Distress proneness and cognitive declinein a population of older persons

Robert S. Wilsona,b,*, David A. Bennetta, Carlos F. Mendes de Leonc,Julia L. Bieniasc, Martha C. Morrisc, Denis A. Evansa,c

aRush Alzheimer’s Disease Center and Department of Neurological Sciences,Rush University Medical Center, Chicago, IL, USAbDepartment of Psychology, Rush University Medical Center, Chicago, IL, USAcRush Institute for Healthy Aging and Department of Internal Medicine,Rush University Medical Center, Chicago, IL, USA

Received 17 October 2003; received in revised form 23 April 2004; accepted 26 April 2004

03do

CeCeTe

KEYWORDSLongitudinal studies;Population studies;Cognitive function;Psychological distress;Personality

06-4530/$ - see front matter Q 200i:10.1016/j.psyneuen.2004.04.005

* Corresponding author. Address: Rnter, Rush University Medical Center, 600 South Paulina, Suite 1038l.: C1-312-942-2354; fax: C1-312-9E-mail address: rwilson@rush.edu (

Summary The association between distress proneness and cognitive decline wasexamined in older residents of a biracial community in Chicago. At baseline,participants completed four cognitive tests that yielded a global measure (baselinemeanZ101.2; standard deviation (SD)Z7.8), and a brief measure of the tendency toexperience negative emotions (meanZ16.5; SDZ6.7) based on the Neuroticism scaleof the NEO Five-Factor Inventory. Cognitive testing was repeated twice at three-yearintervals. In mixed models that controlled age, sex, race, and education, for eachpoint on the distress proneness scale, global cognitive score was 0.12 unit lower atbaseline (p!0.001) and annual rate of decline increased by 0.01 unit (pZ0.002), orabout 2%. Thus, cognitive decline was about 30% faster in a person highly prone todistress (scoreZ24, 90th percentile) compared to the one low in distress proneness(scoreZ9, 10th percentile). This effect was unchanged after controlling for level ofcognitive activity or excluding people with cognitive impairment at baseline, but itwas reduced to a trend (pZ0.059) after controlling for depressive symptoms. Theresults suggest that the tendency to experience psychological distress is associatedwith increased cognitive decline in old age.Q 2004 Elsevier Ltd. All rights reserved.

4 Elsevier Ltd. All rights reserv

ush Alzheimer’s Diseasenter, Armour Academic, Chicago, IL 60612, USA.42-2297.R.S. Wilson).

1. Introduction

People differ in the tendency to experiencenegative emotional states like depression, anger,and shame. Because this trait, variously referred toas distress proneness, neuroticism, or negativeaffectivity (Watson and Clark, 1984; Goldberg,

Psychoneuroendocrinology (2005) 30, 11–17

www.elsevier.com/locate/psyneuen

ed.

R.S. Wilson et al.12

1990; Costa and McCrae, 1992), is very stablethroughout adulthood (Costa and McCrae, 1988;Hultsch et al., 1999), it is an indicator of thecumulative level of distress experienced by anindividual (Watson and Clark, 1984; Depue andMonroe, 1986; Ormel and Wohlfarth, 1991; Magnuset al., 1993). In this article, we use the termdistress proneness because it succinctly specifiesthe central feature of the trait.

Several lines of evidence suggest that distressproneness may contribute to cognitive decline anddementia in old age. Chronic stressful experience inanimals has been associated with structuralchanges in neural systems that regulate thehypothalamic–pituitary–adrenal axis (e.g. hippo-campus) and in forms of memory and cognitionmediated by these systems (Sapolsky et al., 1985;Magarinos et al., 1997; Gould et al., 1998; Bakerand Kim, 2002). In humans, negative events acrossthe life span have been associated with increasedrisk of dementia in old age (Persson and Skoog,1996). In addition, distress proneness has beennegatively associated with level of cognitive func-tion in older persons with and without dementia(Jorm et al., 1993; Meier et al., 2002; Wilson et al.,2004a).

Prospective studies provide a more direct oppor-tunity to observe an association between distressproneness and cognitive decline, but few have beenconducted and their results have been mixed. Onestudy found that older persons who were moreprone to experience psychological distress experi-enced more rapid cognitive decline than personsless prone to distress during five to six years offollow-up (Wilson et al., 2003b), but the associationwas not observed in three other studies (Hultschet al., 1999; Wetherell et al., 2002; Jelicic et al.,2003). However, methodological factors like theduration of the follow-up period and the level offollow-up participation may have contributed tothe null results. Another issue is that most previousstudies have been conducted on selected groups ofolder persons, likely limiting the spectrum ofdistress proneness and the opportunity to detectan association between distress proneness andclinical outcomes.

