focaltask:Askthe

experim

enterforthereturn

ofaperson

albelong

ingat

theendof

thetask

Subjectivereportof

PM:Self-ratin

gson

aqu

estio

nnaire

ofPM

ability

Activity-based

task:N

osign

ificant

associationwith

depression

,state,traitanxiety

Subjectivereportof

PM:D

epression,

state,andtraitanxiety

wereindividu

allyassociated

(not

includ

edin

onemod

el).

Harris

&Cu

mming,

2003

Anxiety;no

n-clinical

undergradu

ates

STAI

Aswords

areread

alou

d,generate

awordthat

pairs

them

aticallywith

each

word

(word-pairassociates)

Event-based,

focaltask:Indicate

iftheword

was

presentedon

acued

listpresentedat

beginn

ingof

stud

yby

writingtheword

insteadof

apairedword

Prospectivecomponent:Ind

ividualswith

high

stateanxiety

performed

worse

than

thosewith

low

stateanxiety;no

sign

ificant

associationwith

traitanxiety

Retrospectivecomponent:N

osign

ificant

associationforstate

ortraitanxiety

Harris

&Menzies,

1999

Depression,

anxiety,stress;

non-clinical

undergradu

ates

Depression,

Anxietyand

Stress

Scales

(DAS

S)Generateword-pairassociates

Event-based,

focaltask:Indicate

bywritinga

symbo

lifwordisapieceof

clothing

ora

partof

thehu

man

body

Anxietywas

asign

ificant

predictorof

performance,after

adjustmentfordepression

andstress

Kliegele

tal.,

2005

Indu

cednegativeor

neutral

mood;

non-clinical

undergradu

ates

Six-minutelong

video

clips

Indicate

ifanimalwordwas

the

sameas

that

presentedtwo

stimulip

reviou

sly(n-back)

Time-based:

Pressatarget

keyeveryon

eminute

Firsthalfof

thetask:Ind

ividualsin

anegativemood

performed

worse

than

thosein

apo

sitivemood

Second

halfof

task:N

osign

ificant

difference

asafunctio

nof

mood

Kliegel&

Jäger,2006

Depression,

anxiety;no

n-clinicalun

dergradu

ates

andhealthycommun

ity-

dwellingvolunteers

HAD

SIndicate

ifapicturematches

theon

eshow

ntwostimuli

previously(n-back)

Event-based,

focaltask:Indicate

ifthepicture

isan

animal

Time-basedtask:P

ress

atarget

keyeverytwo

minutes

Event-basedtask:N

osign

ificant

difference

asafunctio

nof

depression

oranxiety

Time-basedtask:H

igherdepressive

symptom

sperformed

worse,adjustedforanxiety

Anxietywas

notsign

ificantlyassociated

Liet

al.

(2013)

Depression;

undergradu

ates

with

high

orlow

depressive

symptom

sper

BDI-II

BDI-II,DAS

SMem

oryforIntentions

ScreeningTest(M

IST)

Event-based,focaltasks:fou

rtasks,halfwith

a2-min

delayintervalandtheotherhalf

with

a15-m

indelay

Time-basedtasks:four

tasks,halfwith

a2-min

delayintervalandtheotherhalfwith

a15-

min

delay

Event-basedtasks:no

sign

ificant

difference

asafunctio

nof

depressive

symptom

sTime-basedtasks:individu

alswith

high

depressive

symptom

sperformed

worse

Delays:Highdepressive

symptom

sassociated

with

worse

performance

on15-m

indelay,bu

tno

t2-min

delay

Livner

etal.,

2008

Depression;

subsam

pleof

popu

latio

n-basedstud

yof

olderadults

Comprehensive

Psycho

patholog

ical

Ratin

gScale

Mem

orizeaseriesof

orally

presentedwords

Event-based,

non-focaltask:Remindthe

experim

enterto

makeaph

onecallat

the

endof

thestud

y

Event-basedtask:D

epressivesymptom

simpairedthe

retrospectivecompo

nent

ofthetask,b

utno

tthe

prospectivecompo

nent

Rude

etal.,

1999

Depression;

adults

with

orwith

outdiagno

sed

depression

SCID,B

DI

General

know

ledg

etest

Time-basedtask:p

ress

atarget

keyeveryfive

minutes

tocheckthetim

eTime-basedtask:D

epressed

individu

alsperformed

worse

than

non-depressed

Schn

itzspahn

etal.,2014

Indu

cednegative,neutral

orpo

sitivemood;

healthycommun

ity-

dwellingadultsaged

18–

84

Six-minutelong

video

clips

Indicate

ifanimalwordwas

the

sameas

that

presentedtwo

stimulip

reviou

sly(n-back)

Time-basedtask:P

ress

atarget

keyeveryon

eminute

Time-basedtask:Ind

ividualswith

indu

ctionof

negativeand

positivemoodperformed

worse

than

neutralm

oodin

youn

gadults(18–25)

Nosign

ificant

difference

formoodin

olderadults

(59–84

yearsold)

JOURNAL OF CLINICAL AND EXPERIMENTAL NEUROPSYCHOLOGY 695

as a calendar or a diary?”, Cuttler & Graf, 2008).Depressive symptoms have also been related to impair-ments in PM performance, on time-based tasks (Kliegel& Jäger, 2006; Li, Weinborn, Loft, & Maybery, 2013;Rude, Hertel, Jarrold, Covich, & Hedlund, 1999), non-focal tasks (Altgassen et al., 2009), and focal tasks whenstimuli were positively valenced (Altgassen, Henry,Bürgler, & Kliegel, 2011). A limitation of all of thesestudies, however, is a failure to consider the other symp-tom domain. Yet this is critical, as depression and anxietyare highly co-morbid (Hirschfield, 2001), but aredissociable.

