Neuroticism (Not Depressive Symptoms)Predicts Memory Complaints in SomeCommunity-Dwelling Older Adults

Matt R. Merema, B.A. (Hons.), Craig P. Speelman, Ph.D., Jonathan K. Foster, D.Phil.,Elizabeth A. Kaczmarek, Ph.D.

Received JanuarUniversity (MRMNeurosciences UB.A. (Hons.), Sch

� 2013 Amehttp://dx.d

Am J Geriatr Ps

Objectives: To examine whether depressive symptoms are useful predictors of subjec-

tive memory complaints in community-dwelling older adults, beyond the predictive

utility already provided by memory performance and characteristics of personality.

Design: Using hierarchical regression, we examined the relationship between

depressive symptoms and subjective memory complaints, controlling for age, gender,

education, memory performance, conscientiousness, and neuroticism. Participants:Community-dwelling older adults aged 66 to 90 years (N ¼ 177) who responded to

a newspaper advertisement for a memory study in Perth, Western Australia.

Measurements: The General Frequency of Forgetting scale (for memory complaints),

Depression Anxiety Stress Scales (for depressive symptoms), NEO-Five Factor Inven-

tory (for conscientiousness and neuroticism), and the Visual Reproduction and

Logical Memory subtests from the Wechsler Memory Scale—4th Edition (for visual

and verbal memory). Results: The hierarchical regression analysis indicated that

while depressive symptoms significantly predicted memory complaints after variance

associated with age, gender, education, memory performance, and conscientiousness

was partialled out, they accounted for almost none of the variance in complaints

when neuroticism was partialled out. Conclusions: The well-established relationship

between depression and memory complaints may exist in some community-dwelling

older adult populations only on account of the manner in which both are associated

with neuroticism. (Am J Geriatr Psychiatry 2013; 21:729e736)

Key Words: Memory complaints, depression, neuroticism, conscientiousness, personality

hile the nature of the relationship between

Wmemory complaints in older adulthood andobjective memory performance is still being exam-ined and clarified, it is now well-established thatdepression (or characteristics of it) is associated with

y 26, 2012; revised May 10, 2012; accepted June 19, 201, CPS, EAK), Joondalup, Australia; and School of Psychnit, Health Department of West Australia (JKF), Austraool of Psychology & Social Science, Edith Cowan Univerican Association for Geriatric Psychiatryoi.org/10.1016/j.jagp.2013.01.059

ychiatry 21:8, August 2013

memory complaints of greater number and/orseverity.1e6 Symptoms of anxiety, poorer physical andmental health, and less physical activity (all of whichare factors heavily associated with depression7e9) arealso associated with subjective memory complaints.5

2. From the School of Psychology & Social Science, Edith Cowanology & Speech Pathology, Curtin University, Bentley, Australia &lia. Send correspondence and reprint requests to Matt R. Merema,rsity, Joondalup 6027, Australia. e-mail: mmerema@our.ecu.edu.au

729

Depression and Memory Complaints

Indeed, depression is arguably one of the most con-sistent predictors of memory complaints identified todate.10

The scientific literature also suggests depression isa more useful predictor of memory complaints thandemographics, medical conditions (including stroke),memory performance, general cognitive functioning,andApolipoprotein E status.10,11 Depression, however,is not the only factor consistently found to be associatedwith memory complaints. Various studies have identi-fied a link between memory complaints and charac-teristics of personality.12e14 High levels of neuroticismand low levels of conscientiousness, for example, havebeen found to be associated with more frequent orsevere complaints regarding memory.13,14 Given theoverlap between depression and personality,15,16 suchassociations raise an interesting question about (i) therole of personality characteristics and (ii) the uniquecontributions of depressive symptoms, beyondpersonality factors, in mediating memory complaints.

