Aging and Loneliness

7/30/2019 Aging and Loneliness

1/12

.International Journal of Psychophysiology 35 2000 143

154

Lonely traits and concomitant physiological processes:the MacArthur social neuroscience studies

John T. Cacioppoa,U, John M. Ernst b, Mary H. Burlesonc, Martha K.McClintocka, William B. Malarkey d, Louise C. Hawkley d, Ray B.

Kowalewskid, Alisa Paulsend, J. Allan Hobsone, Kenneth Hugdahlf,David Spiegelg, Gary G. Berntsond

aDepartment of Psychology, Uniersity of Chicago, 5848 S. Uniersity Aenue, Chicago, IL 60637 USAbDepartment of Psychology, Illinois Wesleyan Uniersity, P.O. Box 2900, Bloomington, IL 61702 USA

cDepartment of Social and Beha ioral Sciences, Arizona State Uni ersity West, 4701 W. Thunderbird, Phoenix, AZ 85069

USAdDepartment of Psychology, Ohio State Uni ersity, 1885 Neil Aenue, Columbus, OH 43210 USA

eDepartment of Psychiatry, Har ard Medical School, 74 Fenwood Road, Boston, MA 02115 USA

fDepartment of Biological and Medical Psychology, Uniersity of Bergen, Arstadeien 21, N-50009 Bergen, NorwaygDepartment of Psychiatry and Behaioral Sciences, Stanford Uni ersity School of Medicine, Room 2325, MC 5544,

Stanford, CA 94305-5544 USA

Received 23 February 1999; accepted 23 March 1999

Abstract

Loneliness is a complex set of feelings encompassing reactions to unfulfilled intimate and social needs. Although

transient for some individuals, loneliness can be a chronic state for others. Prior research has shown that loneliness isa major risk factor for psychological disturbances and for broad-based morbidity and mortality. We examineddifferences between lonely and socially embedded individuals that might explain differences in health outcomes.Satisfying social relationships were associated with more positive outlooks on life, more secure attachments andinteractions with others, more autonomic activation when confronting acute psychological challenges, and moreefficient restorative behaviors. Individuals who were chronically lonely were characterized by elevated mean salivarycortisol levels across the course of a day, suggesting more discharges of corticotropin-releasing hormone and elevated

UCorresponding author. Tel.: q1-773-702-1962; fax: q1-773-702-0886.

.E-mail address: [email protected] J.T. Cacioppo

0167-8760r00r$ - see front matter 2000 Elsevier Science B.V. All rights reserved. .PII: S 0 1 6 7 - 8 7 6 0 9 9 0 0 0 4 9 - 5


2/12

( )J.T. Cacioppo et al. rInternational Journal of Psychophysiology 35 2000 143154144

activation of the hypothalamicpituitaryadrenocorticol axis. An experimental manipulation of loneliness furthersuggested that the way in which people construe their self in relation to others around them has powerful effects ontheir self concept and, possibly, on their physiology. 2000 Elsevier Science B.V. All rights reserved.

Keywords: Loneliness; Heart rate; Salivary cortisol; Blood pressure; Cardiovascular reactivity; Socially embedded

1. Introduction

Humans are social animals, so much so that abasic need to belong has been posited Baumeis-

ter and Leary, 1995; Gardner et al., in review;.Gardner et al., in press . People form associations

and connections with others from the moment

they are born. The very survival of newbornsdepends on their attachment to and nurturanceby others over an extended period of time. Itshould be no surprise that evolution has sculptedthe human genome to be sensitive to and suc-coring of contact and relationships with others.For instance, caregiving and attachment have

.hormonal e.g. Uvnas-Mosberg, 1997 and neuro- .physiological substrates cf. Carter et al., 1997 .

Communication, the bedrock of complex socialinteraction, is universal and ubiquitous in hu-mans. In the rare instances in which human lan-guage is not modeled or taught, language devel-

ops nevertheless e.g. Goldin-Meadow and My-.lander, 1983, 1984 .

