Assessment of anhedonia in psychological trauma: psychometric and neuroimaging perspectives.

0.77 MB·English

by  Frewen, Paul A.

(cid:1)CLINICAL RESEARCH ARTICLE
Assessment of anhedonia in psychological
trauma: psychometric and neuroimaging
perspectives
Paul A. Frewen1,2,3*, David J. A. Dozois1,2 and Ruth
A. Lanius1,3
1DepartmentofPsychiatry,UniversityofWesternOntario,London,Ontario,Canada;2Departmentof
Psychology, University ofWestern Ontario,London, Ontario, Canada; 3GraduateProgram in
Neuroscience,University ofWestern Ontario, London, Ontario, Canada
Symptomsofanhedonia,ordeficitsintheabilitytoexperiencepositiveaffect,areincreasinglyrecognizedas
anoutcomeoftraumaticstressincludinginindividualswithPTSD.However,littleresearchhasinvestigated
negative affective responses to what would normally be considered pleasant events (e.g., receiving a
compliment or gift, physical affection) in traumatized persons. We demonstrate not only self-reported
decreased positive affect but also increased negative affect in response to positive events in 55 women with
PTSD,incomparisonwith35womenwithoutPTSD,viatheirresponsetoaHedonicDeficit&Interference
Scale(HDIS).TheHDISdemonstratedstronginternalvalidity,convergentandincrementalvalidityrelative
toothermeasuresofanhedonia,anddiscriminantvalidityinrelationtodepressionversusanxietysymptoms
in this sample. In addition, in response to imageryof social versusnon-social positive events, HDIS scores
predictedself-reportpositiveandnegativeaffectiveresponses.Inasub-sampleofparticipantscompletingthe
imagery task while undergoing fMRI (n(cid:1)12), HDIS scores also predicted BOLD response within the left
orbitofrontalcortex,ventromedialprefrontalcortex,amygdala,andcerebellum.Futureresearchandclinical
directions arediscussed.
Keywords: anhedonia;positiveaffect;negativeaffect;negativeaffectiveinterference;PTSD;depression
Fortheabstractorfulltextinotherlanguages,pleaseseeSupplementaryfilesunderReadingTools
online
Received:19 August 2011; Revised:30 November 2011; Accepted:11 December 2011;Published:11January2012
AnhedoniawasdefinedbyRibot(1896)toreferto (criterion C5, American Psychiatric Association, 2000,
adeficiencyinanindividual’scapacitytoexperi- p.464).Researchalsoshowsthatsymptomsofemotional
ence positive affect in situations that should numbing (e.g., Orsillo, Theodore-Oklota, Luterek, &
normally provoke it. Such symptoms have long since Plumb, 2007; Ramirez et al., 2001) may be particularly
been recognized in individualssuffering from psychiatric related to anhedonia. Functional neuroimaging studies
disordersincludingmajordepression,schizophrenia,and suggest that individual differences in anhedonic proces-
substance abuse disorders. More recently, symptoms of sing are partly represented within the ventromedial
anhedonia have also been systematically observed in prefrontal cortex (VMPFC) during visual emotional
individuals with posttraumatic stress disorder (PTSD; processing (Harvey, Pruessner, Czechowska, & Lepage,
e.g., Kashdan, Elhai, & Frueh, 2006, 2007) which 2007; Keedwell, Andrew, Williams, Brammer, & Phillips,
includes as diagnostic criteria a diminished interest or 2005).Theleftorbitofrontalcortex(L-OFC)alsohasan
participation in previously enjoyed activities (criterion established role in neural processing relevant to reward
C4) and a reduced ability to feel emotions, particularly functioning and positive affect (reviewed by Burgdorf &
those associatedwith intimacy, tenderness, and sexuality Panksepp, 2006) as does the amygdala (Murray, 2007)
EuropeanJournalofPsychotraumatology2012. #2012PaulA.Frewenetal.ThisisanOpenAccessarticledistributedunderthetermsoftheCreative 1
CommonsAttribution-Noncommercial3.0UnportedLicense(http://creativecommons.org/licenses/by-nc/3.0/),permittingallnon-commercialuse,distribution,
andreproductioninanymedium,providedtheoriginalworkisproperlycited.
Citation:EuropeanJournalofPsychotraumatology2012,3:8587-DOI:10.3402/ejpt.v3i0.8587
(pagenumbernotforcitationpurpose)
PaulA.Frewenetal.
and cerebellum (e.g., Stoodley & Schmahmann, 2009; affective interference (i.e., negative affective responses to
Turner et al., 2007). positivestimuliandevents)inindividualsasafunctionof
Anhedoniaasasymptomistypicallymeasuredbyself- traumahistoryandtrauma-relatedsymptoms.Wefurther
report (Leventhal & Rehm, 2005) via measures such as examinethepsychometricpropertiesofaHedonicDeficit
the Physical and Social Anhedonia Scales (PSAS; Chap- & Interference Scale (HDIS; Frewen, Dean, & Lanius,
man, Chapman, & Raulin, 1978), the Fawcett-Clark 2012) as a brief method for directly assessing hedonic
PleasureCapacityScale(FCPCS;Fawcett,Clark,Scheft- deficits as distinguished from negative affective interfer-
ner, & Gibbons, 1983) and/or the Snaith-Hamilton ence in women with PTSD predominantly related to
Pleasure Scale (SHAPS; Snaith et al., 1995). The assess- childhood abuse. We investigate the convergent, incre-
ment method taken by each of these instruments is to mental, discriminant, and concurrent criterion-related
querythedegreetowhichrespondentsbelievetheywould validity of the HDIS using a multi-method design
experience positive affect in response to stimuli and (questionnaires, self-report response to emotional ima-
events that normally provoke it (e.g., ‘‘You sit watching gery, and neural response to emotional imagery via
a beautiful sunset in an isolated, untouched part of functional magnetic resonance imaging [fMRI]). As
theworld’’[FCPCSItem#1],‘‘Iwouldenjoymyfavorite well, this is the first study of which the authors’ are
television or radio programme’’ [SHAPS Item #1]). aware to evaluate anhedonia in individuals with PTSD
The assumption underlying this approach is that low using standardized measures (i.e., the SHAPS, Snaith
positiveaffectiveresponsestopleasantstimuliandevents et al., 1995; and the FCPCS, Fawcett et al., 1983).