Interestingly, previous studies have hinted at thepossibility that depression might contribute littleunique variance to the prediction of memory com-plaints beyond that already predicted by personality.Two studies have directly assessed personality char-acteristics (i.e., neuroticism and conscientiousness)and self-esteem as predictors of subjective memorycomplaints.13,14 Data from both studies show consid-erable decreases in the relationship between self-esteem (which is associated with depression17) andspecific memory complaints when neuroticism andconscientiousness are partialled out of the statisticalmodel. Research evaluating more general cognitivecomplaints has provided similar evidence. In a studyof 123 older adults with Aging-Associated CognitiveDecline (AACD) and 291 healthy older adults,neuroticism accounted for a significant amount ofvariance in cognitive complaints for both the AACDand healthy older adult groups.12 Furthermore, in theAACD group, neuroticism accounted for the mostvariance in cognitive complaints, while depressionwas not found to predict a significant amount ofunique statistical variance. Together, these findingshint at the possibility that depression may not alwaysbe directly associated with memory complaints in oldage but rather may sometimes be a surrogate forpersonality characteristics that can account well forboth levels of memory complaints and depressivesymptoms.

730

On the basis of these previous results together withresearch that suggests personality characteristicsplay a considerable role in depression vulnerability,15

the current study examined whether the associationbetween depressive symptoms and memory com-plaints in community-dwelling older adults existsbeyond personality characteristics, or whether itmay constitute a cum hoc, ergo propter hoc fallacy18

(whereby depression and memory complaints cor-relate only because both are shaped by personalitycharacteristics). To achieve this goal, the currentstudy employed a correlational design to examinethe relationship between depressive symptomsand memory complaints, both alone and in thecontext of characteristics of personality (i.e., neu-roticism and conscientiousness). On the basis ofprevious literature,13,14 we hypothesized that therelationship between depressive symptoms andmemory complaints would diminish considerablyafter partialling out statistical variance associatedwith characteristics of personality.

METHODS

Participants

The sample comprised 177 English-speaking olderadults (115 women, 62 men) aged between 66 and 90years (M ¼ 73.62, SD ¼ 6.33) residing within thePerth metropolitan area of Western Australia. Yearsof formal primary, secondary, and postsecondaryeducation ranged from 8 to 24 years (M¼ 12.56, SD ¼2.86). History of vision or hearing impairments,medications, previous head injuries, and stroke wererecorded, although participants were only excludedfrom the current study when hearing or visionimpairments did not allow for a valid assessment tobe conducted. All participants were recruited via anadvertisement in local newspapers promotinga study on memory. For the purposes of reviews andmeta-analyses that may be undertaken, it should benoted that data from 121 of these participants wereused in another recently published study.19

Materials

Subjective memory complaints were assessed usingthe General Frequency of Forgetting (GFF) scale fromthe Memory Functioning Questionnaire.20 The GFF

Am J Geriatr Psychiatry 21:8, August 2013

Merema et al.

scale is a 33-item, 7-point Likert scale designed tomeasure self-reports of memory problems acrossvarious tasks, contexts, and time frames. Objectivememory performance was assessed using the visualreproduction (VR) and logical memory (LM) subtestsfrom the Wechsler Memory Scale—4th Edition (WMS-IV).21 The subtests assess visual and verbal remem-bering, respectively; each providing a measure ofimmediate recall (VR-I and LM-I), delayed recall (VR-IIand LM-II), and recognition memory (VR-R and LM-R). Depressive symptoms were measured using theDepression Scale from the Depression Anxiety StressScales (DASS),22 a 14-item, 4-point Likert scale (rangingfrom “Did not apply to me at all” through to “Appliedto me very much, or most of the time”) designed tomeasure symptoms of self-disparagement, pessimism,life dissatisfaction, and inertia over the previous week.The Depression Scale of the DASS was developed foruse in Australia and has been used in both clinical andcommunity contexts as a measure of depression.23e25

It demonstrates acceptable psychometric properties26

and clearly delineates between individuals withand without Major Depressive Disorder.27 Neuroti-cism and conscientiousness were measured usingthe NEO-Five Factor Inventory (NEO-FFI).28 Bothscales are composed of 12 items and are scored ona 5-point Likert scale, ranging from “strongly disagree”to “strongly agree.”