The need to belong does not stop at infancybut rather affiliation and nurturant social rela-tionships appear to be essential for physical andfor psychological well-being across the lifespan.Disruptions of social connections, whetherthrough ridicule, separation, divorce, or bereave-ment, are among the most stressful events people

.must endure Gardner et al., in press . Berkman

.and Syme 1979 , for instance, operationalizedsocial connections as marriage, contacts withfriends and extended family members, churchmembership, and other group affiliations. Theyfound that adults with fewer social connectionssuffered higher rates of mortality over the suc-ceeding 9 years even after accounting for self-reports of physical health, socioeconomic status,smoking, alcohol consumption, obesity, race, lifesatisfaction, physical activity, and preventive

.health service usage. House et al. 1982 repli-cated these findings using physical examinationsto assess health status. In their review of five

.prospective studies, House et al. 1988 concludedthat social isolation was a major risk factor formorbidity and mortality from widely varyingcauses. This relationship was evident even after

statistically controlling for known biological riskfactors, social status, and baseline measures ofhealth. The negative health consequences of so-cial isolation were particularly strong among someof the fastest growing segments of the population:the elderly, the poor, and minorities such as

African-Americans. Astonishingly, the strength ofsocial isolation as a risk factor was comparable tosmoking, high blood pressure, obesity, and seden-tary lifestyles.

Social isolation and loneliness are associatedwith poorer mental as well as physical well-beinge.g. Perkins, 1991; Gupta and Korte, 1994; Ernst

.and Cacioppo, 1999 . Conversely, people who re-port having contact with five or more intimatefriends in the prior 6 months are 60% more likelyto report that their lives are very happy, ascompared to those who do not report such con-

.tact Burt, 1986 . People appear to be cognizantof the importance of social relationships. Whenasked what is necessary for happiness? mostrated relationships with friends and family as

.being the most important factor Berscheid, 1985 .

Although social isolation is multi-dimensional,it generally appears that the intimacy and emotio-nal nourishment provided by at least one otherindividual are key to buffering the effects of themajority of stressors. In their seminal review,

.House et al. 1988 concluded that:

... the mere presence of, or sense of relatedness with,

another organism may have relatively direct motivational,

emotional, or neuroendocrinal effects that promote health


3/12

( )J.T. Cacioppo et al. rInternational Journal of Psychophysiology 35 2000 143154 145

either directly or in the face of stress or other healthhazards but that operate independently of cognitive ap-

.praisal or behavioral coping and adaptation p. 544 .

Our goal here is to provide a preliminary re-port of an ongoing study of the psychological andphysiological differences between individuals dif-fering in social embeddedness. Although thereare gripping states of loneliness that everyoneexperiences transiently in specific circumstancesor interactions, some individuals live in the devas-tating clutches of loneliness even though they arenot physically and socially isolated. Our target

.sample was college undergraduates because athis is an active period of dating, mate selection,

.and sexual activity; b habitual patterns of health

behaviors and of interacting with others are being .established; and c social relationships undergo

an upheaval when individuals first go to collegeyet ample opportunities exist for making newacquaintances and friends.

2. Loneliness on college campus

We tested 2632 male and female undergradu-ates at the Ohio State University to determinetheir feelings of loneliness and their living cir-cumstances. Loneliness was unrelated to variousfeatures of college life such as the number ofroommates with whom they lived, the percent

who belonged to organizations, the number ofquarters of college completed, the number ofhours in college in which they were enrolled, thenumber of quarters needed to graduate, and thenumber of hours they exercised weekly. Despitethese superficial similarities, lonely individualsclearly felt less connected to the people around

them. Loneliness, for instance, was negativelycorrelated with the extent to which roommatesreduced feelings of isolation, and positively re-lated to feelings of dysphoria and fears of publicspeaking. Lonely individuals rated their primaryroommate less positively than embedded individu-als, and lonely individuals who lived in dormito-ries reported having fewer positive relationships

with suitemates, floormates, and hallmates thanembedded individuals.

Lonely individuals were slightly though signifi-cantly less likely to consume alcohol than socially

.embedded i.e. low in loneliness individuals aresult that is consistent with a study reported

.recently by Eccles et al. 1997 . They conducted a6-year study of teens free-time activities,academic performance, and alcohol use in 125910th graders from 12 school districts around De-troit. They found that high school athletes hadhigher grades and stayed in college longer but

were also more likely to use alcohol and drugs.The authors noted that athletes tend to formpopular cliques who party, and alcohol is part ofthat entertainment. The same appears true ofsocially connected students in college.