indirectly indicate the presence of anhedonia, that is, the
inability to experience positive affect in response to such Method
events. However, a limitation of this approach is that
it fails to directly measure the perceived inability to Participants
experience positive affect apart from individual differ- Ninety women took part in this study. Participantswere
ences in the intensity with which respondents experience recruited over a 30-month period via advertisements
positive affect. placed in local community, hospitals, and newspapers
Furthermore,inadditiontodifficultiesinexperiencing targeting individualswho had experienced traumatic life
pleasure, individuals with PTSD often experience ele- events. Participants either reported no current or past
vated negative emotions including anxious hyperarousal psychiatric history or history of child maltreatment (the
(Pole, 2007), anger (Orth & Wieland, 2006), guilt and psychologically healthy control group, n(cid:1)35), or met
shame (Kim,Talbot, &Cicchetti, 2009;Leskela,Dieper- DSM-IV diagnostic criteria for current chronic PTSD
ink, & Thuras, 2002). Emotion-regulation perspectives (n(cid:1)55).PTSDdiagnosticstatusandsymptomfrequency
(e.g., McCullough et al., 2003) suggest that a key task is and severity was determined by the Clinician Adminis-
to determinewhether positive stimuli and events may be tered PTSD Scale (CAPS; Blake et al., 1995), and
responded to not only with less than expected positive comorbid diagnoses were determined by the Structured
affect, potentially reflecting a hedonic deficit, but also Clinical Interview for DSM-IV (SCID-I; First, Gibbon,
with an increase in negative affect. In other words, Spitzer, & Williams, 1996), which are widely considered
individuals may not only exhibit deficient positive gold-standard measures. Childhood trauma history was
affective responses to positive events (e.g., responding measured by the Childhood Trauma Questionnaire
with disinterest, dullness, blunting) but they may also Short-form (CTQ; Bernstein & Fink, 1998); described
experience interfering negative affect (e.g., anxiety, guilt, below.Demographicinformation,inadditiontodescrip-
shame, disgust). A limitation of present measures of tive information pertaining to clinical severity and
anhedonia is that they only assess the degree of positive comorbid psychiatric diagnosis, is reported in Table 1.
affect experienced in response to positive stimuli, not Differences in mean age between groups did not alter
taking account of negative affective responses that may interpretation of the principal results of this study as
occurtopositivestimuli.Bysolelyassessingthedegreeof examined by covariance analyses (not reported). Poorer
hedonic deficit experienced in response to positive employmentstatusandeducationlevelaswellasmarital/
stimuli, one cannot know whether anhedonic symptoms relational problems are among the recognized long-term
are associated only with low positive affect (e.g., disin- associationsofchildhoodtraumaandwere thereforenot
terest,dullness,blunting)andlowpleasantness(e.g.,little covaried in analyses.
happiness or pleasure), or are also accompanied by
interfering negative affective responses such as distress Measures
or disgust (Tellegen, Watson, & Clark, 1999). Hedonic Deficit & Interference Scale (HDIS; Frewen,
The purpose of the present study was to examine Dean, & Lanius, 2012). The HDIS was administered in
hedonic deficits (i.e., self-reported difficulties experien- order to directly assess negative affect interference (i.e.,
cing positive affect) as distinguished from negative secondarynegative affectiveresponsestopositivestimuli
2
Citation:EuropeanJournalofPsychotraumatology2012,3:8587-DOI:10.3402/ejpt.v3i0.8587
(pagenumbernotforcitationpurpose)
Assessmentofanhedoniainpsychologicaltrauma
Table1. Demographicanddiagnostic information
PTSDSample(n(cid:1)55) Controls(n(cid:1)35)
Demographics
MeanAge(SD) 39.85(8.03)* 30.40(12.86)
Caucasian 91% 68%
Married/common-law 24% 39%
Separated/divorced 46%* 4%
Single 33%* 57%
Completedsomeorcurrentlycompleting 72% 69%
post-secondaryeducation
Completedsecondary-school 92% 100%
Employed(fullorparttime)orcurrentstudent 42%* 87%
SeverityofPTSD
MeanCAPS(SD),range 77.42(16.52),54(cid:2)120* 1.13(4.32),0(cid:2)22
ChildhoodTraumaQuestionnaire:M(SD),percentileequivalent
ofMean)1
Emotionalabuse 18.37(5.84),97thpercentile* 6.57(1.99),29thpercentile
Emotionalneglect 17.17(5.37),93rdpercentile* 7.21(2.48),25thpercentile
Physicalabuse 12.93(6.07),96thpercentile* 5.76(1.46),37thpercentile
Physicalneglect 12.02(4.38),98thpercentile* 5.19(0.59),31thpercentile
Sexualabuse 15.19(7.93),98thpercentile* 5.33(1.10),36thpercentile
Comorbidaxisiconditions(n) Past Current *
Alcoholabuse 7 1
Substanceabuse 3 1
Majordepressivedisorder 6 12
Dysthymia 0 4
Panicdisorderw/woagoraphobia 1 15
Agoraphobiawopanicdisorder 0 8
Socialphobia 1 13
Specificphobia 0 14
Obsessivecompulsivedisorder 3 6
Generalizedanxietydisorder 0 7
Somatizationdisorder 0 2
Undifferentiatedsomatoformdisorder 0 7
Paindisorder 0 2
Hypochondriasis 0 2
Anorexianervosa 3 0
Bulimianervosa 6 0
EatingdisorderNOS 1 1
Note:*significantlydifferentbetweengroups(pB0.05).1Percentilesrelativetothenormativepopulationoffemalehealthmanagement
organizationmembers(N(cid:1)1187)describedinBernstein&Fink(1998,Table4.5).DSM-IVDisordersnotlistedwerenotpresentinthe
sample.‘‘SD’’,standarddeviation,‘‘CAPS’’,clinicianadministeredPTSDScale.ComormidpsychiatricconditionsassessedviatheSCID-I;
Pastorpresentpsychiatricdiagnosiswasanexclusionarycriterionforthecontrolgroup.