Procedure

Once informed consent was obtained, participantsprovided details regarding age, education, vision andhearing difficulties, medication use, and history of anyhead injuries and stroke. The first component of eachof the WMS-IV subtests (VR-I and LM-I) was thencompleted, after which 25 minutes was spent filling inthe three questionnaires (Memory Functioning Ques-tionnaire, NEO-FFI, and DASS). Participants typicallydid not complete all three questionnaires within theallocated 25-minute period, in which case they weremomentarily put aside. The second components of theWMS-IV subtests (VR-II, VR-R, LM-II, and LM-R)were completed next, followed by any unfinishedsections of the questionnaires.

Analysis

Multiple regressions were used to examine theunique contribution of depressive symptoms as

Am J Geriatr Psychiatry 21:8, August 2013

predictors of memory complaints beyond personality.However, given strong relationships between predictorvariables in a multiple regression analysis can distortregression coefficients and that (based on previousliterature21) we expected reasonably strong intercorre-lations between the measures of memory, a prelimi-nary factor analysis was carried out to determinewhether a smaller number of latent variables couldadequately account for memory performance. Rawscores from the immediate, delayed, and recognitionmeasures of visual and verbal memory were subjectedto Principal Axis factoring and a Direct Oblimin rota-tion. A nonorthogonal rotation was chosen because weexpected factors to be correlated to some degree,regardless of the underlying structure. Factor scoresfrom each of the factors derived were then utilized asmeasures of memory in the multiple regressionanalysis.

A three-stage hierarchical regression analysis wasthen carried out to examine the predictive utility ofdepressive symptoms beyond personality, while alsocontrolling for memory performance and demo-graphics. Depression Scale scores, memory factorscores, age, gender, and years of education wereentered into the model first (model 1) to examine thepredictive utility of depressive symptoms beyonddemographics and memory performance but irre-spective of personality. Previous literature suggestsconscientiousness might be marginally more stronglyrelated to memory complaints than neuroticism,13,14

so conscientiousness scores were entered into themodel next (model 2). Neuroticism scores wereentered into the model last (model 3).

RESULTS

Preliminary Factor Analysis

The Kaiser-Meyer-Olkin measure suggested goodto very good sampling adequacy (0.78, individuallyall six measures of memory exceeded 0.72), indicatingfactor analysis was appropriate. Bartlett’s test ofsphericity also suggested sufficiently large correla-tions between the measures, c2[15] ¼ 560.82, p <0.001.Two factors with eigenvalues exceeding 1 wereidentified, which individually explained 57.3% and20.5% of the variance in memory performance scores.The 3 measures of verbal memory loaded heavily on

731

Depression and Memory Complaints

the first factor; namely, LM-I (0.84), LM-II (0.98), andLM-R (0.65). The measures of visual memory loadedheavily on the second factor, that is; VR-I (0.89), VR-II(0.87), and VR-R (0.68). All other loadings rangedbetween �0.07 and 0.05, suggesting that a singlemeasure of memory was not appropriate, and thatseparate measures of visual and verbal memoryshould be utilized. The correlation of 0.545 observedbetween the two factors was not strong enough towarrant concern for multicollinearity. Zero-ordercorrelations between memory complaints, factorscores on the visual and verbal memory factors,depressive symptoms, characteristics of personality,and demographic variables are provided in Table 1.