In the present study, we used data from theChicago Health and Aging Project to examine therelation of a brief measure of distress proneness tocognitive decline in old age. Older residents of abiracial community on the south side of Chicagounderwent cognitive function testing at three-yearintervals for a mean of 5.3 years. At baseline, theycompleted a brief measure of distress proneness.We used mixed models to test the association ofdistress proneness with baseline level of cognition

and annual rate of cognitive decline controllingselected demographic and clinical variables.

2. Methods

2.1. Participants

Beginning in 1993, we conducted a census of allhouseholds in a geographically defined communityin Chicago as part of the Chicago Health and AgingProject (Evans et al., 2003; Morris et al., 2002;Wilson et al., 1999, 2002b). Persons aged 65 yearsor older were asked to participate in an in-homeinterview, and 6158 of 7826 eligible persons (79%)did so. The interview was subsequently repeatedtwice, at approximately three-year intervals. Thestudy was approved by the Institutional ReviewBoard of Rush University Medical Center.

Of 6158 persons at baseline, 1175 died before thefirst follow-up. Of the remaining 4983 persons, 4392(88%) completed at least one follow-up, with amean of 2.6 evaluations per person during a mean of5.3 years of observation. They had a mean age of73.9 years (standard deviation (SD)Z6.5) and amean of 12.0 years of education (SDZ3.7); 62%were women, and 62% were black.

2.2. Assessment of distress proneness

Distress proneness was assessed at baseline withfour items from the 12-item Neuroticism scale ofthe NEO Five-Factor Inventory (Costa and McCrae,1992): I am not a worrier; I often feel tense andjittery; I often get angry at the way people treatme; I often feel helpless and want someone else tosolve my problems. Participants rated their level ofagreement with each statement. Item scores,which ranged from 0 to 4 with higher scoresindicating more distress proneness, were summedand then multiplied by three to make the scoremore comparable to the original 12-item version ofthe scale.

2.3. Assessment of cognitive function

Cognitive function was assessed with four tests ateach evaluation. The Symbol Digit Modalities Testis a measure of perceptual speed in whichparticipants try to match as many symbol digitpairs as possible in 90 s (Smith, 1982). Personswere read a brief story and then asked to recall itimmediately and again after an approximately3-min delay (Albert et al., 1991). Scores forimmediate and delayed retention were the

Distress proneness and cognitive decline in a population of older persons 13

number of the 12 story ideas recalled. The fourthtest was the Mini-Mental State Examination, ameasure of global cognitive function (Folstein etal., 1975). As the tests loaded on a single factor ina previous factor analysis (Wilson et al., 1999), weused a composite measure in analyses. It wasconstructed by converting raw scores on each testto a standard scale with a mean of 100 and SD of10, and then computing the mean of thesestandard scores.

2.4. Other data collection

Depressive symptoms were assessed at baselinewith a 10-item version (Shrout and Yager, 1989;Kohout et al., 1993) of the Center for EpidemiologicStudies Depression (CES-D) scale (Radloff, 1977).Participants were asked if they had experiencedeach of 10 symptoms much of the time in the pastweek. The symptoms are: I felt everything I did wasan effort; My sleep was restless; I felt depressed; Iwas happy; I felt lonely; People were unfriendly; Ienjoyed life; I felt sad; I felt that people dislikedme; I could not get going. The score is the numberof symptoms experienced.

At baseline, people rated their frequency ofparticipation in seven cognitively stimulatingactivities (e.g. reading the newspaper) on a five-point scale. The ratings were averaged to yield acomposite measure of cognitive activity frequency.Information on the metric properties of thismeasure and its relation to cognitive decline andAlzheimer’s disease is published elsewhere (Wilsonet al., 1999, 2002b,c, 2003a).

2.5. Data analysis

We used mixed models to characterize individualpaths of change in cognition and to test the relationof distress proneness score to baseline level ofcognition and annual rate of change (Laird and Ware,1982). In this approach, variation is assumed to bedue to people following different underlying pathsand to the observed measurements varying from theunderlying paths. Each individual path was assumedto follow the mean path of the population except forrandom effects, which caused the baseline level ofcognition to be higher or lower and the rate ofchange to be faster or slower. A strength of thisapproach is that initial level of cognition is modeledas a source of random variability and a possiblecorrelate of rate of change in cognition.