To our knowledge, only three studies have includedboth anxiety and depression symptoms in one regressionmodel with the outcome of PM performance. In evaluat-ing performance on time-based tasks, higher depressivesymptoms, adjusted for trait anxiety symptoms, wereassociated with worse performance (Kliegel & Jäger,2006). For focal PM task performance, one study reportedthat trait anxiety, but not stress and depression in thesame regression model, was associated with worse per-formance on a task that involved writing down pairedassociates to words read aloud (ongoing task) andresponding differently to words that belonged to thecategory of clothes or body parts (Harris & Menzies,1999). In contrast, another study of a focal PM perfor-mance reported that neither depressive nor trait anxietysymptoms were associated with performance on a taskthat required a different response from the ongoing n-back task when an animal picture appeared (Kliegel &Jäger, 2006). In a study of non-focal PM performance,high state anxiety was associated with poorer perfor-mance on a non-focal task that asked participants toidentify whether a colored word matched the color ofpreviously presented rectangles and respond differently ifone of the words matched a list of cued words (Arnold,Bayen, & Böhm, 2015).

In general, the results of the few existing studiesincluding both depression and anxiety symptoms inregression models suggest that anxiety affects the abilityto store or correctly recall the PM instruction during anongoing task, while depression affects the ability to dis-engage from the ongoing task to perform the PM task.Anxiety, but not depression, was associated with focal(Harris & Menzies, 1999) and non-focal (Arnold et al.,2015) event-based PM performance. Moreover, depres-sion, but not anxiety, was associated with impaired time-based PM performance (Kliegel & Jäger, 2006). All threeof these studies assessed nonclinical undergraduate parti-cipants, although Kliegel and Jäger (2006) also includedhealthy community-dwelling participants. Thus, thesestudies suggest that anxiety affects the ability to storeand recall information, while depression affects the ability

to disengage from the ongoing task. These studies havetaken important first steps to disambiguate the overlapbetween depressive and anxious symptoms in PM per-formance, which is critical given evidence that depressionand anxiety are highly comorbid (Hirschfield, 2001).Limiting these studies, however, is a lack of considerationof affective state.

In clinical psychology research, Clark and Watson(1991) have attributed the strong correlation betweenanxiety and depression to general distress (i.e., negativeaffect). In their model, anxiety is characterized by highlevels of physiological hyperarousal and high levels ofnegative affect, whereas depression is characterized bylow levels of positive affect and high levels of negativeaffect (Watson et al., 1995). This theory gained furthercredence after an exploratory factor analysis (Watsonet al., 1995) and a subsequent confirmatory factoranalysis (Joiner, 1996) demonstrated psychometric sup-port for the tripartite model in nonclinical collegesamples. These findings suggest that the inclusion ofan affective state measure is informative when attempt-ing to tease apart the theoretical distinction betweendepression and anxiety, even within otherwise healthy,nonclinical college cohorts.

Previous clinical PM studies of anxiety and depres-sion have not accounted for negative affect in theirmodels. Yet, assessing affect is important for PM stu-dies. An individual’s emotional state can determine theamount of attentional resources that can be devoted toa task at any given time (Ellis & Ashbrook, 1988). Insupport of this, experiments involving negative moodinduction have been shown to impair performance onboth a PM task (e.g., a time-based PM task,Schnitzspahn et al., 2014) and an ongoing task (e.g.,an n-back working memory task; Kliegel et al., 2005)compared to individuals induced into a positive orneutral mood. Therefore, it is plausible that naturallyoccurring affective states may also impact PM perfor-mance. Moreover, including a measure of negativeaffect in PM studies of anxiety and depression may behelping in parsing apart the highly overlapping con-structs anxiety and depression.

In the present study, we examined the contributions ofanxious and depressive symptoms on PM performanceusing a focal, event-based task. Additionally, we assessedhow these symptoms affect PMbeyondwhat they share interms of general distress/negative affect using one cohe-sivemodel – the tripartite model (Clark &Watson, 1991).It is important to examine measures of affect and psycho-pathology together if we are to disentangle their uniqueand shared effects. This approach also allows for compar-isons of the relative contributions of different areas ofaffect and psychopathology to prospective memory

696 M. A. BOWMAN ET AL.

performance, as well as the incremental predictive powerof specific phenomena (e.g., anxiety) over and abovemore general domains (e.g., negative affect).We reasonedthat utilizing this model in a pilot study of clinical influ-ences on prospective memory performance might help usto begin to clarify whether general distress or uniquecomponents of depression or anxiety drive poor PMperformance and provide a more nuanced understandingof the possible associations between clinical symptomsand diminished performance on future-oriented tasks.