and events) in addition to, and as distinct from, hedonic when good things in your life happen’’ (Items 6(cid:2)10;
deficits (i.e, difficulties in experiencing positive affect). Hedonicdeficit[HD]subscale),and11itemsaskwhether
Five items assess positive emotionality (HDIS Positive interfering negative affective consequences tend to occur
Emotionality[PE]subscale),fiveitemsasktheinformant whenpositiveeventshappenintheindividual’slife(Items
whetherheorshe‘‘can’t(youarenotableto)experience 11(cid:2)21; HDIS Negative Affective Interference [NAI]
[insertingseparatelyeachofthesamefivedistinctpositive subscale, i.e., whether participants commonly felt numb,
affectsusedinthePEscale]evenwhenyoutry,andeven dissociative, anxious, fearful, guilty, self-critical, shame,
3
Citation:EuropeanJournalofPsychotraumatology2012,3:8587-DOI:10.3402/ejpt.v3i0.8587
(pagenumbernotforcitationpurpose)
PaulA.Frewenetal.
disgust, emotionally empty, lifeless, and/or purposely inventoryofsymptomsofdepression(e.g.,‘‘Ifeltthatlife
attempted to suppress positive feelings, specifically in was meaningless’’), anxiety (e.g., ‘‘I felt I was close to
response to prototypically positive events). Items are panic’’), and stress (e.g., ‘‘I found it difficult to relax’’).
ratedforfrequencywithinthepastmonthonan11-point Studiesattest tothereliabilityandvalidityoftheDASS-
ratingscalerangingfrom0(‘‘NotatAllorNeverTrue’’) 21 (e.g., Antony, Bieling, Cox, Enns, & Swinson, 1998).
to 10 (‘‘Completely true or very frequent [Always or The DASS-21-Depression subscale was administered to
Almost Always the Case]’’), with 5 referring to ‘‘Moder- assessconcurrentcriterion-relatedvalidityfortheHDIS,
ately True or Moderately Frequent’’ and no other item whereas the DASS-21-Anxiety scale wasadministered to
anchors. Excellent psychometric characteristics were assessthediscriminantvalidityoftheHDIS,onthebasis
reportedfortheHDISin99undergraduateswithvariable that current theoretical models of mood and anxiety
trauma histories (Frewen, Dean, & Lanius, 2012), how- symptomatology hypothesize that anhedonia is more
ever, the present study is the first to evaluate the strongly associated with depressed mood than with
psychometric characteristics of the HDIS in individuals anxious hyperarousal (e.g., the tripartite model; Clark
with clinician-diagnosed PTSD. & Watson, 1991). There were no specific predictions
Snaith-HamiltonPleasureScale(SHAPS;Snaithetal., made for the DASS-21-Stress scale and therefore such
1995).TheSHAPSwasadministeredtoassessanhedonia associationswere not investigated.
inPTSDwithanestablishedmeasure,aswellastoassess Childhood Trauma Questionnaire (cid:2)Short Form (CTQ-
convergent validity for the HDIS. The SHAPS is a 14- SF; Bernstein & Fink, 1998). The CTQ-SF is a widely-
item scale requiring respondents to indicate their per- used and standardized retrospective measure of adults’
ceived ability to experience pleasure in response to a list exposureto traumatic events during theirchildhood and
ofsituationsifthesesituationshadoccurredoverthelast adolescence. The CTQ-SF has five subscales: Emotional
few days (e.g., ‘‘I would enjoy my favorite television or Neglect(reverse-scored;e.g.,‘‘Iknewtherewassomeone
radioprogram’’,‘‘Iwouldenjoybeingwithmyfamilyor to take care of me and protect me’’), Emotional Abuse
close friends’’). Lower experience of pleasure in these (e.g., ‘‘Peoplein my familycalled methings like ‘stupid’,
circumstances is considered to indirectly reflect the ‘lazy’, or ‘ugly’’’), Sexual Abuse (e.g., ‘‘Someone tried to
inability to experience pleasure (i.e., anhedonia) as touch me in a sexual way, or tried to make me touch
described above. Items were scored in such a way that them’’),PhysicalAbuse(e.g.,‘‘Peopleinmyfamilyhitme
highscoresreflectedloweragreementthattherespondent so hard that it left me with bruises or marks’’), and
would enjoy each circumstance (i.e., greater presumed Physical Neglect (e.g., ‘‘I didn’t have enough to eat’’).
anhedonia). Psychometric support for the SHAPS was Excellentpsychometriccharacteristicshavebeenreported
provided in previous studies (Franken, Rassin, & Muris, by Bernstein et al. (2003).
2007; Gilbert, Allan, Brough, Melley, & Miles, 2002; Affective Response Test (cid:2) Positive Version (ART-P).
Leventhal, Chasson, Tapia, Miller, & Pettit, 2006). Thistaskwasadministeredtoassessresponsestopositive
Fawcett-Clark Pleasure Capacity Scale (FCPCS; Faw- stimuli and events within an experimental context,
cett et al., 1983). The FCPCS was also administered to affording measurement of associated self-report and
assess anhedonia in PTSD using an established measure functional neural responses. The primary results from
(toassessconvergentvalidityfortheHDIS).TheFCPCS this sample regarding performance of the ART-P have
is a 36-item questionnaire that also requires respondents been reported previously (Frewen et al., 2010). In brief,
to imagine themselves in various pleasurable situations participants listened to and imagined twelve 30-second
(e.g., ‘‘You are listening to beautiful music in peaceful audio-scripted vignettes happening to themselves, and
surroundings’’)andthenratethedegreeofpleasurethey attendedtowardtheiremotionalresponsestothescripts.