We constructed an initial model to examine theassociation of distress proneness with baseline levelof global cognition and annual rate of change. The

model included terms for time (in years sincebaseline), distress proneness, the interaction ofdistress proneness with time, and terms to controlfor the effects of age, sex, race, and education. Theterm for time indicates the mean annual change inthe cognitive score for a typical participant. Theterm for distress proneness is the mean differencein baseline cognitive score associated with one uniton the distress proneness measure. The interactionis the mean difference in annual rate of cognitivechange associated with one unit of distressproneness.

In additional analyses, we examined factors withthe potential to influence the association betweendistress proneness and cognitive decline. Werepeated the core model with terms added fordistress proneness!race and for distress prone-ness!race!time. We constructed similar modelsto examine whether distress proneness interactedwith sex, age or education. To examine whether thepresence of cognitive impairment in the populationmight have affected results, we repeated theanalysis excluding persons at or below the fifthpercentile of global cognition at baseline, and thenexcluding those at or below the 10th and 15thpercentiles. We repeated the core model withterms added for cognitive activity and its inter-action with time and then again with terms forCES-D score and its interaction with time.

Programming was done in SAS (SAS Institute Inc.,2000). All models were validated graphically andanalytically.

3. Results

3.1. Distress proneness measure

Scores on the measure of distress pronenessranged from 0 to 45 (meanZ16.5; SDZ6.7) withhigher scores indicating being more prone toexperience distress. The internal consistency ofthe measure was modest (coefficient alphaZ0.47)reflecting the brevity of the scale. Distressproneness was positively correlated with age andnegatively correlated with education (Table 1).Level of distress proneness was higher in women(meanZ17.2) than men (meanZ15.4; p!0.001)and in African-Americans (meanZ16.8) than whitepersons (meanZ15.7; p!0.001). A higher level ofdistress proneness was associated with a lowerlevel of cognitive activity and a higher level ofdepressive symptomatology on the CES-D scale(Table 1).

Table 1 Correlations among distress proneness and selected covariates.

Distress Age Education Cognitiveactivity

CES-Da

Distressproneness

– 0.07 K0.19 K0.20 0.45

Age – K0.13 K0.16 0.09Education – 0.46 K0.22Cognitiveactivity

– K0.26

CES-Da score –

For all correlations, p!0.001.a CES-D is for Center for Epidemiologic Studies Depression scale.

R.S. Wilson et al.14

3.2. Distress proneness and cognitive decline

At baseline, scores on the global measure ofcognitive function ranged from 65.0 to 115.4(meanZ101.2; SDZ7.8), with higher scores indi-cating better cognitive function. We examined therelation of distress proneness to baseline cognitionand annual rate of change in a mixed modeladjusted for age, sex, race, and education(Table 2, model 1). The global cognitive scoredeclined an average of 0.61 unit annually, as shownby the term for time in the table. With each onepoint on the distress proneness scale, globalcognition was an average of 0.12 unit lower atbaseline, as shown by the term for distress prone-ness. Annual rate of global cognitive declineincreased by 0.01 unit or about 2%, with eachadditional point on the distress proneness scale, asshown by the interaction term. On an average,therefore, a person with a high level of distressproneness (scoreZ24, 90th percentile) declinedabout 30% faster than a person with a low level ofdistress proneness (scoreZ9, 10th percentile).

Table 2 The relation of distress proneness to baseline lechange estimated from mixed models that controlled fordepressive symptoms (model 2).

Model term Model 1

Estimate SE

Time K0.61 0.04Distress proneness K0.12 0.01Distress proneness!time K0.01 !0.01

Results show the mean cognitive change per year of time in the studyon baseline cognition and annual rate of change in cognition.