We recruited approximately 5% from our origi-nal sample of 2632 students for more intensivefollow-up study. Inclusion criteria included theirbeing enrolled in at least 6 credit hours in thequarter they were tested, none were first quarterfreshman or last quarter seniors, none scoredhigher than mildly dysphoric on the Beck Depres-

sion Inventory, none were obese body mass index.-27 , none were speech or needle phobic, none

were married or living with a significant other,and all were US citizens. We recruited an equalnumber of male and female students whose scores

on the UCLA loneliness scale Russell et al.,. .1980 fell into the top lonely group , middle

. normal group , or bottom quintile socially em-.bedded group . The middle group was included to

allow us to determine whether differencesbetween lonely and socially embedded reflectedsomething special about lonely individuals, some-thing unique about embedded individuals, orsomething that varied monotonically betweenthese two extreme groups.

3. Differences in social capital

We first examined the hypothesis that individu-als who were lonely simply had less social capitalto offer than others. By social capital, we meanthe resources they bring to a social interaction their physical attractiveness, intelligence, height,

weight, age, socioeconomic status, or scholastic


4/12


achievements. Analyses revealed the groups didnot differ on any of these variables.

We next examined the hypothesis that individu-als who were lonely reported more traumatic life

histories, were dealing with more major life events,or were suffering from more intrusive events.

Again, no differences were found among thegroups. Instead, the differences among groups

were in the way they appraised people and eventsaround them and the way in which they related toothers. For instance, lonely individuals reportedhigher levels of perceived stress, more frequentand more severe hassles, and less severe upliftsthan embedded individuals. The lonely and nor-mal participants responded similarly on most of

these measures, however, it was the socially em-bedded group who appeared the least stressed.The same pattern was found on the Profile ofMoods States, which revealed embedded individu-als reported feeling more vigor and less tension,hostility, confusion, fatigue, and dejection thanthe normal and lonely groups; of these states, thenormal and lonely individuals differed only onfeelings of dejection.

The social world also emerged as a less reward-ing place for lonely individuals. Lonely relative to

embedded individuals, reported comparable so-cial desirability motives but had less secure adultromantic attachment styles. Lonely individuals

were also characterized by greater anxiety, anger .including anger-in , and shyness, less sociability,less optimism, and poorer social skills, and theyexpressed stronger fears of negative evaluation.Lonely individuals were no less assertive thanembedded individuals in terms of the likelihoodof a behavior, but they reported much higherlevels of discomfort and anxiety about being as-

sertive.This does not mean that lonely individuals do

not know how to interact socially; they do sofairly well with their most important contactthough less well with their most frequent con-

.tact . Several different measures revealed thatlonely, relative to embedded, individuals weremore likely to attribute problems in social rela-tionships to others, to view themselves as victims

who are already giving as much as they can to

their relationships with others. The normal groupgenerally fell between the lonely and embeddedgroups on these measures.

What happens when you draw upon social capi-

tal? Does seeking social support and assistancedraw down the balance subsequently available,like a bank account, or does it renew itself throughmutually reinforcing interactions? The latter wassuggested when we found that lonely individualsdrew less upon social capital than normals andembedded individuals. To examine how our par-ticipants generally coped with stressors in theirlives, we had the participants complete the COPE

.scale Carver et al., 1989 . The COPE contains 15subscales of four items each: active coping, plan-

ning, seeking instrumental social support, seekingemotional social support, suppression of compet-ing activities, religion, positive reinterpretationand growth, restraint coping, acceptance, focus on

venting of emotions, denial, mental disengage-ment, behavioral disengagement, alcoholrdruguse, and humor. The groups were found to differon four of these subscales: lonely individuals wereless likely to actively cope, seek instrumental sup-port from others, or seek emotional support fromothers and were more likely to behaviorally disen-gage than were embedded individuals. Appar-ently, the balance of social capital is not governedlike bank accounts, with the total available de-creasing with each request for assistance. Instead,they suggest that being a friend in need, espe-cially when the assistance is effective and dulyappreciated, may build closer ties between thehelper and requester.

4. Differences in autonomic activation

If anxiety and anger are especially powerful inlonely individuals, then they might be expected toshow greater autonomic activation than embed-ded individuals. If, however, behavioral or emo-tional disengagement from the social environ-ment is especially powerful in lonely individuals,then they might be expected to show less au-tonomic activation than embedded individuals. Toexamine this issue, participants were tested at the


5/12


General Clinical Research Center at The OhioState University Hospitals at approximately thesame time in the late afternoon.