Based on the research outlined above, we hypothe-sized that this study would provide preliminary evi-dence that higher levels of (1) anxiety and (2)negative affect would negatively correlate with PMperformance. In line with previous research suggest-ing that depression impairs time-based PM perfor-mance and not focal, event-based PM performance,we did not expect (3) depression to be associated withPM impairment here. Importantly, to further probehypotheses 1–3, we used the tripartite model to pro-vide preliminary evidence for a framework distin-guishing between the relative contributions ofanxiety, depression, and negative affect. Given itsnovelty in PM tasks specifically and memory researchmore generally, we used the tripartite model in anexploratory manner to evaluate whether symptomsunique to anxiety or depression were associated withPM performance. If the tripartite model aligns withprevious findings in the literature concerning anxiety,depression, and PM performance, it would suggestthat the symptoms unique to anxiety, but not depres-sion, would be associated with PM performance.

Method

Participants

Twenty-seven University of Notre Dame undergraduatestudents (19 female; ranging from 18 to 30 years of age;mean age of 19.9) participated as part of a larger ongoingstudy assessing sleep and memory. Participants werecompensated with cash payment or class credit. TheUniversity of Notre Dame Institutional Review Board(IRB) approved all testing procedures, and written con-sent was obtained before the experiment. All participantswere instructed to refrain from using tobacco, caffeine,alcohol, and recreational drugs for 24 hours prior toparticipation in the study. They were all fluent Englishspeakers and had normal or corrected-to-normal vision.Participants who reported substance abuse, a previouslydiagnosed major mental illness, prior sleep disorders, orthe use of medications affecting the central nervous sys-tem were excluded from the study. Given that

participants with psychological diagnoses were ineligibleto participate in the larger ongoing study, our results hereexamined clinical symptoms in an otherwise healthypopulation.

Procedure

Participants were randomly assigned to either a morn-ing (n = 15) or evening (n = 12) condition to controlfor time of day effects on performance. The morninggroup arrived at the laboratory at 9:00 AM, while theevening group arrived at the laboratory at 9:00 PM.Although performance differences as a function of timeof day have never been observed in our laboratory, wenevertheless include these conditions because success-ful memory retrieval has been shown to change as afunction of time of day in some studies (e.g., Folkard &Monk, 1980). All other tasks and instructions wereidentical between conditions. Participants filled out aseries of self-report questionnaires and engaged inthree cognitive tasks (living/nonliving decision, lexicaldecision, and semantic categorization; Scullin &McDaniel, 2010) on a desktop computer with a 17-inch monitor running E-Prime software (Schneider,Eschman, & Zuccolotto, 2002). The three ongoingtasks occurred in the same order for all participants.The PM instruction was an event-based focal task.

In the living/nonliving decision task, participantswere presented with an object word and were askedto determine whether the object was living (e.g., “cat”)or non-living (e.g., “couch”), as quickly and accuratelyas possible. In the lexical decision task, participantswere asked to determine whether a string of lettersformed a word or a non-word. In the semantic cate-gorization task, participants were presented with anitem (e.g., “pizza”) and were asked to determine if itmatched the category presented next to it (e.g., “food”).There were 164 trials of each task. After completion ofall three of these tasks, participants were given the PMinstruction. Specifically, participants were instructed topress the “Q“ key on the keyboard whenever the words“table” or ”horse” appeared in any of the three tasks inthe next session (Scullin & McDaniel, 2010). Afterreceiving these instructions, all participants completeda 20-minute distractor task involving a second set ofself-report and demographic questionnaires.

Participants then engaged in a second round of allthree ongoing cognitive tasks, which included a test ofPM for the critical words. Each of the three tasksshowed the words “table” and “horse” only once.Therefore, participants were given opportunities toremember the PM task instructions and press the “Q”key on the keyboard following the presentation of these

JOURNAL OF CLINICAL AND EXPERIMENTAL NEUROPSYCHOLOGY 697

words a maximum of six times. Once again, there were164 trials of each ongoing task during the second ses-sion. PM performance was measured for each partici-pant as the proportion of successful attempts recallingand completing the instruction to press the letter “Q”in response to the words “horse” or “table” over thetotal number of times these words were presentedduring the tasks.

Measures

The Beck Depression Inventory (BDI-II; Beck, Steer, &Brown, 1996) is a 21-item self-report measure of depres-sion in adolescents and adults. Respondents rate theirexperience of each symptom (e.g., “I don’t have enoughenergy to do anything”) over the past two weeks using afour-point intensity scale (from 0–3, with unique descrip-tors for each question). Scores below 13 are interpreted asindicating minimal depression, 14–19 as indicating milddepression, 20–28 as indicating moderate clinical depres-sion, and above 29 as indicating severe clinical depression(Beck et al., 1996).

The Beck Anxiety Inventory (BAI; Beck, Epstein,Brown, & Steer, 1988) is a 21-item self-report measureof anxiety that focuses on somatic content to minimizethe overlap with depression. Respondents rate theirexperience of each symptom (e.g., “fear of the worsthappening”) over the past week using a four-pointintensity scale (0 = “it did not bother me at all,” 3 =“I almost could not stand it”). Scores between 8 and 15are interpreted as indicating mild anxiety, 16–25 asindicating moderate anxiety, and 25 and above as indi-cating severe anxiety. The BAI was included in thestudy as a measure of trait anxiety, or how anxiousrespondents tend to feel in general, given evidence forhigh test–retest reliability across time (Beck et al., 1988;but see Brown et al., 1997, for evidence that BAI scorescan change with intervention).