experience in consequence on a 5-point Likert scale Half(n(cid:1)6)ofthescriptstendtoelicitpositiveemotional
ranging from ‘‘No pleasure at all’’ to ‘‘Extreme and experiencesofmildtomoderateself-reportedintensityin
lastingpleasure.’’Itemswereagainscoredsuchthathigh healthy individuals, whereas the remaining scripts de-
scores reflected a lower experience of pleasure (i.e., scribescenariosthatonaveragearouseexperienceslower
greater presumed anhedonia). Several psychometric stu- in emotional intensity (neutral scripts). Analyses for this
dies support the use of the FCPCS (see Leventhal & study were restricted to the patently emotional scripts,
Rehm, 2005; Leventhal et al., 2006, for reviews). Keed- as there were no specific predictions for anhedonia to
well et al. (2005) demonstrated that FCPCS scores be associated with imagery for relatively emotionally-
predict response within the VMPFC during recall of insignificant (i.e., neutral) events. Scripts were further
positiveeventsandviewingofhappyfacialexpressionsin divided in terms of those wherein the positive affect
a fMRI studyof 12 individualswith major depression. generated primarily occurs within the context of inter-
DepressionAnxietyStressScales-Shortform(DASS-21; personal interaction (social-positive; e.g., receiving a
Lovibond& Lovibond,1995a, b). Theshort-form ofthe warm greeting or compliment), and those wherein inter-
DASS-21 (Lovibond & Lovibond, 1995a, b) is a 21-item personal interaction is either absent or not emphasized
4
Citation:EuropeanJournalofPsychotraumatology2012,3:8587-DOI:10.3402/ejpt.v3i0.8587
(pagenumbernotforcitationpurpose)
Assessmentofanhedoniainpsychologicaltrauma
(e.g., a solitary walkon the beach, or enjoying a bubble http://www.fil.ion.ucl.ac.uk/spm), creating contrast
bath) (Frewen et al., 2011). images.Participants’HDISscoreswereregressedontheir
Following each imagined situation, participants were individual contrast images to identify clusters of activa-
asked a series of questions concerning what they experi- tion associated with anhedonia (k [cluster-size] ] 32
enced,includingthedegreetowhichtheyexperiencedthe voxels [representing approximately 25% the width of the
following affective responses presented in this order: default smoothing kernel [8mm] used in SPM2, voxels
happy,increasedself-esteem,relaxation,physicalpleasure being resampled at 2mm3 within SPM2]; except within
(all coded as positive affective responses) and fear, theamygdalawhereinthecluster-sizethresholdwask ]
anxiety, sadness, shame, anger, disgust, and ‘‘feeling 10 voxels). Alphavalues for all analyses were set at pB
emotionally numb’’ (all coded as negative affective 0.005 balancing risk of Type I and II errors (see
responses). Providing that our previous research sug- Liebermann & Cunningham, 2009). Analyses employed
gested that HDIS scores relate differentially to specific a random-effects model wherein degrees of freedom
negative emotional responses (Frewen, Dean, & Lanius, represent the number of participants (n(cid:1)12) less one.
2012), however, we also examined the negative affective Coordinates are in accordance with the stereotaxic
responses separately. The affective response ratings were system of the Montreal Neurological Institute (MNI).
given on scales from 0 to 3, where zero indicated ‘‘No Wewereparticularlyinterestedtoexaminecorrelations
increase in emotion’’, and ratings one, two, and three between HDIS scores and the following regions of
referenced the participants’ perception that they ‘‘felt interest: bilateral amygdala, insula, and temporal pole
slightly/somewhat’’, ‘‘felt moderately strong’’, and ‘‘felt (wherewepreviouslyobservedstatepositiveandnegative
strongly or very strongly’’ each particular affective emotional ratings predicted response during non-social
response. Participants were also asked whether they positive emotional imageryinthe present group;Frewen
wished to avoid experiencing positive and negative et al., 2010), dorsomedial prefrontal cortex (where
emotional events during imagery using the same item healthywomenrespondedmorestronglythanthepresent
anchors: ‘‘No avoidance’’, ‘‘Slightly/somewhat avoided’’, group on average during social positive imagery; Frewen
‘‘moderately strongly avoided’’, and ‘‘strongly or very et al., 2010), ventromedial prefrontal cortex (where
strongly avoided’’. Note that descriptive information previousstudieshaveobservedcorrelationswithanhedo-
regarding self-report responses, BOLD responses, and niasymptoms;Harveyetal.,2007;Keedwelletal.,2005),
thecorrelationbetweenthesemeasuresinresponsetothe orbitofrontalcortex(knowntobeinvolvedinresponseto
ART-P has been reported previously (Frewen et al., reward; Burgdorf & Panksepp, 2006), and cerebellum
2010); the present manuscript represents a follow-up (increasingly recognized as being involved in emotional
investigation specifically regarding the association be- processing; e.g., Stoodley & Schmahmann, 2009; Turner
tween such measures and self-reported anhedonia symp- et al., 2007).
toms as measured by the HDIS.
Functional Magnetic Resonance Imaging (fMRI). All Procedure
bloodoxygenationleveldependent(BOLD)imagingdata Allparticipantsprovidedwritteninformedconsentbefore
were collected on a 4 Tesla Varian UNITYINVOAwhole participating and were debriefed afterward. Upon con-
body scanner equipped with Siemens Sonata gradients tactingresearchpersonnelbytelephonewithanintentto
and a quadrature hybrid birdcage radiofrequency (RF) participate, participantswere pre-screened for likelihood
head coil. Prior to functional imaging, for anatomical of child maltreatment history and psychiatric diagnostic
registration, high resolution T1-weighted images were status,followingwhichrecruitmentstrategiesfavoredthe
acquired with a 3D GE pulse sequence with spiraled inclusion of participants with very high or very low
gradient waveforms (256(cid:4)256 matrix size, 64(cid:4)2.5 mm likelihoods of meeting diagnostic criteria for moderate-
slices,TR(cid:1)50msec.,TE(cid:1)3msec.,TI(cid:1)1300msec.,flip to-severe PTSD. Participants were then tested individu-
angle(cid:1)208). fMRI was conducted as follows: 25 con- ally and the HDIS was administered as an interview
tiguous slices, 5-mm thick, were acquired using an during a session in which the CAPS (Blake et al., 1995)
interleaved, two-segment gradient echo (GE) pulse se- and SCID-I (First et al., 1996) were also conducted to
quence with spiraled gradient waveforms (FOV(cid:1)22 cm, formally assess diagnostic status, and the Childhood
64(cid:4)64 matrix size, TR(cid:1)1.5 sec, TE(cid:1)15 msec, flip Trauma Questionnaire (CTQ; Bernstein & Fink, 1998)
angle(cid:1)608). Please see Frewen et al. (2010) for descrip- was administered to assess child maltreatment history.