Participants completed two social speeches e.g.

asking someone out for a date, describing why.youre a likable person , two non-social speeches

e.g. describing the objects in the room, describ-ing the route from your residence to your first

.class of the week , and a mental arithmetic task .Cacioppo et al., 1995 . These five tasks wererandomly ordered for each participant. Beforeeach speech task there was a 2-min sitting restperiod, and preceding the verbal mentalarithmetic task there was a 4-min sitting restperiod. The experimental tasks then concluded

with the orthostasis task followed by a 4-minsitting rest period and a speech task about one .being accused of shoplifting Saab et al., 1989 .

The results, which are displayed in Fig. 1, re-vealed that across tasks the lonely individualswere characterized by lower basal heart rate andlower heart rate reactivity than normal or embed-ded individuals. Basal blood pressure was compa-rable across the tasks, although again cardiovas-cular reactivity tended to be lowest for the lonelyindividuals. These data are consistent with thenotion that lonely individuals are more emotio-nally withdrawn than normal or embedded indi-

viduals when in a new social setting.

5. Differences in attentional control

Why might lonely individuals be more emotio-nally withdrawn in new social settings? Personal-ity differences such as shyness, sociability, nega-tivity, and fear of negative evaluation provide apartial explanation. New social settings can be

overwhelming, however, and people must exertvoluntary control over their attentional focus tobe effective. Do lonely individuals differ in theirability to voluntarily control their attentional fo-cus?

To explore this possibility, participants per- .formed a dichotic listening task Hugdahl, 1995

while at the Clinical Research Center. The di-chotic listening task requires that participantsidentify the consonantvowel pair that was pre-

.Fig. 1. Basal heart rate a and heart rate reactivity to four .classes of stressors b .

sented to their right or left ear. Because theauditory system is predominantly crossed and be-cause language is left-lateralized in most right-handed individuals, right-handed individuals tendto perform better when verbal stimuli are pre-

sented to the right than left ear. All of the.participants in our sample were right-handed.

Superimposed on this general right-ear advan-tage, however, are the effects of attention, asindividuals generally perform better whenever

verbal stimuli are presented to the ear to whichthey are focusing their attention, as well.

As expected, there was an overall right-earadvantage across groups and instructional condi-

.tions see Fig. 2 . In addition, a significant maineffect for attentional instruction showed that in-dividuals performed better with left-ear stimuli

when they were instructed to focus on stimulipresented to their left ear than in the otherconditions. The two-way interaction between at-


6/12


Fig. 2. Dichotic listening performance under no attentional

. .instructions a , instructions to attend to the right ear b , and .instructions to attend to the left ear c .

tentional instruction and ear was also significant,showing a significant right-ear advantage duringthe no-instruction and focus on right-ear condi-tions and a significant left-ear advantage duringthe focus on left-ear condition.

As can also be seen in Fig. 2, lonely individualstended to show the strongest right-ear advantage

in the no-instruction condition, presumably re-flecting the potency of bottom-up i.e. stimulus-

.driven attentional processing. More interestingly,lonely individuals apparently failed to shift to an

a priori predicted left-ear advantage in the focuson left-ear condition, despite showing a largeright-ear advantage when instructed to shift at-tention to the right ear. Specific planned contrastsconfirmed that all three groups showed a signifi-cant right-ear advantage during the focus onright-ear condition, but only the normal and em-bedded individuals were able to revert to a sig-nificant left-ear advantage in the focus on left-earcondition. Thus, attentional control appearedcomparable in lonely and embedded individuals

until voluntary attentional control conflicted withautomatic attentional processes, at which pointlonely individuals showed an attentional deficit.This result raises the possibility that lonely indi-

viduals feel overwhelmed and withdraw from thesocial environment, especially new or complexsocial environments, because they have less con-

trol over the focus of their attention i.e. they are.more distractable .