3.3. Influences on the distressproneness–cognitive decline association

In view of the association of distress proneness withdemographic variables at baseline, we consideredthe possibility that the association of distressproneness with cognitive decline might vary indemographic subgroups. Because information aboutracial differences in negative emotions is limited(Barnes et al., 2004; Blazer et al., 1998; Kubzanskyet al., 2000), we first examined whether theassociation differed in African-American comparedto white persons. We repeated the original modelwith terms for the interaction of race with distressproneness and for the triple interaction of race,distress proneness, and time. The triple interactionwas not significant, meaning that the associationbetween distress proneness and cognitive declinedid not vary by race. We constructed identicalmodels separately for sex, age, and education. Wefound no evidence that sex, age, or educationmodified the association of distress proneness withcognitive decline (all p>0.500).

vel of global cognition and to annual rate of cognitiveage, sex, race, and education (model 1) and also for

Model 2

p-value Estimate SE p-value

!0.001 K0.58 0.04 !0.001!0.001 K0.11 0.01 !0.001

0.002 K0.01 !0.01 0.059

and the mean effect of one unit on the distress proneness scale

Distress proneness and cognitive decline in a population of older persons 15

Because the community population includedsome persons with substantial cognitive impair-ment, we conducted additional analyses to see ifthis subgroup affected results. We first repeatedthe analysis of distress proneness and change incognitive function excluding persons whose base-line global cognitive score was at or below the fifthpercentile. The association of distress pronenesswith cognitive decline was not substantially chan-ged (estimateZK0.01, SE!0.01, pZ0.011). Theeffect of distress proneness was similar whenwe excluded those with cognition at or below the10th (estimateZK0.01, SE!0.01, pZ0.012) or15th (estimateZK0.01, SE!0.01, pZ0.043)percentile.

Because frequency of cognitive activity has beenrelated to rate of cognitive decline in this (Wilsonet al., 2003a) and other (Hultsch et al., 1999;Wilson et al., 2002c) cohorts, and because distressproneness was negatively related to frequency ofcognitive activity, we repeated the analysis withterms added for cognitive activity and its inter-action with time. The association of distressproneness with cognitive decline was unchangedin this analysis (estimateZK0.01, SE!0.01, pZ0.010), and higher level of cognitive activitywas still related to reduced cognitive decline(estimateZ0.12, SEZ0.03, p!0.001).

Persons prone to distress tend to experiencemore depressive symptoms and other forms ofnegative effect than persons not prone to distress(Ormel and Wohlfarth, 1991; Magnus et al., 1993),and depressive symptoms have been associated withcognitive decline (Bassuk et al., 1998; Paternitiet al., 2002; Yaffe et al., 1999; Wilson et al., 2002a)and risk of Alzheimer’s disease (Devanand et al.,1996; Wilson et al., 2002a, 2003c). Therefore, werepeated the analysis with a term added for CES-Dscore and its interaction with time. The meannumber of depressive symptoms on the CES-D scalein this population was 1.5 (SDZ2.0), which iscomparable to an earlier study of older personsresiding in the community of East Boston, MA(Coroni-Huntley et al., 1986). In this analysis(Table 2, model 2), the association of distressproneness with cognitive decline was reduced to atrend (estimateZ0.01, SE!0.01, pZ0.059), andthe association of CES-D score with cognitive declineremained (estimateZ0.02, SEZ0.01, pZ0.035).

4. Discussion

In this biracial community of older persons, wefound that level of distress proneness was associ-ated with rate of cognitive decline during about fiveyears of observation. The average rate of cognitive

decline in people highly prone to experiencepsychological distress was about 30% greater thanthe rate in those not prone to distress. The resultssuggest that distress proneness is associated withincreased risk of cognitive decline in old age.

Previous prospective studies of the relation ofdistress proneness to cognitive decline have yieldedmixed results, as noted above. In one (Wilson et al.,2003b), distress proneness was associated withmore rapid cognitive decline and an increased riskof dementia, consistent with the present results.Another study (Hultsch et al., 1999) found noassociation between distress proneness and cogni-tive decline during a six-year period, but analyseswere based on two data points and only about halfof the cohort participated in the second evaluation.In another study with negative results (Wetherellet al., 2002), about one quarter of the cohort waslost to follow-up and cases of incident dementiawere excluded from analyses, possibly limiting thepower to detect an association. A third prospectivestudy with negative results was conducted on apopulation-based sample (Jelicic et al., 2003), butdistress proneness was only assessed in a smallsubset of the sample and rate of participation infollow-up was indeterminate, making the resultsdifficult to interpret.