tion of subtraction analyses. In brief, differences in TheART-PwasadministeredbycomputerusingE-Prime
location and intensity of BOLD response during the Softwareduringa subsequent testing sessionheldwithin
positive event script-driven imagery task relative to two weeks of the interview session, at which time the
baseline scanning (30-seconds preceding each script other paper-and-pencil questionnaires (SHAPS, FCPCS,
onset) were ascertained by use of standard subtraction DASS-21) were also completed. Participants who met
analyses using Statistical Parametric Mapping 2 (SPM2: diagnostic criteria for PTSD as measured by the CAPS
5
Citation:EuropeanJournalofPsychotraumatology2012,3:8587-DOI:10.3402/ejpt.v3i0.8587
(pagenumbernotforcitationpurpose)
PaulA.Frewenetal.
were recruited to complete the ART-P within an fMRI comparisons. As predicted, women with PTSD scored
environment if: 1) they were not currently being treated significantly lower in HDIS-Positive Emotionality (d(cid:1)
fortheirpsychiatriccondition(s)bypsychotropicmedica- 1.91), and significantly higher in HDIS-Hedonic Deficit
tions, and 2) they met standard safety precautionary (d(cid:1)1.88)andHDIS-NegativeAffectiveInterference(d(cid:1)
criteria for MRI; 14 participants met these inclusion 1.65),relativetowomenwithoutcurrentorpastpsychia-
criteria,althoughHDISscoreswerenotcollectedfortwo tricproblems.Infact,inallocatingparticipantstogroups
participants, leaving 12 available for fMRI analysis. The (PTSD vs. control) an HDIS-Hedonic Deficit score
remaining participants completed the ART-P outside of merely(cid:3)1.0 exhibits 90.00% sensitivity and 94% specifi-
the MRI environment in atypical office setting. city. In comparison, an HDIS-Negative Affective Inter-
ference score merely(cid:3)1.0 exhibits 91% sensitivity and
91% specificity.
Results
InternalvalidityoftheHDIS ConvergentvalidityoftheHDIS
Table 2 presents the alpha coefficients obtained for the HDIS(cid:2)PositiveEmotionalityandHDIS(cid:2)HedonicDeficit
HDIS subscales separately for women with versus with- scores were significantly correlated with SHAPS scores:
out PTSD. Coefficient alphas for the HDIS-Positive r(cid:1)-.26, pB0.05, r(cid:1)(cid:6).36, pB0.05, respectively, in the
Emotionality, HDIS-Hedonic Deficit, and HDIS-Nega- PTSD group. However, HDIS-Negative Affective Inter-
tive Affective Interference subscales were high in the ference scores were not significantly correlated with
PTSD group. Coefficient alphas for the HDIS-Positive SHAPS scores, r(cid:1)(cid:6).14, ns. FCPCS scores were sig-
EmotionalityandHDIS-HedonicDeficitscaleswerealso nificantly correlated with each of the HDIS subscales:
high in healthy women, but low for the HDIS-Negative HDIS-Positive Emotionality, r(cid:1)(cid:5).52, pB0.01, HDIS-
Affective Interference scale; the latter may indicate a Hedonic Deficit, r(cid:1)(cid:6).44, pB0.01, and HDIS-Negative
multi-factorial structure to the item content in healthy Affective Interference, r(cid:1)(cid:6).28, pB0.05.
women. In women with PTSD, correlations between
scores on the three HDIS subscales were generally
IncrementalvalidityoftheHDISinthepredictionof
moderate or lesser in magnitude, agreeing with their
PTSDsymptoms
proposed discriminability: HDIS-Positive Emotionality
Although SHAPS and FCPCS scores significantly differ
with HDIS-Hedonic Deficit, r(cid:1)(cid:5).42, p(cid:1)0.001; HDIS-
betweenthePTSDgroupandcontrols,withinthePTSD
Positive Emotionality with HDIS-Negative Affective
group SHAPS and FCPCS scores were not significantly
Interference, r(cid:1)(cid:5).20, p(cid:1)0.07; HDIS-Hedonic Deficit
correlated with PTSD symptom severity as indexed by
with HDIS-Negative Affective Interference, r(cid:1)(cid:6).67,
CAPS total scores: r(cid:1).05, p(cid:1).38, and r(cid:1)(cid:5).01, p(cid:1)
pB0.001.
0.48, respectively. In contrast, Table 3 indicates that
HDIS-Hedonic Deficit and HDIS-Negative Affective
Groupdifferencesinanhedoniabetweenwomenwith Interference scores correlated not only with CAPS total
vs.withoutPTSD scores but also with both CAPS PTSD cluster C
The PTSD group scored significantly higher on the (Avoidance, Numbing) and D (Hyperarousal) scores,
FCPCS (M(cid:1)156.40, SD(cid:1)25.14, vs. M(cid:1)121.00, SD(cid:1) but not CAPS PTSD cluster B (re-experiencing) scores.
11.92, t[61](cid:1)7.49, pB0.001, d(cid:1)1.41) and the SHAPS A multiple regression analysis with HDIS-Hedonic
(M(cid:1)3.89, SD(cid:1)3.37, vs. M(cid:1)0.48, SD(cid:1)1.45, t[61](cid:1) Deficit and HDIS-Negative Affective Interference en-
5.52,pB0.001,d(cid:1)1.01).Table2alsoreportsdescriptive tered in step 1 as predictors of CAPS total scores was
statistics for the HDIS subscales, separately for women statisticallysignificant,R2(cid:1)0.16,F(2,34)(cid:1)3.23,p(cid:1)0.05,
with vs. without PTSD, as well as the results of group whereas the addition of SHAPS and FCPCS scores
Table2. Internal validity& descriptivestatistics forthe hedonicdeficit&interferencescales
Controlgroup(n(cid:1)35) PTSD(n(cid:1)55)
a M SD a M SD t(61) d
HDIS-PE .84 7.38 1.54 .85 3.32 2.09 10.01 1.91
HDIS-HD .85 0.41 1.11 .89 5.16 2.60 10.52 1.88
HDIS-NAI .56 0.41 0.44 .93 4.21 2.33 9.45 1.65
Note: All between-group differences have p’sB0.001. a(cid:1)coefficient alpha; HDIS, hedonic deficit & interference scale; PE, positive
emotionalitysubscale;AD,hedonicdeficitsubscale;NAI,negativeaffectiveinterferencesubscale.