Multiple Regression Analysis

Overall, the multiple regression analysis indicatedthat model 1 accounted for 14.1% of the variancein GFF scores, F[6, 170] ¼ 4.649, p <0.001. Verbalmemory factor scores (partial r ¼ 0.240, p ¼ 0.002)and Depression Scale scores (partial r ¼ �0.207,p ¼ 0.006) accounted for a significant amount of vari-ance in GFF scores, although visual memory factorscores (partial r ¼ �0.051, p ¼ 0.510), age (partial r ¼0.036, p¼ 0.640), years of education (partial r¼�0.075,p¼ 0.326), and gender (partial r¼ 0.144, p¼ 0.059) didnot. After entering conscientiousness, model 2accounted for an additional 8.0% (p <0.001) of thevariance in GFF scores and a total of 22.1% of thevariance, F[7, 169]¼ 6.864, p<0.001. Conscientiousnessscores (partial r ¼ 0.306, p <0.001), verbal memoryfactor scores (partial r ¼ 0.266, p <0.001), and

TABLE 1. Zero-order Correlations Among Study Variables

Measure 1 2 3 4

1. GFF —

2. Dep �0.220a —

3. Consc 0.336b �0.129 —

4. Neurot �0.375b 0.549b �0.315b —

5. VisMem 0.117 �0.157c 0.117 �0.0946. VerMem 0.279b �0.152c 0.047 �0.117 07. Age �0.037 �0.058 �0.009 �0.098 �08. Edu �0.022 �0.072 �0.080 �0.163c 09. Gender 0.156c 0.125 0.186c 0.146 0

Notes: Reported values are Pearson correlations (df ¼ 175). Gender isForgetting scale; Dep: depressive symptoms; Consc: conscientiousness; Nverbal memory factor scores; Edu: education.

ap <0.01, two-tailed.bp <0.001, two-tailed.cp <0.05, two-tailed.

732

Depression Scale scores (partial r¼�0.173, p¼ 0.024)accounted for a significant amount of variance in GFFscores, although visual memory factor scores (partialr¼�0.090, p¼ 0.240), age (partial r¼ 0.021, p¼ 0.790),education (partial r ¼ �0.053, p ¼ 0.488), and gender(partial r ¼ 0.084, p ¼ 0.273) did not.After entering neuroticism, model 3 accounted for

an additional 5.7% (p <0.001) of the variancein GFF scores and a total of 27.8% of the variance,F[8, 168] ¼ 8.092, p <0.001. Neuroticism scores (partialr ¼ �0.270, p <0.001), verbal memory factor scores(partial r ¼ 0.256, p ¼ 0.001), and conscientiousnessscores (partial r ¼ 0.214, p ¼ 0.005) accounted fora significant amount of variance in GFF scores,although Depression Scale scores (partial r ¼ �0.019,p ¼ 0.806, 95% CI [�0.166, 0.129]) no longeraccounted for a significant amount of variancewith neuroticism scores included in the model.Likewise, visual memory factor scores (partialr ¼ �0.095, p ¼ 0.216), age (partial r ¼ �0.009,p ¼ 0.912), education (partial r ¼ �0.103, p ¼ 0.182),and gender (partial r ¼ 0.133, p ¼ 0.083) also failed toaccount for a significant amount of variance in GFFscores in model 3.

DISCUSSION

The current study examined whether the associa-tion between depressive symptoms and memorycomplaints in community-dwelling older adultsexists beyond personality characteristics, or whether

5 6 7 8 M SD

160.00 26.024.95 6.95

34.00 5.6719.10 7.38

— 0.00 0.94.590b — 0.00 0.96.600b �0.423b — 73.62 6.33.066 0.161c �0.048 — 12.56 2.86.065 0.161c �0.075 �0.048 — —

coded as male ¼ 1 and female ¼ 2. GFF: General Frequency ofeurot: neuroticism; VisMem: visual memory factor scores; VerMem:

Am J Geriatr Psychiatry 21:8, August 2013

Merema et al.

depression and memory complaints correlate onlybecause both are shaped by personality characteris-tics. As hypothesized, the relationship betweendepressive symptoms and memory complaintsdiminished after variance associated with personalitycharacteristics (predominantly neuroticism) waspartialled out. Depressive symptoms significantlypredicted memory complaints when memoryperformance, conscientiousness, age, education, andgender were taken into account. However, the vari-ance uniquely associated with depressive symptomstended toward zero after neuroticism was alsoincluded in the model. Based on the assumption thatmore enduring personality characteristics can influ-ence depression vulnerability,15 these results providesupport for the notion that the relationship betweendepression and memory complaints in community-dwelling older adults may not be a direct one butmay instead constitute an example of the cum hoc,ergo propter hoc fallacy, induced by the manner inwhich both depression and memory complaints arerelated to neuroticism.