The State-Trait Anxiety Inventory (STAI; Spielberger,Gorsuch, Lushene, Vagg, & Jacobs, 1983) assesses howanxious respondents tend to feel in general (i.e., trait anxi-ety) and how anxious they are feeling currently (i.e., stateanxiety). This study used only the state anxiety subscale, asthat was the only subscale administered. This subscaleconsists of 20 items (e.g., “I feel over-excited and rattled”).Respondents rate their current experience of each symp-tom using a four-point intensity scale (1 = “not at all,” 4 =“very much so”), and higher scores reported on the STAIindicate greater state anxiety.

The Positive and Negative Affect Schedule (PANAS;Watson, Clark, & Tellegen, 1988) is a 20-item self-report affect measure comprising two scales: PositiveAffect (PA) and Negative Affect (NA). The NA scale

contains 10 negative affect terms (e.g., “upset”), whilethe PA scale contains 10 positive affect terms (e.g.,“enthusiastic”). This study used present-moment timeinstructions for the PANAS, such that respondentsrated the extent to which they were experiencing eachaffect term “right now” using a five-point scale (1 =“very slightly or not at all,” 5 = “extremely”). The PAand NA scales both show strong internal consistencyunder present-moment time instructions and havebeen shown to be independent of each other (Watsonet al., 1988).

Statistical analysis

We first examined whether memory performance dif-fered between morning versus evening conditionsusing a one-way analysis of variance (ANOVA), andif memory performance differed as a function of theongoing task in paired sample t-tests. Second, weexamined the bivariate correlations among the self-report measures of affect, anxiety, and depression todetermine whether these correlations were consistentwith prior research and theory. In subsequent correla-tions, we evaluated the association between these clin-ical constructs and PM performance. Finally, to test thetripartite model, we conducted hierarchical linearregression analyses to examine whether BDI-II, BAI,or STAI scores were associated with differences in PMperformance over and above PANAS negative affect. Inthese models, we first included PANAS negative affectas an independent variable, and in a second step, weincluded the BAI, the STAI, or the BDI-II. These hier-archical regression models allow us to evaluate theunique role of anxiety or depression on prospectivememory performance, after partialling out the varianceexplained by negative affect. We did not include posi-tive affect in our models, as positive affect is not theo-rized to explain the overlap between anxious anddepressive symptoms (Clark & Watson, 1991). In sen-sitivity analyses, we examined whether these associa-tions differed as a function of the ongoing task.

Results

There were no differences in memory performancebetween the morning and evening conditions, F(1, 25) =0.69, p > .40. Moreover, there were no differences in anyof the self-report measures between the morning andevening conditions (all ps > .20). Thus, we collapsedacross participants for all subsequent analyses. In pairedsample t-tests, there was no significant difference in PMperformance across all ongoing tasks (all ps > .4). Samplecharacteristics are presented in Table 2.

698 M. A. BOWMAN ET AL.

The basic elements of the tripartite model were repli-cated in this dataset, such that BAI anxiety correlatedpositively with negative affect (r = 0.63, p < .001) andwas independent of positive affect (r = 0.02, p = .90),whereas BDI-II depression correlated positively withnegative affect (r = 0.57, p = .002) and negatively withpositive affect (r = −0.42, p = .03). The STAI’s stateanxiety pattern of correlations with the PANAS wasmore similar to depression (i.e., negatively correlatedwith positive affect, positively correlated with negativeaffect) than to anxiety assessed by the BAI, which isconsistent with criticisms that the STAI measuresdepression more so than it measures anxiety (Creamer,Foran, & Bell, 1995). Associations among self-reportmeasures are reported in Table 3.

Trait and state anxiety were associated with worseoverall PM performance as measured both by the BAI(r = −0.62, p< .001) and the STAI (p = .03), respec-tively. Negative affect was also associated with poorerPM performance (r = −0.45, p = .02). However, depres-sion scores (r = −0.27, p = .20) and positive affectscores (r = −0.21, p = .30) were not associated withPM performance (Table 3).

In sensitivity analyses, we examined whether the PMperformance from any specific task (i.e., living/non-livingtask, lexical decision-making task, and semantic categor-ization task; two opportunities in each task) was drivingthis relation. We found that anxiety assessed by BAI was

associated with PM performance within all three tasksindividually (all ps < 0.001). Anxiety assessed by the STAIwas significantly associated with PM performance withinthe lexical decision-making task only (r = −0.45, p = .02).Negative affect from the PANAS was significantly asso-ciated with PM performance on the lexical decision-mak-ing task (r = −0.41, p = .04) and the semanticcategorization task (r = −0.48, p = .01), but not theliving/nonliving task. Depression scores and positiveaffect scores were not significantly associated with PMperformance on any of the three tasks individually.

We next ran a series of hierarchical linear regressionanalyses to clarify the relative contributions of BAI(trait) and STAI (state) anxiety, BDI-II depression,and PANAS negative affect to overall PM performance.In the first step, PANAS negative affect was included.In the second step, BAI, STAI, or BDI-II was included(Table 4). Only BAI anxiety was associated with PMperformance over-and-above PANAS negative affect(β= −0.56, p = .01, R2 = 0.40). Moreover, adding BAIscores to the model nearly doubled the amount ofvariance explained and diminished the contributionof negative affect (Figure 1). In a model adjusting forboth BDI-II depression scores and PANAS negativeaffect, BAI scores were significantly associated withPM performance (β = −0.57, p = .01), suggesting thatthe trait component of elevated anxiety is the strongestcorrelate of diminished PM.