6
Citation:EuropeanJournalofPsychotraumatology2012,3:8587-DOI:10.3402/ejpt.v3i0.8587
(pagenumbernotforcitationpurpose)
Assessmentofanhedoniainpsychologicaltrauma
Table3. Correlation betweenHDISsubscalesand PTSD, ^R2(cid:1)0.17, F(2,32)(cid:1)3.34, pB0.05. These findings sup-
depression, andanxiety symptomsaswellaschildhood port the incremental validityof the HDIS in accounting
trauma history for variance in PTSD symptom severity within a PTSD
sample relative to standard measures of anhedonia.
HDIS-PE HDIS-HD HDIS-NAI
AssociationsbetweentheHDISandchildhood
traumahistory
PTSD(CAPS)
Table 3 shows that HDIS scores were predicted by
Total (cid:5).42 .32 .42
severity of childhood trauma history as measured by
B(Re-experiencing) (cid:5).36 .06 .10
the Childhood Trauma Questionnaire. All associations
C(Avoidance-numbing) (cid:5).36 .28 .50
were statistically-significant and moderate in magnitude.
D(Hyperarousal) (cid:5).18 .36 .28
Depression,anxiety,stress
Concurrentcriterion-related&discriminantvalidityof
scale(DASS-21)
theHDIS:mono-method(questionnaires)
Depression (cid:5).51 .44 .42
Table 3 also reports correlations between the HDIS and
Anxiety (cid:5).01 .20 .30
theDASS-21.Aspredicted,HDIS-PositiveEmotionality
Childhoodtrauma(CTQ) scoresweremorestronglycorrelatedwithDASS-Depres-
Emotionalabuse (cid:5).58 .60 .60 sionthanwithDASS-Anxiety,Z(cid:1)3.49,pB0.001,aswas
Emotionalneglect (cid:5).56 .61 .60 the case for HDIS-Hedonic Deficit scores, Z(cid:1)1.71, pB
Physicalabuse (cid:5).39 .52 .46 0.05. These findings support the discriminant validityof
Physicalneglect (cid:5).53 .60 .58 the HDIS scales (cf. Clark & Watson, 1991). However,
Sexualabuse (cid:5).44 .58 .52 HDIS-Negative Affective Interference scores were not
significantly more strongly correlated with DASS-De-
Note: For CAPS & DASS-21, n(cid:1)55 (PTSD sample only). For pression than with DASS-Anxiety, Z(cid:1)0.87, ns. This
CTQ, n(cid:1)90 (full sample). r ] .26 corresponds to p 5 .05. finding is consistent with the discriminant validity of
HDIS(cid:1)HedonicDeficit&InterferenceScale,PE(cid:1)PositiveEmo- the HDIS-Negative Affective Interference subscale rela-
tionality Subscale, AD(cid:1)Hedonic deficit Subscale, NAI(cid:1)Nega- tive to the HDIS-Hedonic Deficit subscale, and suggests
tiveAffectiveInterferenceSubscale.
negativeaffectiveresponsestopositiveeventsmaycovary
with anxiety symptoms.
failed to significantly improve prediction, ^R2(cid:1)0.02,
F(2,32)(cid:1)0.30, p(cid:1)0.74. In comparison, a multiple re-
Concurrentcriterion-relatedvalidityoftheHDIS:
gression analysis using SHAPS and FCPCS as scores in multi-method(ART-Pself-report)
step 1 was not statistically significant, R2(cid:1)B .01, Table 4 presents Pearson correlation coefficients bet-
F(2,34)(cid:1)0.05, p(cid:1)0.95; however, the addition of HDIS- ween positive and negative emotional responses to the
Hedonic Deficit and HDIS-Negative Affective Interfer- ART-P Emotion scripts and each of the HDIS sub-
ence scores in step 2 significantly improved prediction, scales for women with PTSD. All associations were in
Table4. Multi-method concurrent criterion-relatedvalidityofthe hedonicdeficit &interference scaleswithself-report
responses tosocial positiveandnonsocial positiveemotionscripts ofthe affectiveresponse test
PA Avoid-PA Anger Anxiety Fear Disgust Sad Numb Shame Avoid-NA
Social
HDIS-PE .25 (cid:5).18 (cid:5).16 (cid:5).32 (cid:5).30 (cid:5).19 (cid:5).21 .08 (cid:5).40 (cid:5).08
HDIS-HD (cid:5).46 .37 .29 .43 .30 .38 .22 .30 .30 .25
HDIS-NAI (cid:5).52 .35 .38 .47 .38 .55 .37 .33 .41 .38
Non-Social
HDIS-PE .20 (cid:5).38 (cid:5).29 (cid:5).43 (cid:5).23 (cid:5).35 (cid:5).19 .03 (cid:5).38 (cid:5).23
HDIS-HD (cid:5).13 .21 .00 .27 .24 .26 .06 .09 .18 .37
HDIS-NAI (cid:5).25 .27 .09 .29 .24 .27 .16 .12 .33 .33
Note:r].26correspondstop50.05.ResultsfromparticipantswhocompletedthetaskoutsideofthefMRIscanner.PA.positiveaffect;
Avoid-PA,attemptedavoidanceofpositiveaffect;Anx,Anxiety;Avoid-NA,attemptedavoidanceofnegativeaffect.HDIS,hedonicdeficit
&interferencescale;PE,positiveemotionalitysubscale;HD,hedonicdeficitsubscale;NAI,negativeaffectiveinterferencesubscale.
7
Citation:EuropeanJournalofPsychotraumatology2012,3:8587-DOI:10.3402/ejpt.v3i0.8587
(pagenumbernotforcitationpurpose)
PaulA.Frewenetal.
the predicted direction, and many were statistically- responserelativetobaselinewithinboththeleftandright
significant. The magnitude of correlations between each OFC during social positive imagery (Fig. 1-A), and
oftheHDIS-HedonicDeficitandHDIS-NegativeAffec- within the left OFC only during non-social positive
tive Interference scores with the ART-P measures were imagery (Fig. 2-A).Regarding response to social positive
not significantly different, although in nearly all cases imagery, HDIS-Positive Emotionality scores also posi-
correlationswere stronger with Negative Affective Inter- tively predicted responsewithin medial prefrontal cortex
ference scores. Associations for both measures were (including the anterior cingulate), cerebellum, and
somewhat stronger in response to imagery of positive occipital cortex (Fig. 1-B), and negatively predicted
events that were explicitly social as comparedwith non- response within right middle temporal cortex. In com-
social events, irrespective of the type of negative emo- parison, regarding response to non-social positive ima-
tional response (i.e., anger, anxiety, fear, disgust, numb- gery, HDIS-Positive Emotionality scores positively
ing, sadness, or shame). predicted response within the right insula (Fig. 2-B),
and negatively predicted response within bilateral
Concurrentcriterion-relatedvalidityoftheHDIS: precuneus, and right superior parietal and middle temp-
multi-method(ART-PfMRI-BOLDresponse) oral cortex.