In the current study, we refer tomemory complaintsas the subjective reports of problems independent ofobjective memory performance. While depressivesymptoms may not show a direct relationship withsuch complaints beyond neuroticism, depressivesymptoms exhibited a weak negative association withverbal (and visual)memory performance (see Table 1),which in-turn was weakly associated with complaintsreported on theGFF scale. Thus,while depressionmaynot always contribute to complaints directly, it maycontribute indirectly via its association with memoryperformance. However, we did not test for this in thecurrent study because (despite a weak negative asso-ciation between depressive symptoms and memoryperformance) the DASS primarily assesses mood-related symptoms of depression and so would notadequately capture the changes in brain structure indepression that largely underlie associated memorydysfunction (e.g., hippocampal changes29).

On a clinical level, the results of the current studyhave two important implications for mental health inolder adults. Firstly, they have implications formental health screening, given previous literaturesuggests that older adults complaining of memoryproblems should be assessed for underlying depres-sion as well as memory dysfunction.11 We do notargue that screening for depression is not warranted

Am J Geriatr Psychiatry 21:8, August 2013

in older adults complaining of memory problems;depression is unquestionably related to memorycomplaints. However, given memory complaintsmay materialize primarily through more enduringfeatures of personality in some circumstances, olderadults complaining of memory problems may exhibitan ongoing risk for depression (given the relationshipbetween neuroticism and risk of depression onset),even if depression is not present at the time ofassessment. This could be investigated longitudinallyby observing empirically whether incidence ofdepression at follow-up is associated with memorycomplaints at baseline in a group of older adults withno previous history of depression, with our currentfindings suggesting that such incidence may beinfluenced by an individual’s level of neuroticism.

A second implication of the results presented hereis that they may help to further explain why memoryand/or cognitive complaints (independent of objec-tive cognitive performance) have been reported asbeing predictive of subsequent dementia.10 In addi-tion to the relationship between concurrent depres-sion and memory complaints already raised in theliterature, the results presented here indicate thatolder adults with memory complaints likely exhibitan increased risk of subsequent depression, even ifdepressive symptoms are not present when thememory complaint is made. In combination with theincreased risk of concurrent depression, greatervulnerability to future depression may further help toexplain the higher rates of subsequent dementia,given recent research suggesting that a diagnosis ofdepression may subsequently double the risk fora dementia diagnosis in some contexts.30

While some research14 suggests self-esteem (whichis inversely associated with depression) does notuniquely account for a significant amount of varianceinmemory complaints beyondneuroticism, it has beenfound to account for a significant amount of uniquevariance in an earlier study.13 Likewise, while otherresearch has indicated that depression does notaccount for a significant amount of unique variance incognitive complaints in an AACD group,12 the samestudy reported that it did account for a significantamount of unique variance in memory complaints ina healthy control group (more so, even, than neuroti-cism). This result was replicated in another study,31

although presumably the results are based, in part,on the same data (from the Interdisciplinary Study on

733

Depression and Memory Complaints

Adult Development). Furthermore, several longitu-dinal studies have demonstrated that successfultreatment of depression is associatedwith decreases inmemory complaints.32,33

We suspect the discrepancy between studies thatdo and do not find depression (or associated features)to be an important predictor of age-related memory(or cognitive) complaints might be largely attribut-able to sample characteristics. For instance, neuroti-cism appears to play more of a role in determiningmemory complaints than depression in individualsexperiencing cognitive decline, which may reflectgreater levels of worry in study volunteers at anincreased risk of dementia (given their generallypoorer memory performance, relative to the pop-ulation at large).12 Neuroticism scores were notabnormally high in the current study (mean neurot-icism scores were on the 50th and 49th percentiles formen and women, respectively). However, given thatwe recruited through newspaper advertisements forthis study, it is unlikely that the sample includedmany individuals with symptoms of depression thatwere severe enough to warrant a formal diagnosis ofa Major Depressive Episode.