In sensitivity analyses, this same pattern of resultsoccurred when each of the three ongoing tasks (living/non-living task, lexical decision-making task, andsemantic categorization task) was considered individu-ally. Specifically, after adjusting for PANAS negativeaffect, BAI was a significant or trending correlate ofPM performance during two of the ongoing tasks –

Table 2. Sample characteristics.

Measure Minimum Maximum MeanSt.

Deviation

PM performance Task 0 6 3.9 2.0Living-nonliving Task 0 2 1.2 0.8Lexical decision Task 0 2 1.2 0.8Semantic Task 0 2 1.3 0.8

Anxiety BAI 0 27 8.9 6.9Anxiety STAI 24 62 37.7 7.9Depression BDI-II 0 23 4.4 4.9Positive affect PANAS 10 40 22 7.8Negative affect PANAS 10 26 12.7 3.8

PM: prospective memory; BAI: Beck Anxiety Inventory; STAI: State-TraitAnxiety Inventory; BDI-II: Beck Depression Inventory-II; PSQI: PANAS:Positive and Negative Affect Schedule.

Table 3. Bivariate associations among anxiety, depression, andmood measures.

PM Score BAI STAI BDI-II PANAS PA

BAI −0.62STAI −0.41 0.65BDI-II −0.27 0.43 0.62PANAS PA −0.21 0.02 −0.55 −0.42PANAS NA −0.45 0.63 0.63 0.57 −0.12

Correlation coefficients that are significant at p< 0.05 are indicated in bold.Performance indicates performance on the PM task.PM Score: PM Score; BAI: Beck Anxiety Inventory; STAI: State-TraitAnxiety Inventory; BDI-II: Beck Depression Inventory-II; PANAS PA:Positive and Negative Affect Schedule, Positive Affect; PANAS NA:Positive and Negative Affect Schedule, Negative Affect.

Table 4. Incremental predictive power of BAI, BDI-II, and STAIover PANAS NA.

β R2

Step 1:PANAS NA

−0.45* 0.20

Step 2:PANAS NA

−0.09

BAI −0.56* 0.40+

Step 1:PANAS NA

−0.45* 0.20

Step 2:PANAS NA

−0.32

STAI −0.21 0.23Step 1:PANAS NA

−0.45* 0.20

Step 2:PANAS NA

−0.44

BDI-II −0.02 0.20

p < .05 = *; significant R2 change at p < .05 = +.BAI: Beck Anxiety Inventory; STAI: State-Trait Anxiety Inventory; BDI-II:Beck Depression Inventory-II; PANAS NA: Positive and NegativeAffect Schedule, Negative Affect.

JOURNAL OF CLINICAL AND EXPERIMENTAL NEUROPSYCHOLOGY 699

semantic categorization (β = −0.48, p = .03), and living/non-living (β = −0.48, p = .06). Although the directionof results was the same for lexical decision-making(β = −0.34, p = .16) it was not significant, which maybe due to low power. BDI-II and STAI were not sig-nificant or trending correlates of PM performance dur-ing any of the three ongoing tasks after adjustment forPANAS negative affect (ps > 0.3).

Discussion

The current study examined the relationship between PMand anxiety, depression, and negative affect in a sample ofnon-diagnosed college students. Using the tripartitemodel as a framework to begin to disambiguate therelative contributions of various symptoms, we foundpreliminary evidence that, above and beyond negativeaffect, anxious symptoms assessed by the BAI are asso-ciated with poor PM performance. Because STAI was nota significant correlate of PM performance, the results ofthis study may suggest that trait, but not state, anxiety isthe more important component for successful executionof future tasks than the more affective componentsassessed by the STAI. Additionally, it may be that theBAI assesses the physiological arousal associated withanxiety, whereas the STAI assesses the more affectivecomponents of anxiety.

Anxiety, depression, and negative affect

Our analyses replicated the basic structure of the tri-partite model in the present sample, as theorized by

Clark and Watson (1991). That is, BAI anxiety waspositively related to PANAS negative affect and linearlyindependent of PANAS positive affect, whereas BDI-IIdepression was positively related to PANAS negativeaffect and negatively related to PANAS positive affect.Furthermore, PANAS positive and negative affectscales were linearly independent of each other.

Our hierarchical regression analyses indicated thattrait anxiety as measured by the BAI was associatedwith PM impairment over-and-above the contributionof PANAS negative affect; furthermore, including BAIscores in the model nearly doubled the amount ofvariance explained. The contribution of negative affectto diminished PM appears to be driven almost entirelyby anxiety, as the effect dropped to null when anxietyscores from the BAI were included in the model. Incontrast, neither the STAI nor the BDI-II providedincremental prediction over-and-above the contribu-tion of PANAS negative affect. This test using theframework of the tripartite model suggests that thesymptoms unique to anxiety are associated withimpaired PM performance.