The basis of the association between distressproneness and cognitive decline is uncertain. Theeffect was reduced after controlling for depress-ive symptomatology, which has been associatedwith cognitive decline in this (Wilson et al.,2004b) and other (Bassuk et al., 1998; Paternitiet al., 2002; Wilson et al., 2002a; Yaffee et al.,1999) cohorts, consistent with the idea thatdistress proneness is associated with cognitivedecline by virtue of its association with negativeemotional states. Negative mood states could bean early sign of a neurodegenerative conditionlike Alzheimer’s disease. In recent clinicopatho-logic studies, however, neither distress pronenessnor depressive symptomatology was related toquantitative measures of Alzheimer’s diseasepathology or modified the relation of thatpathology to clinical disease (Wilson et al.,2003b,c). We think a more likely possibility isthat chronic psychologic distress eventually com-promises neural systems that regulate the hypo-thalamic–pituitary–adrenal axis, including regionsin the hippocampus, amygdala, and prefrontalcortex that also contribute to memory andcognition. This hypothesis is especially intriguingbecause animal research suggests that anti-depressants may prevent much of the deleteriousimpact of stress on the central nervous system(Malberg et al., 2000; Shakesby et al., 2002).

R.S. Wilson et al.16

Clinicopathologic studies will be needed toelucidate the neurobiologic basis of the associ-ation of psychological distress with cognitivedecline.

Confidence in these findings is strengthened byseveral factors. Participants were drawn from adefined community, increasing the likelihood thatthere was a broad spectrum of distress pronenessand cognitive ability. The availability of more than2.5 observations per person during an average ofmore than five years of follow-up with a high rate offollow-up participation and a psychometricallyestablished measure of cognition enhanced ourability to characterize individual paths of cognitivedecline. The study also has important limitations.We used a brief measure of distress pronenesswhich lacks the measurement precision of longerscales (Costa and McCrae, 1992); we also used ameasure of global cognition but prior researchsuggests that distress proneness is more stronglyassociated with some forms of cognition than others(Wilson et al., 2003b, 2004a). For both thesereasons, the current results may underestimatethe magnitude of the association between distressproneness and cognitive decline in old age.

Acknowledgements

The authors thank the residents of Morgan Park,Washington Heights, and Beverly who participatedin the study. They also thank Ms Ann Marie Lane forcommunity development and oversight of projectcoordination, Ms Michelle Bos, Ms Holly Hadden,Mr Flavio LaMorticella, and Ms Jennifer Tarpey forcoordination of the study, Mr Todd Beck andMs Liping Gu for analytic programming, and thestaff of the Rush Institute for Healthy Aging. Thisresearch was supported by National Institute onAging grants AG11101 and AG10161 and by NationalInstitute of Environmental Health Sciences grantES10902.

References

Albert, M.S., Smith, L.A., Scherr, P.A., Taylor, J.O., Evans, D.A.,Funkenstein, H.H., 1991. Use of brief cognitive tests toidentify individuals in the community with clinically diag-nosed Alzheimer’s disease. Int. J. Neurosci. 57, 167–178.

Baker, K.B., Kim, J.J., 2002. Effects of stress and hippocampalNMDA receptor antagonism on recognition memory in rats.Learn. Mem. 9, 58–65.

Barnes, L.L., Mendes de Leon, C.F., Wilson, R.S., Bienias, J.L.,Bennett, D.A., Evans, D.A., 2004. Racial differences inperceived discrimination in a community population ofolder blacks and whites. J. Aging Health 16, 315–337.

Bassuk, S.S., Berkman, L.F., Wypij, D., 1998. Depressivesymptomatology and incident cognitive decline in an elderlycommunity sample. Arch. Gen. Psychiatry 55, 1073–1081.

Blazer, D.G., Landerman, L.R., Hays, J.C., Simonsick, E.M.,Saunders, W.B., 1998. Symptoms of depression amongcommunity-dwelling elderly African-American and whiteolder adults. Psychol. Med. 28, 1311–1320.

Cornoni-Huntley, J., Brock, D.B., Ostfeld, A.M., Taylor, J.O.,Wallace, R.B. (Eds.), 1986. Established Populations forEpidemiologic Studies of the Elderly Resource Data Book. USGovernment Printing Office, Washington, DC.

Costa, P.T., McCrae, R.R., 1988. Personality in adulthood: asix-year longitudinal study of self-reports and spouse ratingson the NEO personality inventory. J. Pers. Soc. Psychol. 54,853–863.

Costa, P.T., McCrae, R.R., 1992. Revised NEO PersonalityInventory (NEO-PI-R) and NEO Five-Factor Inventory (NEO-FFI) Professional Manual. Psychological AssessmentResources, Odessa, FL.