6. Differences in neuroendocrine activation

As noted above, a major objective in this studywas to explore potential mechanisms underlyingthe relationship between social isolation and mor-bidity and mortality. Our data indicated thatlonely, relative to embedded, individuals engaged

in comparable or better health behaviors e.g.cigarette smoking, alcohol consumption, drug

.abuse, and exercise . Autonomic reactivity toacute psychological stressors also tended to bediminished in lonely, relative to embedded, indi-

viduals. Prior research has shown autonomic acti-vation to vary with the quality of a persons socialrelationship with others, but this prior researchhas tended to use either more powerful manipu-

lations of social relationships see review by.Gardner et al., in press or older individuals for

whom the cumulative stress of loneliness is higher .see review by Uchino et al., 1996 . Thus, bothhealth behaviors and heightened sympathetictonus remain possible contributors to differences


7/12


in morbidity and mortality in later life, but thesedo not play much of a role in young adults.

Given the behavioral and autonomic evidencethat lonely individuals disengage from the social

environment and the suggestive evidence of theirrelative inability to override automatic attentionalprocessing, we hypothesized that lonely individu-als would show either elevated tonic activation ofthe hypothalamicpituitaryadrenocortical sys-tem or a muted diurnal pattern. If so, this wouldprovide a potential neurophysiological linkbetween loneliness and later-life morbidity andmortality because the pituitary and adrenal hor-mones and other neuropeptides play an impor-tant role in the modulation of the immune system

.Munck et al., 1984 .Cortisol is released in a pulsatile fashion, withmajor secretory episodes appearing during theearly morning with the frequency and amplitudeof secretory episodes declining over the course of

.the day Van Cauter, 1990 . Given the pulsatilenature of cortisol releases, multiple assessmentsover the course of the day may provide a morereliable assessment of HPA activity. Participantsin our study were beeped at nine random inter-

vals during a normal day, at which time they wereasked to sit, place a roll of dental cotton in theirmouth, and complete a short questionnaire. Aftereach such period, the cotton roll was returned toa test tube, which participants returned to us thefollowing day. Subsequently, assays were per-formed to quantify salivary cortisol, which is un-

.bound biologically active and correlates well withplasma and urinary assessments Kirschbaum and

.Hellhammer, 1994 .The salivary cortisol levels across the day, which

are depicted in Fig. 3, reflected the typical diurnalrhythm, with no differences among groups. Spear-

man correlations revealed that the mean dailycortisol level was significantly correlated with theparticipants perceived stress as well as the num-ber of major life events with which they weredealing. Loneliness was related to mean salivarycortisol levels across the course of a normal daybut only when loneliness was chronic. UCLAloneliness scores and state loneliness scores werepositively but non-significantly correlated withmean salivary cortisol levels, whereas trait loneli-

Fig. 3. Spearman correlations between salivary cortisol andthree different measures of loneliness. Salivary cortisol wasassayed up to nine random times during a normal day. Morn-ing represents the correlation between the mean cortisol

during the morning and each loneliness measure. Similarly,Afternoon and Evening represent the correlation betweenthe mean cortisol during the afternoon and evening, respec-tively, and each loneliness measure. Mean represents thecorrelation between the mean salivary cortisol across theentire day and each loneliness measure.

ness scores were positively and significantly corre-lated with mean cortisol levels. The stress ofloneliness and associated events may simply taketime to alter HPA function.

Our participants attended classes during the

morning and afternoon but were left to their owndevices in the evenings. Students study during theevening but this is also the most popular time forstudents to socialize or to think about the absenceof satisfactory social interactions. Inspection ofFig. 3 reveals that it was during the evening thatour participants trait loneliness was most highlyrelated to salivary cortisol levels despite HPAactivity being at its nadir during the evening.

7. Differences in the salubrity of sleep

The salubrity of nourishing social relationshipshas been documented previously, but the benefi-cial effects of nurturant or intimate social rela-tionships are usually attributed to direct assis-

tance or stress-buffering Cohen and McKay,.1984; see Gardner et al., in press . For example,

medical students undergoing exams showedstress-induced decrements in immune function-