Our findings suggest that anxiety and negative affectare related to impairments in PM performance on a focal,event-based task, while depressive symptoms had noeffect. This may be because depressive symptoms havebeen related to PM when the instruction is an event-based non-focal task (Altgassen et al., 2009; Livner,Berger, Karlsson, & Bäckman, 2008) or a time-basedPM task (Kliegel & Jäger, 2006), whereas our instructionwas an event-based focal task (depression was non-sig-nificant in two other studies with focal, event-based tasks;Altgassen et al., 2009; Li et al., 2013). Consistent with our

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

NA NA + BDI-II NA + STAI NA + BAI

Par

tial

Cor

rela

tion

with

Pro

spec

tive

Mem

ory

*

*

Figure 1. Contributions of anxiety and depression to negative affect.The variance in PM performance explained by negative affect (NA), as well as the contribution of BAI, STAI, and BDI-II above and beyond NA.Notes .*significant at p < .05NA: Positive Affect and Negative Affect Schedule, Negative Affect; BDI-II:Beck Depression Inventory-II; STAI: State-Trait Anxiety Inventory;BAI: BeckAnxiety Inventory.

700 M. A. BOWMAN ET AL.

results, another study that included anxiety and stress inthemodel also found that depression was not a significantcorrelate (Harris & Menzies, 1999). Therefore, the formof prospective memory tested may be important (i.e.,ones that require substantial disengagement from theongoing task), or previous studies that have found anassociation between depression and PM in focal, event-based tasks may have done so only because of its strongcorrelation with anxiety. As such, we would expect thatan individual experiencing both anxiety and depressionmight have worse PM than someone experiencing onlydepression. Using the tripartite model as a framework,our findings suggest that the effect of negative affect onPM is driven by the unique components of anxiety, suchthat previously obtained significant findings for relateddomains (depression, general negative affect) may actu-ally be an indirect function of those domains’ correlationswith anxiety or having an impact on other forms of PM.Due to power limitations, these results should be viewedas preliminary, yet our findings highlight this is an inter-esting area for future work on psychopathology and PMfunction.

Measurement considerations and limitations

The prospective memory tasks were selected to replicate aprevious study examining the role of sleep and prospec-tive memory (Scullin & McDaniel, 2010). This task wasevent-based and focal to the ongoing task. However,depression has been shown in previous studies to parti-cularly impact performance on time-based PM tasks(Kliegel & Jäger, 2006; Li et al., 2013; Rude et al., 1999),and non-focal tasks (Altgassen et al., 2009). Therefore,our task selection may be one reason that we do notreport an effect of depression. While the study did notdissociate prospective from retrospective components ofthe task (e.g., Arnold et al., 2015), it should be noted thatthe retrospective demands of the task wereminimal in thesense that participants only remembered two high-fre-quency nouns. In sensitivity analyses, we reported thatthere were significant associations between clinical mea-sures and PM performance on some, but not all, ongoingtasks. This was unexpected, as all PM cues were focal tothe ongoing tasks. Future workmight also evaluate differ-ences across PM tasks of the same type (focal versus non-focal) to determine if this is a unique feature of thecurrent study.

In the current study, participants had the opportu-nity to execute the PM task a total of six times. Only18% of our sample correctly executed all six of thetasks. Although Graf and Uttl (2001) questionedwhether a binary outcome (success or failure) has thevariability necessary to accurately reflect real-world PM

performance, Kelemen, Weinberg, Alford, Mulvey, andKaeochinda (2006) reported that studies with highvariability in performance had better alternate formreliability than those experiments where the majorityof participants executed all of the PM tasks. Based onthis evidence, our task may have been sufficiently chal-lenging to be a reliable measure of individuals’ capacityfor PM. However, we only tested participants in onesession, and so do not have the ability to directlymeasure reliability.

Our results for two measures of anxiety were incon-sistent. Although the BAI and STAI both purport tomeasure anxiety, our key findings were obtained onlyfor BAI scores. This difference in results may bebecause the BAI measures trait anxiety and physiolo-gical arousal, whereas the STAI measured state anxietyand affective components of anxiety. This discrepancycould also indicate construct validity issues with theSTAI. We found that patterns of STAI relations withPANAS scales resembled those that would be expectedfor a depression scale (positive relations with negativeaffect and negative relations with positive affect), ratherthan those that would be expected for an anxiety scale(positive relations with negative affect and no relationto positive affect). These findings are consistent withcriticisms (e.g., Bieling, Antony, & Swinson, 1998) thatthe STAI includes substantial depression content (e.g.,“I am regretful”), which necessarily compromises itsability to measure anxiety with fidelity. Anxiety mea-sures with more optimal construct validity (e.g., theBAI) than the STAI may be better recommended forfuture studies investigating associations between anxi-ety and cognitive performance.

The current study tested the tripartite model toexamine the association of emotional states and PM.Future work might evaluate the role of emotion on theprocess model of PM, which posits that there are fourstages to successful PM performance: an individualmust form an intention to do something (intentionformation), retain the intention during other ongoingactivities (intention retention), initiate the intention atthe appropriate time (intention initiation), and executethe action as planned (Kliegel, Martin, McDaniel, &Einstein, 2002). Based on the literature reviewed inTable 1, we think it is plausible that anxiety may inter-fere with intention retention (e.g., Kliegel, Altgassen,Hering, & Rose, 2011; Kliegel et al., 2002), that depres-sion may interfere with the intention initiation compo-nent (Kliegel & Jäger, 2006; Li et al., 2013; Rude et al.,1999), and that affect may be related to intentionexecution (Kliegel et al., 2005; Schnitzspahn et al.,2014). This integration of the tripartite model withthe process model of PM is speculative based on

JOURNAL OF CLINICAL AND EXPERIMENTAL NEUROPSYCHOLOGY 701

current literature. The current study is a preliminarystep to merge the cognitive concept of prospectivememory with the clinical framework of the tripartitemodel. A larger-scale study designed to test how theprocess model of PM and the tripartite model might bethe next step in bringing these two fields together.