Depue, R.A., Monroe, S.M., 1986. Conceptualization andmeasurement of human disorder in life stress research: theproblem of chronic disturbance. Psychol. Bull. 99, 36–51.

Devanand, D.P., Sano, M., Tang, M-X., Taylor, S., Gurland, B.J.,Wilder, D., Stern, Y., Mayeux, R., 1996. Depressed mood andthe incidence of Alzheimer’s disease in the elderly living inthe community. Arch. Gen. Psychiatry 53, 175–182.

Evans, D.A., Bennett, D.A., Wilson, R.S., Bienias, J.L.,Morris, M.C., Scherr, P.A., Herbert, L.E., Aggarwal, N.A.,Beckett, L.A., Joglekar, R., Berry-Kravis, E., Schneider, J.A.,2003. Incidence of Alzheimer’s disease in a biracial urbancommunity. Arch. Neurol. 60, 185–189.

Folstein, M.F., Folstein, S.E., McHugh, P.R., 1975. Mini-mentalstate: a practical method for grading the mental state ofpatients for the clinician. J. Psychiatr. Res. 12, 189–198.

Goldberg, L.R., 1990. An alternative description of personality:the big-five structure. J. Pers. Soc. Psychol. 59, 1216–1229.

Gould, E., Tanapat, P., McEwen, B.S., Flugge, G., Fuchs, E.,1998. Proliferation of granule cell precursors in the dentategyrus of adult monkeys is diminished by stress. Proc. NatlAcad. Sci. USA 95, 3168–3171.

Hultsch, D., Hertzog, C., Small, B., Dixon, R., 1999. Use it or loseit: engaged lifestyle as a buffer of cognitive decline in aging?.Psychol. Aging 14, 245–263.

Jelicic, M., Bosma, H., Ponds, R.W.H.M., van Boxtel, M.P.J.,Houx, P.J., Jolles, J., 2003. Neuroticism does notaffect cognitive functioning in later life. Exp. Aging Res.29, 73–78.

Jorm, A.F., Mackinnon, A., Christensen, H., Henderson, S.,Scott, R., Korten, A., 1993. Cognitive functioning andneuroticism in an elderly community sample. Pers. Individ.Diff. 15, 721–723.

Kohout, F.J., Berkman, L.F., Evans, D.A., Cornoni-Huntley, J.,1993. Two shorter forms of the CES-D depression symptomsindex. J. Aging Health 5, 179–193.

Kubzansky, L.D., Berkman, L.F., Seeman, T.E., 2000. Socialconditions and distress in elderly persons: findings from theMacArthur studies of successful aging. J. Gerontol.: Psychol.Sci. 55B, P238–P246.

Laird, N., Ware, J., 1982. Random-effects models for longitudi-nal data. Biometrics 36, 963–973.

Magarinos, A.M., Verdugo, J.M.G., McEwen, B.S., 1997. Chronicstress alters synaptic terminal structure in hippocampus.Proc. Natl Acad. Sci. USA 94, 14002–14008.

Magnus, K., Diener, E., Fujita, F., Pavot, W., 1993. Extraversionand neuroticism as predictors of objective life event: alongitudinal analysis. J. Pers. Soc. Psychol. 65, 1046–1053.

Distress proneness and cognitive decline in a population of older persons 17

Malberg, J.E., Eisch, A.J., Nestler, E.J., Duman, R.S., 2000.Chronic antidepressant treatment increases neurogenesis inadult rat hippocampus. J. Neurosci. 20, 9104–9110.

Meier, B., Perrig-Chiello, P., Perrig, W., 2002. Personality andmemory in old age. Aging Neuropsychol. Cogn. 9, 135–144.

Morris, M.C., Evans, D.A., Bienias, J.L., Tangney, C.C.,Bennett, D.A., Aggarwal, N.T., Wilson, R.S., Scherr, P.A.,2002. Dietary intake of antioxidant nutrients and the risk ofincident Alzheimer’s disease. JAMA 287, 3230–3237.

Ormel, J., Wohlfarth, T., 1991. How neuroticism, long-termdifficulties, and life situation change influence psychologicaldistress: a longitudinal model. J. Pers. Soc. Psychol. 60,744–755.

Paterniti, S., Verdier-Taillefer, M.-H., Dufouil, C.,Alperovitch, A., 2002. Depressive symptoms and cognitivedecline in elderly people. Br. J. Psychiatry 181, 406–410.