Lower depression severity in our sample maytherefore account for why our results differed fromthose of several previous studies,12,13,31 suggestingthat depression is a useful predictor of memorycomplaints in addition to the predictive utilityoffered by personality characteristics. These previousstudies pursued participants more actively than wedid for the current study, and therefore we may haverecruited fewer and/or less depressed individuals.We do not regard this as a limitation to the currentstudy, however, as we argue that the participantsample used here represents a sample of community-dwelling older adults who may be likely to present tohealth professionals or memory clinics with subjec-tive memory concerns. Therefore, while weacknowledge that depression may be a usefulpredictor of memory complaints in some samples ofolder adults, memory complaints may be relatedprimarily to neuroticism in older adults who moreproactively seek to address their subjective memoryproblems. This issue deserves further attention, andcould be examined by evaluating depression andneuroticism scores as predictors of memorycomplaints in groups that do and do not actively seekhelp for their subjective memory problems. Indeed, it

734

may be the case that depression increases memorycomplaints but decreases the likelihood of thesecomplaints being reported to a health professional.

While zero-order and partial correlations betweendepressive symptoms and memory complaintsobserved in the present study were weaker than thecorresponding correlations between personality char-acteristics and memory complaints, additionalconsideration needs to be given to this comparison.Depression Scale scores in our own data showeda considerable positive skew (see Table 1; note thatDepression Scale scores on the DASS can range from0 to 42), which is attributable to the manner in whichsymptoms of depression are characterized in theDASS. Although depression severity ultimately variesalong a continuum, a large proportion of people incommunity samples typically report few symptoms ofdepression or no symptoms at all (this proportionmaybe even higher in samples where individuals arerequired to actively seek out the researchers to partic-ipate, such as was the case in the current study). Weattempted to correct for this via various trans-formations, converting our data to ranks and groupingindividuals into depressed and nondepressed groupsbased on the median score and on clinical cutoffsprovided in the DASS manual.22 However, theseprocedures all returned weaker correlations fordepression symptoms with memory complaints thanwas the case for the original raw scores.

Skewed data for measures of depression symptomsare evident in a number of studies,4,34 with thesedistributions likely impacting on the strength ofcorrelations with memory complaints as well as withother variables. In contrast, measures of personalitycharacteristics are less likely to return skewed data incommunity samples, because personality measuresare typically designed for use in nonclinical pop-ulations. Correlations between depression scores,personalitymeasures, andmemory complaints shouldtherefore be interpretedwith this inmindand carefullyconsidered when making comparisons regarding thepredictive validity of memory complaints with char-acteristics of personality versus symptoms of depres-sion. One may be able to avoid this potential problemwith community samples in future by using wider-ranging measures of mood symptoms that vary from“extreme negative” to “extreme positive,” rather thanby applying measures that simply capture negativefeatures of mood that are associated with depression.

Am J Geriatr Psychiatry 21:8, August 2013

Merema et al.

A limitation of the current study is that, while thereis evidence to suggest that characteristics of person-ality influence depression vulnerability,15 we cannotrule out the possibility that, conversely, depressivesymptoms had an influence on reporting of thepersonality measures. That is, although depressiveepisodes may play little role in determining long-termaspects of personality, it is possible that participantsin the current study reported higher levels ofneuroticism because they were experiencing symp-toms of depression.35 Also, although the DASS clearlydifferentiates between individuals with and withouta formal diagnosis of depression,27 it is unlikely thatmany individuals with more severe depressionparticipated in the current study, given our recruit-ment method. Consequently, the possibility ofgeneralizing these results to a diagnostic contextlikely requires further consideration and investiga-tion. Finally, it should also be noted that, in clinicaltesting, the Verbal Paired-Associates subtest of theWMS-IV is usually administered between the imme-diate and delayed components of the VR and LMsubtests, while this was replaced with questionnaires(Memory Functioning Questionnaire, NEO-FFI, and

Am J Geriatr Psychiatry 21:8, August 2013

DASS) in the current study. However, we expect thischange in procedure had little effect on the key rela-tionships of interest reported here.