Our study has three critical limitations that must benoted. First, the small sample size limited our powerwhen including more than one independent variable inour regression models, so these results should be con-sidered preliminary and interpreted with caution at thisstage. Second, the sample used here was non-clinicalgroup of college-aged students. Although the decisionto use such a sample was consistent with our initialgoal, clinical symptoms reported were only in the mildto moderate range. This was particularly true fordepression symptoms, in which the average BDI-IIwas 4.4 (4.9) with a range of 0–23. This reducedrange could have had an effect on our correlationanalysis and may have contributed to the findingsthat depression did not have an effect on this task. Assuch, it will be necessary to replicate the findingsreported here with larger samples, both in non-diag-nosed and clinical populations to fully determine theirscope and validity. Finally, our study is correlational innature, and thus cannot be used to understand direc-tionality of associations.

In spite of these limitations, our study is novel inthat it is the first to examine PM performance with theframework of the tripartite model, and while prelimin-ary, it is hypothesis-generating for future studies of itskind. Additionally, it is important to note that thefindings reported here fall directly in line with previousresearch in a non-clinical sample suggesting that anxi-ety symptoms impair PM performance in focal, event-based tasks while depression symptoms do not (Harris& Menzies, 1999). As such, our findings provide addi-tional evidence of possible differential associationsbetween clinical symptoms and PM performance, andsuggest that accounting for negative affect may beimportant for disentangling these symptoms.

Conclusion

To our knowledge, this study is the first to compare therelative contributions of self-reported anxiety, depres-sion, and negative affect to PM performance. Using theframework of the tripartite model, our findings providepreliminary evidence that the unique components ofanxiety (i.e., somatic tension and arousal) and, at leastindirectly, overall negative affect, have a meaningfulimpact on future-oriented cognitive functioning.Methodologically, results from this study highlight the

value of incorporating clinical psychology frameworkswith cognitive psychology tasks. By including perspec-tives from psychometrics and clinical science, we wereable to examine clinical variables with greater theore-tical precision and thereby consider more generalizableapplications for cognitive research. More widespreaduse of this approach in research might improve ourunderstanding of psychological phenomena and thepotential applications of our findings.

Acknowledgments

The authors thank Dr. Nathan Rose for his scientific advice,as well as the research assistants in the Sleep, Stress, andMemory Lab for data collection.

Data availability statement

These data are not from a publicly available dataset. Foraccess to script, output, and data, please contact the corre-sponding author.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This study was funded by a grant from the University ofNotre Dame: Institute for Scholarship in the Liberal ArtsFounders’ and Directors’ 30th Anniversary ResearchAward;University of Notre Dame: Institute for Scholarshipin the Liberal Arts Founders’ and Directors’ 30thAnniversary Research Award.

References

Altgassen, M., Henry, J. D., Bürgler, S., & Kliegel, M. (2011).The influence of emotional target cues on prospectivememory performance in depression. Journal of Clinicaland Experimental Neuropsychology, 33(8), 910–916.

Altgassen, M., Kliegel, M., & Martin, M. (2009). Event-basedprospective memory in depression: The impact of cuefocality. Cognition and Emotion, 23(6), 1041–1055.

Arnold, N. R., Bayen, U. J., & Böhm, M. F. (2015). Is pro-spective memory related to depression and anxiety? Ahierarchical MPT modelling approach. Memory, 23(8),1215–1228.

Beck, A. T., Epstein, N., Brown, G., & Steer, R. A. (1988). Aninventory for measuring clinical anxiety: Psychometricproperties. Journal of Consulting and Clinical Psychology,56(6), 893–897.

Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Beckdepression inventory-II. San Antonio, 78(2), 490–498.

Bhui, K. S., Dinos, S., Stansfeld, S. A., & White, P. D. (2012).A synthesis of the evidence for managing stress at work: Areview of the reviews reporting on anxiety, depression,

702 M. A. BOWMAN ET AL.

and absenteeism. Journal of Environmental and PublicHealth. doi:10.1155/2012/515874

Bieling, P. J., Antony, M. M., & Swinson, R. P. (1998). Thestate-trait anxiety inventory, trait version: Structure andcontent re-examined. Behaviour Research and Therapy, 36(7), 777–788.

Brown, G. K., Beck, A. T., Newman, C. F., Beck, J. S., & Tran,G. Q. (1997). A comparison of focused and standardcognitive therapy for panic disorder. Journal of AnxietyDisorders, 11(3), 329–345.

Clark, L. A., & Watson, D. (1991). Tripartite model of anxietyand depression: Psychometric evidence and taxonomic impli-cations. Journal of Abnormal Psychology, 100(3), 316–336.

Creamer, M., Foran, J., & Bell, R. (1995). The Beck AnxietyInventory in a non-clinical sample. Behaviour Researchand Therapy, 33(4), 477–485.

Cuttler, C., & Graf, P. (2008). Sub-clinical checking compul-sions are related to impaired prospective memory inde-pendently of depression, anxiety and distractibility.Journal of Anxiety Disorders, 22(4), 642–654.