Persson, G., Skoog, I., 1996. A prospective population study ofpsychosocial risk factors for late onset dementia. Int.J. Geriatr. Psychiatry 11, 15–22.

Radloff, L.S., 1977. The CES-D scale: a self report depressionscale for research in the general population. Appl. Psychol.Meas. 1, 385–401.

Sapolsky, R.M., Krey, L.C., McEwan, B.S., 1985. Prolongedglucocorticoid exposure reduces hippocampal neuron num-ber: implication for aging. J. Neurosci. 5, 1222–1227.

SAS Institute Inc., 2000. SAS/STATw User’s Guide, Version 8. SASInstitute Inc., Cary, NC.

Shakesby, A.C., Anwyl, R., Rowan, M.J., 2002. Overcoming theeffects of stress on synaptic plasticity in the intacthippocampus: rapid actions of serotonergic and antidepress-ant agents. J. Neurosci. 22, 3638–3644.

Shrout, P.E., Yager, T.J., 1989. Reliability and validity ofscreening scales: effect of reducing scale length. J. Clin.Epidemiol. 42, 69–78.

Smith, A., 1982. Symbol Digit Modalities Test Manual-Revised.Western Psychological Press, Los Angeles, CA.

Watson, D., Clark, L.A., 1984. Negative affectivity: the disposi-tion to experience aversive emotional states. Psychol. Bull.96, 465–490.

Wetherell, J.L., Reynolds, C.A., Gatz, M., Pedersen, N.L., 2002.Anxiety, cognitive performance and cognitive decline innormal aging. J. Gerontol.: Psychol. Sci. 57B, P246–P255.

Wilson, R.S., Bennett, D.A., Beckett, D.A., Morris, M.C.,Gilley, D.W., Bienias, J.L., Scherr, P.A., Evans, D.A., 1999.Cognitive activity in older persons from a geographicallydefined population. J. Gerontol.: Psychol. Sci. 54B, P155–P160.

Wilson, R.S., Barnes, L.L., Mendes de Leon, C.F., Aggarwal, N.T.,Schneider, J.A., Bach, J., Pilat, J., Beckett, L.A.,Arnold, S.E., Evans, D.A., Bennett, D.A., 2002. Depressivesymptoms, cognitive decline, and risk of AD in older persons.Neurology 59, 364–370.

Wilson, R.S., Bennett, D.A., Bienias, J.L., Aggarwal, N., Mendesde Leon, C.F., Morris, M.C., Schneider, J.A., Evans, D.A.,2002. Cognitive activity and incident AD in a population-basedsample of older persons. Neurology 59, 1910–1914.

Wilson, R.S., Mendes de Leon, C.F., Barnes, L.L.,Schneider, J.A., Bienias, J.L., Evans, D.A., Bennett, D.A.,2002. Participation in cognitively stimulating activities andrisk of incident Alzheimer’s disease. JAMA 287, 742–748.

Wilson, R.S., Bennett, D.A., Bienias, J.L., Mendes de Leon, C.F.,Morris, M.C., Evans, D.A., 2003. Cognitive activity andcognitive decline in a biracial community population.Neurology 61, 812–816.

Wilson, R.S., Evans, D.A., Bienias, J.L., Mendes de Leon, C.F.,Schneider, J.A., Bennett, D.A., 2003. Proneness to psycho-logical distress is associated with risk of Alzheimer’s disease.Neurology 61, 1479–1485.

Wilson, R.S., Schneider, J.A., Bienias, J.L., Arnold, S.E.,Evans, D.A., Bennett, D.A., 2003. Depressive symptoms,clinical AD, and cortical plaques and tangles in older persons.Neurology 61, 1102–1107.

Wilson, R.S., Fleischman, D.A., Myers, R.A., Bennett, D.A.,Bienias, J.L., Gilley, D.W., Evans, D.A., 2004. Premorbidproneness to distress and episodic memory impairment in AD.J. Neurol. Neurosurg. Psychiatry 75, 191–195.

Wilson, R.S., Mendes de Leon, C.F., Bennett, D.A., Bienias, J.L.,Evans, D.A., 2004. Depressive symptoms and cognitivedecline in a community population of older persons.J. Neurol. Neurosurg. Psychiatry 75, 126–129.

Yaffe, K., Blackwell, T., Gore, R., Sands, L., Reus, V.,Browner, W.S., 1999. Depressive symptoms and cognitivedecline in nondemented elderly women. Arch. Gen. Psychia-try 56, 425–430.