In summary, the results of the present studysuggest that, outside of the predictive utilitycontributed by neuroticism, depressive symptomsmay contribute little as unique predictors of memorycomplaints in some community-dwelling older adultpopulations. In older adults who actively seek outhelp for subjective memory problems, the “well-established” relationship between depression andmemory complaints may instead represent a cum hoc,ergo propter hoc fallacy, whereby this relationshipexists only on account of both factors being stronglyassociated with neuroticism. However, the impact ofnonnormal distributions of depression data incommunity-based samples and the question ofwhether or not these results generalize to a diagnosticcontext require further investigation.

This research was supported by an Edith CowanUniversity Postgraduate Research Scholarship (ECU-PRS) awarded to Matt Merema. The authors have nodisclosures to report.

References

1. Minett TS, Da Silva RV, Ortiz KZ, et al: Subjective memorycomplaints in an elderly sample: a cross-sectional study. Int JGeriatr Psychiatry 2008; 23:49e54

2. Mendes T, Gino S, Ribeiro F, et al: Memory complaints in healthyyoung and elderly adults: reliability of memory reporting. AgingMent Health 2008; 12:177e182

3. Jessen F, Wiese B, Cvetanovska G, et al: Patterns of subjectivememory impairment in the elderly: association with memoryperformance. Psychol Med 2007; 37:1753e1762

4. Minett TS, Dean JL, Firbank M, et al: Subjective memory complaints,white-matter lesions, depressive symptoms, and cognition in elderlypatients. Am J Geriat Psychiat 2005; 13:665e671

5. Lautenschlager NT, Flicker L, Vasikaran S, et al: Subjectivememory complaints with and without objective memoryimpairment: relationship with risk factors for dementia. Am JGeriatr Psychiatry 2005; 13:731e734

6. Zimprich D, Martin M, Kliegel M: Subjective cognitive complaints,memory performance, and depressive affect in old age: a change-oriented approach. Int J Aging Hum Dev 2003; 57:339e366

7. Achat H, Kawachi I, Spiro A, et al: Optimism and depression aspredictors of physical and mental health functioning: thenormative aging study. Ann Behav Med 2000; 22:127e130

8. Camacho TC, Roberts RE, Lazarus NB, et al: Physical activity anddepression: evidence from the Alameda County study. Am JEpidemiol 1991; 134:220e231

9. Hirschfeld RMA: The comorbidity of major depression andanxiety disorders: recognition and management in primary care.Primary Care Companion to J Clin Psychol 2001; 3:244e254

10. Reid LM, MacLullich AM: Subjective memory complaints andcognitive impairment in older people. Dement Geriatr CognDisord 2006; 22:471e485

11. Harwood DG, Barker WW, Ownby RL, et al: No associationbetween subjective memory complaints and apolipoprotein Egenotype in cognitively intact elderly. Int J Geriatr Psychiatry2004; 19:1131e1139

12. Kliegel MK, Zimprich D, Eschen A: What do subjective cognitivecomplaints in persons with aging-associated cognitive declinereflect? Int Psychogeriatr 2005; 17:499e512

13. Pearman A, Storandt M: Predictors of subjective memory in olderadults. J Gerontol B Psychol Sci Soc Sci 2004; 59B:4e6

14. Pearman A, Storandt M: Self-discipline and self-consciousnesspredict subjective memory in older adults. J Gerontol B PsycholSci Soc Sci 2005; 60B:153e157