DiMatteo, M. R., Lepper, H. S., & Croghan, T. W. (2000).Depression is a risk factor for noncompliance with med-ical treatment: Meta-analysis of the effects of anxiety anddepression on patient adherence. Archives of InternalMedicine, 160(14), 2101–2107.

Ellis, H. C., & Ashbrook, P. W. (1988). Resource allocationmodel of the effects of depressed mood states on memory.In K. Fiedler & J. Forgas (Eds.), Affect, cognition and socialbehavior (pp. 25–43). Toronto, UK: Hogrefe.

Folkard, S., & Monk, T. H. (1980). Circadian rhythms inhuman memory. British Journal of Psychology, 71(2),295–307.

Graf, P., & Uttl, B. (2001). Prospective memory: A new focusfor research. Consciousness and Cognition, 10(4), 437–450.

Harris, L. M., & Cumming, S. R. (2003). An examination ofthe relationship between anxiety and performance on pro-spective and retrospective memory tasks. AustralianJournal of Psychology, 55(1), 51–55.

Harris, L. M., & Menzies, R. G. (1999). Mood and prospec-tive memory. Memory, 7(1), 117–127.

Hirschfield, R. M. (2001). The comorbidity of major depres-sion and anxiety disorders: Recognition and managementin primary care. Primary Care Companion to the Journal ofClinical Psychiatry, 3(6), 244.

Joiner, T. E. (1996). A confirmatory factor-analytic investiga-tion of the tripartite model of depression and anxiety incollege students. Cognitive Therapy and Research, 20(5),521–539.

Kelemen, W. L., Weinberg, W. B., Alford, H. S., Mulvey, E.K., & Kaeochinda, K. F. (2006). Improving the reliabilityof event-based laboratory tests of prospective memory.Psychonomic Bulletin & Review, 13(6), 1028–1032.

Kliegel, M., Altgassen, M., Hering, A., & Rose, N. S. (2011). Aprocess-model based approach to prospective memory

impairment in Parkinson‘s disease. Neuropsychologia, 49(8), 2166–2177.

Kliegel, M., & Jäger, T. (2006). The influence of negative emo-tions on prospective memory: A review and new data.International Journal of Computational Cognition, 4(1), 1–17.

Kliegel, M., Jäger, T., Phillips, L., Federspiel, E., Imfeld, A.,Keller, M., & Zimprich, D. (2005). Effect of sad mood ontime-based prospective memory. Cognition and Emotion,19(8), 1199–1213.

Kliegel, M., Martin, M., McDaniel, M. A., & Einstein, G. O.(2002). Complex prospective memory and executive con-trol of working memory: A process model. PsychologicalTest and Assessment Modeling, 44(2), 303.

Li, Y. R., Weinborn, M., Loft, S., & Maybery, M. (2013).Patterns of prospective memory impairment among indi-viduals with depression: The influence of cue type anddelay interval. Journal of the InternationalNeuropsychological Society, 19(6), 718–722.

Livner, Å., Berger, A. K., Karlsson, S., & Bäckman, L. (2008).Differential effects of depressive symptoms on prospectiveand retrospective memory in old age. Journal of Clinicaland Experimental Neuropsychology, 30(3), 272–279.

McDaniel, M. A., & Einstein, G. O. (2007). Prospective mem-ory: An overview and synthesis of an emerging field.Thousand Oaks, CA: Sage Publications.

Rude, S. S., Hertel, P. T., Jarrold, W., Covich, J., & Hedlund,S. (1999). Depression-related impairments in prospectivememory. Cognition & Emotion, 13, 267–276.

Schneider, W., Eschman, A., & Zuccolotto, A. (2002). E-Prime: User‘s guide. Pittsburgh, PA: Psychology SoftwareIncorporated.

Schnitzspahn, K. M., Thorley, C., Phillips, L., Voigt, B.,Threadgold, E., Hammond, E. R., . . . Kliegel, M. (2014).Mood impairs time-based prospective memory in youngbut not older adults: The mediating role of attentionalcontrol. Psychology and Aging, 29(2), 264.

Scullin, M. K., & McDaniel, M. A. (2010). Remembering toexecute a goal: Sleep on it! Psychological Science, 21(7), 1028–1035.

Spielberger, C. D., Gorsuch, R. L., Lushene, R., Vagg, P. R., &Jacobs, G. A. (1983). Manual for the state-trait anxietyinventory. Palo Alto, CA: Consulting Psychologists Press.

Watson, D., Clark, L. A., & Tellegen, A. (1988). Developmentand validation of brief measures of positive and negativeaffect: The PANAS scales. Journal of Personality and SocialPsychology, 54(6), 1063.

Watson, D., Clark, L. A., Weber, K., Assenheimer, J. S.,Strauss, M. E., & McCormick, R. A. (1995). Testing atripartite model: II. Exploring the symptom structure ofanxiety and depression in student, adult, and patient sam-ples. Journal of Abnormal Psychology, 104(1), 15–25.

Zigmond, A. S., & Snaith, R. P. (1983). The hospital anxietyand depression scale. Acta Psychiatrica Scandinavica, 67(6), 361–370.

JOURNAL OF CLINICAL AND EXPERIMENTAL NEUROPSYCHOLOGY 703