Please see Table 5 and Figs. 1 and 2 for SPM2 results. By contrast, HDIS-Hedonic Deficit scores did not
Consistent with hypotheses, within women with PTSD, correlate significantly with response to social positive
HDIS-Positive Emotionality scores positively predicted events.However,HDIS-HedonicDeficitscorespositively
Table5. Multi-method concurrent criterion-relatedvalidityofthe hedonicdeficit &interferencescaleswithself-report
responses to positiveemotion scriptsof the affectiveresponse test
Correlation ROI MNI k Z p
Social
HDIS-PE (cid:6) R-Orbitofrontalcortex 26,32,(cid:5)18 106 3.52 B0.001
(cid:6) L-Orbitofrontalcortex (cid:5)30,34,(cid:5)22 75 3.24 0.001
(cid:6) L-Orbitofrontalcortex (cid:5)30,34,(cid:5)22 54 3.14 0.001
(cid:6) ROccipitalcortex 16,(cid:5)106,4 49 3.33 B0.001
(cid:6) L-Cerebellum(posteriorlobe) (cid:5)18,(cid:5)60,(cid:5)36 327 3.25 0.001
(cid:6) LCerebellum(posteriorlobe) (cid:5)28,(cid:5)80,(cid:5)44 61 3.12 0.001
(cid:6) L-Medialprefrontalcortex (cid:5)4,38,8 144 2.91 0.003
(cid:5) R-Middletemporalcortex 46,(cid:5)34,(cid:5)22 49 3.90 B0.001
HDIS-HD (cid:6) (nosignificantresults) (cid:5) (cid:5) (cid:5) (cid:5)
(cid:5) (nosignificantresults) (cid:5) (cid:5) (cid:5) (cid:5)
HDIS-NAI (cid:6) (nosignificantresults) (cid:5) (cid:5) (cid:5) (cid:5)
(cid:5) RCerebellum(posteriorlobe) 40,(cid:5)58,(cid:5)22 113 3.68 B0.001
(cid:5) RTemporal-parietaljunction 56,(cid:5)52,12 145 3.62 B0.001
(cid:5) LCerebellum(posteriorlobe) (cid:5)34,(cid:5)58,(cid:5)22 101 3.48 B0.001
(cid:5) RCerebellum(posteriorlobe) 24,(cid:5)48,(cid:5)48 132 3.41 B0.001
(cid:5) RMiddletemporalgyrus 48,(cid:5)52,2 64 2.97 0.001
Non-Social
HDIS-PE (cid:6) L-Orbitofrontalcortex (cid:5)24,40,(cid:5)14 44 3.79 B0.001
(cid:6) RInsula 52,12,20 183 3.21 0.001
(cid:5) R-Superiorparietalcortex 46,(cid:5)74,18 55 3.76 B0.001
(cid:5) R-Middletemporalcortex 60,(cid:5)32,6 47 3.42 B0.001
(cid:5) R-Precuneus 10,(cid:5)72,42 33 3.25 0.001
(cid:5) L-Precuneus (cid:5)10,(cid:5)74,48 57 3.12 0.001
HDIS-HD (cid:6) LPre-cuneus (cid:5)8,(cid:5)78,42 83 3.45 B0.001
(cid:6) RCerebellum(anteriorlobe,vermis) 6,(cid:5)50,0 127 3.33 B0.001
(cid:5) (nosignificantresults) (cid:5) (cid:5) (cid:5) (cid:5)
HDIS-NAI (cid:6) RCerebellum(posteriorlobe) 28,(cid:5)82,(cid:5)42 42 3.43 B0.001
(cid:6) RMiddlefrontalgyrus 44,16,18 90 3.34 B0.001
(cid:6) LAmygdala (cid:5)18,(cid:5)4,(cid:5)12 15 3.04 0.001
(cid:5) (nosignificantresults) (cid:5) (cid:5) (cid:5) (cid:5)
8
Citation:EuropeanJournalofPsychotraumatology2012,3:8587-DOI:10.3402/ejpt.v3i0.8587
(pagenumbernotforcitationpurpose)
Assessmentofanhedoniainpsychologicaltrauma
Negative Affective
Positive Emotionality: Hedonic Deficit: Interference:
Positive Correlation Non-significant Positive Correlation
A C
D
B
P < .005 P > .005 P < .005
Fig. 1. BOLDResponse to Imageryof SocialPositiveEvents as Predicted byHDISScores.
Source:A:y(cid:1)34,Orbitofrontalcortex(crosshairs:lefthemisphere);B:x(cid:1)(cid:5)10,MedialPrefrontalCortex(cross-hairs)
and Cerebellum; C: x(cid:1)56, Right Temporoparietal Junction, D: z(cid:1)(cid:5)22, Cerebellum (crosshairs: left lobule).
predictedresponsewithinthecerebellarvermis(Fig.2-C/ Discussion
D) and precuneus during non-social positive imagery. This study identified symptoms of anhedonia in women
Finally, HDIS-Negative Affective Interference scores with PTSD using standard measures (SHAPS, FCPCS),
negatively predicted response to social positive imagery extendingtheresultsofpreviousstudies(Kashdanetal.,
within the right temporoparietal junction (Fig. 1-C), 2006, 2007). We have suggested, however, that one can
bilateralcerebellum(posteriorlobes,Fig.1-D),andright distinguishbetweenanhedonicsub-processes,specifically
middle temporal gyrus. In contrast, HDIS-Negative those relating to the perceived incapacity to experience
AffectiveInterferencescorespositivelypredictedresponse positive affect in response to positive stimuli and events
within the left amygdala (Fig. 2-E), right cerebellum (hedonic deficits) in contrast to the tendency to experi-
(posteriorlobe,Fig.2-F),andrightmiddlefrontalgyrus. ence interfering negative affect in response to positive
Negative Affective
Positive Emotionality: Hedonic Deficit:
Interference:
Positive Correlation Positive Correlation
Negative Correlation
A C E
B D F
P < .005 P < .005 P < .005
Fig. 2. BOLDResponse to Imageryof Non-social PositiveEventsas Predicted byHDISScores.