15. Barnhofer T, Chittka T: Cognitive reactivity mediates the rela-tionship between neuroticism and depression. Behav Res Ther2010; 48:275e281

16. Weber K, Giannakopoulos P, Bacchetta J-P, et al: Personality traitsare associated with acute major depression across the age spec-trum. Aging Ment Health 2011:001e009

17. Gayman MD, Lloyd DA, Ueno K: The history and timing ofdepression onset as predictors of young adult self-esteem. J ResAdolesc 2011; 21:691e702

18. Pirie M: How to win every argument: the use and abuse of logic.London, UK, Continuum International Publishing Group, 2006

19. Merema MR, Speelman CP, Kaczmarek EA, et al: Age and pre-morbid intelligence suppress complaint-performance congruency

735

Depression and Memory Complaints

in raw score measures of memory. Int Psychogeriatr 2012; 24:397e405

20. Gilewski MJ, Zelinski EM, Schaie KW: The memory functioningquestionnaire for assessment of memory complaints in adulthoodand old age. Psychol Aging 1990; 5:482e490

21. Wechsler D: WMS-IV: technical and interpretive manual. SanAntonio, TX, Pearson, 2009

22. Lovibond SH, Lovibond PF:Manual for theDepression Anxiety StressScales; Psychology Foundation. 2nd ed. Sydney, Australia, 1995

23. Ownsworth T, Fleming J, Haines T, et al: Development ofdepressive symptoms during early community reintegration aftertraumatic brain injury. J Int Neuropsychol Soc 2010; 17:112e119

24. Bayram N, Bilgel N: The prevalence and socio-demographiccorrelations of depression, anxiety and stress among a group ofuniversity students. Soc Psychiatry Psychiatr Epidemiol 2008; 43:667e672

25. Dear BF, Titov N, Schwencke G, et al: An open trial of a brieftransdiagnostic internet treatment for anxiety and depression.Behav Res Ther 2011; 49:830e837

26. Gloster AT, Rhoades HM, Novy D, et al: Psychometric propertiesof the Depression Anxiety Stress Scale-21 in older primary carepatients. J Affect Disord 2008; 110:248e259

27. Antony MM, Bieling PJ, Cox BJ, et al: Psychometric properties ofthe 42-item and 21-item versions of the depression anxiety stressscales in clinical groups and a community sample. Psychol Assess1998; 10:176e181

736

28. Costa PT, McCrae RR: Professional manual: revised NEOPersonality Inventory (NEO PI-R) and NEO Five-Factor Inventory(NEO-FFI). Lutz, FL, Psychological Assessment Resources, Inc1992

29. Hickie I, Naismith S, Ward PB, et al: Reduced hippocampalvolumes and memory loss in patients with early- and late-onsetdepression. Brit J Psychiat 2005; 186:197e202

30. Byers AL, Covinsky KE, Barnes DE, et al: Dysthymia and depres-sion increase risk of dementia and mortality among olderveterans. Am J Geriat Psychiat 2012; 20:664e672

31. Kliegel M, Zimprich D: Predictors of cognitive complaints inolder adults: A mixture regression approach. Eur J Ageing 2005;2:13e23

32. Coleman EA, Sackeim HA, Prudic J, et al: Subjective memorycomplaints prior to and following electroconvulsive therapy. BiolPsychiatry 1996; 39:346e356

33. Antikainen R, Hanninen T, Honkalampi K, et al: Mood improve-ment reduces memory complaints in depressed patients. EurArch Psychiatry Clin Neurosci 2001; 251:6e11

34. Zelinski EM, Burnight KP, Lane CJ: The relationship betweensubjective and objective memory in the oldest old: Comparisonsof findings from a representative and convenience sample.J Aging Health 2001; 13:248e266

35. Morey LC, Shea T, Markowitz JC, et al: State effects of majordepression on the assessment of personality and personalitydisorder. Am J Psychiatry 2010; 167:528e535

Am J Geriatr Psychiatry 21:8, August 2013