Source:A:y(cid:1)40,LeftOrbitofrontalcortex;B:y(cid:1)6,RightInsula;C&D:Cerebellum(C:x(cid:1)6,D:y(cid:1)(cid:5)50),E:y(cid:1)(cid:5)
6, Left Amygdala, F: z(cid:1)(cid:5)42, Cerebellum (crosshairs(cid:1)right lobule).
9
Citation:EuropeanJournalofPsychotraumatology2012,3:8587-DOI:10.3402/ejpt.v3i0.8587
(pagenumbernotforcitationpurpose)
PaulA.Frewenetal.
stimuliandevents(negativeaffectiveinterference).Inthe thattheemotionalresponsetaskusedinthepresentstudy
presentstudyweprovideevidenceofeachsub-processof relied too heavily on concentration and imagery which
anhedonia in women with PTSD via their responses to maybeproblematicprovidingthatmanyindividualswith
the Hedonic Deficit and Interference Scale (HDIS; PTSD are alexithymic (e.g., review by Frewen, Dozois,
Frewen, Dean, & Lanius, 2012). Lanius,Neufeld,&Lanius,2008)andthereforemayhave
TheHDISdemonstratedexcellentinternalconsistency difficulties with imagery tasks. In fact, Sifneos (1987)
in women with PTSD, convergence with previously speculated that whereas all anhedonic individuals may
validated anhedonia scales (SHAPS, FCPCS), incremen- not be alexithymic, all alexithymic individuals are likely
tal validity in predicting PTSD severity (CAPS) scores anhedonic; a study of healthy participants, however,
beyondtheSHAPSandFCPCS,andexcellentsensitivity found that symptoms of anhedonia and alexithymia
and specificity for the diagnosis of PTSD (relative to load on distinct factors (Loas, Fremaux, & Boyer,
psychological health) with very minimal scores. As 1997). In our first study of the HDIS (Frewen, Dean &
predicted, the HDIS-Positive Emotionality and HDIS- Lanius, 2012), we evaluated a simpler methodology: we
Hedonic Deficit subscales were more strongly correlated modified the FCPCS to ask about negative affective
with depressive symptoms than anxious hyperarousal, responsestopositiveevents,andfoundthatHDISscores
supportingtheconcurrentcriterion-related anddiscrimi- predict such responses. However, other methods for
nate validity of the HDIS within the context of psycho- assessing anhedonic responses to positive stimuli should
logical models that differentiate depression from anxiety also be tested (e.g., response to pleasant pictures;
symptomatology primarily in terms of anhedonia and Leventhal et al., 2006, and reward tasks; Elman et al.,
low positive affect (e.g., Clark & Watson, 1991). HDIS- 2005, 2009; Hopper et al., 2008). In addition, although
Negative Affective Interference scores, however, were the present study illustrates the applicability of differ-
significantlycorrelatedwith both depression and anxiety entiating between hedonic deficits and negative affective
symptoms, indicating the discriminant validity of asses- interference in PTSD, a disorder that is frequently
sing negative affective responses to positive events, out- associated with anhedonic symptomatology (e.g., Kash-
comes apparently also related to anxiety symptoms. dan et al., 2006, 2007), the present study did not assess
Interestingly,HDIS-HedonicDeficitandHDIS-Negative therelevanceoftheseconstructstoanhedonicsymptoms
Affective Interference correlated particularly with PTSD presentinotherpsychiatricpopulationswhereanhedonia
avoidance-numbingandhyperarousalsymptoms,butnot is more often studied, and which are variably also
with re-experiencing symptoms, suggesting that they are associated with trauma exposure (e.g., schizophrenia,
best understood as affective problems perhaps indepen- mood disorders). A study examining hedonic deficits
dent of a focus on intrusive memories. Additionally, and negative affective interference in individuals with
anhedonia symptoms may have their origin in early schizophreniaandmooddisorders,whovarywithrespect
learning providing that self-reported severity of child- to trauma exposure, in comparison with symptom
hood abuse predicted both increasing HDIS-Hedonic severity in PTSD, would be helpful in determining
Deficits and HDIS-Negative Affective Interference. whether trauma exposure plays an etiological role in
HDIS scores also predicted self-reported emotional symptoms ofanhedonia. Furthermore, the present study
responses during imagery of positive events, particularly was limited to women with the primary diagnosis of
thosewith an explicit interpersonal focus. Finally, HDIS PTSD who had considerable psychiatric comorbidity,
scores concurrently predicted subjective and functional and additional studies will be necessary to ascertain the
metabolicresponsestoimageryofprototypicallypositive generalizability of the present findings to men, and to
events within several regions of interest including the PTSD specifically versus other disorders with which
orbitofrontal cortex, medial prefrontal cortex, insula, PTSD is frequently comorbid. The internal consistency
amygdala, and cerebellum. Furthermore, the neural foundfortheHDIS-NegativeAffectiveInterferenceitems
correlatesofanhedonicsymptomsdifferedbetweensocial was also unacceptably low in healthy women, indicating
relative to non-social positive imagery. For example, either the items may not measure a valid construct in
response within both the medial prefrontal cortex and healthy women, or aremulti-factorial in healthy women.
the right temporoparietal junction, both known to be Itwillalsobeimportanttoassessthetemporalreliability
involved in social cognitive processing (e.g., review by of the HDIS in the future.
Van Overwalle, 2009), was predicted by anhedonic Anumberofopenquestionsremain,suchasregarding
symptoms during social but not during non-social the etiology and development of hedonic deficits and
positive imagery. Nevertheless, these findings must be negative affective interference, and the prognostic sig-
considered preliminarydueto small sample sizes andwe nificance of these measures for treatment. The present
highlight the need for replication. studysuggeststhatitmaybeusefultosupplementtheuse
Therearelimitationsofthepresentstudythatwillneed of traditional measures of anhedonia with measures of
tobeaddressedbyfurtherresearch.Oneconcernmaybe negative affective interference such as the HDIS. It will
10
Citation:EuropeanJournalofPsychotraumatology2012,3:8587-DOI:10.3402/ejpt.v3i0.8587
(pagenumbernotforcitationpurpose)