8/12


ing, but this decline was particularly pronouncedfor the medical students who reported being

lonely Kiecolt-Glaser et al., 1984; Glaser et al.,.1992 . The association between trait loneliness

and salivary cortisol, particularly in the evening,could be explained in terms of stress buffering ordirect effects of social support. Preliminary analy-ses of diaries completed at random intervals byparticipants during the course of their day sug-gested yet another possible means by which socialembeddedness contributes to lower levels of sali-

vary cortisol and long-term well being.We had anticipated that lonely individuals

would be socially isolated and would be partici-pating in fewer restorative behaviors e.g. sleep,

exercise, going to the movies, spending time with.others, reading, relaxing, praying or meditating .In fact, lonely undergraduates were just as ormore likely to engage in these putatively restora-tive acts than were normal and embedded individ-uals. Yet they did not show the salubrious effectsnormals and socially embedded individualsseemed to show. This led us to hypothesize that

restorative behaviors were more potent i.e. salu-.brious when individuals felt socially connected

than when they felt lonely.To test the hypothesis that feeling socially em-

bedded enhanced the salubrious effects ofrestorative behaviors, we measured sleep aquintessential act of restoration that is performed

without immediate social contact, indeed withoutmuch explicit awareness at all. Survey data indi-cated that lonely individuals slept as many hoursper day as normal and socially embedded individ-uals. We used the Pittsburgh Sleep Quality Index .PSQI; Buysse et al., 1989 to assess our partici-pants reports of sleep disturbances, and we usedthe Nightcap to measure sleep of a subset of our

participants during the night they spent in theClinical Research Center.

Results from the PSQI, which are depicted inFig. 4, revealed that lonely individuals reportedpoorer sleep quality, longer sleep latency, longerperceived sleep duration, and greater daytimedysfunction due to sleepiness than embedded in-dividuals. Data from the Nightcap confirmed thatlonely individuals slept less efficiently, tookslightly longer to fall asleep, evidenced longer

Fig. 4. Sleep disturbance as indexed by the Pittsburgh SleepQuality Index, a self-report measure that measures sleep

.disruptions a , and Nightcap recordings during the night .

participants slept in the Clinical Research Center b .

REM latency, and awoke more frequently duringthe night than embedded and normal individuals .see Fig. 4 . Importantly, the Nightcap recordingsrevealed that total time asleep did not differacross the groups. Thus, the restorative act ofsleep was more efficient and effective in sociallyembedded than lonely individuals.

To what extent are feelings of loneliness orsocial embeddedness causal in producing the ef-

fects outlined above? Are the elevated anxietyand cortisol levels in individuals high in traitloneliness the result of cumulative stressors orare individuals who are characterized by anxietyor elevated HPA activation likely to feel sociallydisconnected? Both are possible, and interesting,and either could help account for the highermorbidity and mortality found in socially isolated

.individuals House et al., 1988 . If only the latterwere true, however, there would be little reason


9/12


to intervene, to try to overcome the feelings ofloneliness, to try to reach out and connect withothers. Fortunately, there is preliminary evidencefor both processes operating.

8. Manipulating feelings of loneliness

.Uchino et al. 1996 reviewed intervention stud-ies that have been performed to increase socialsupport. A meta-analysis of these studies showedthat improving social support and connectedness,especially within families, had beneficial physio-

logical effects e.g. lowered blood pressure in

.hypertensives . Such findings are important be-cause they indicate that feelings of loneliness orsocial embeddedness are not simply markers orepiphenomena, and that loneliness is not an in-

variant trait but rather is subject to some manipu-lation.

Although people develop characteristic ways ofappraising, attaching, and interacting with others,it is conceivable that how we feel about ourselvesin relation to others is the bedrock upon whichthese appraisals, attachment, and interactive stylesare built. If so, then feelings of loneliness mayprime memories, associations, and behaviors ofalienation and social awkwardness, whereas feel-ings of connectedness may prime the opposite.That is, how individuals construe and feel aboutthemselves in terms of those around them may be

so central to some dimensions of personality e.g.shyness, sociability, extraversion, agreeableness,

.fear of negative evaluation that such construalsrepresent a deep structure that influences thesurface manifestations of these personality di-mensions. To test this hypothesis, we used hypno-

sis to manipulate feelings of loneliness and em-beddedness within-subjects. Twenty high hypnoti-zable individuals at Stanford University wereidentified based on high scores on the Harvard

.Group Scale scores 912 and the Hypnotic In- .duction Profile scores 810 .

Following each hypnotic induction, participantsreceived the suggestion that they felt lonely or

socially embedded the order of which was coun-.terbalanced across participants . In the lonely

condition, for instance, participants were told to:Think of a time in which you felt isolated. Youfelt lonely. Perhaps you felt like you just didntbelong that you had no friends.... In the

embedded condition, participants were told to:Think of a time in which you felt a sense ofbelonging. Perhaps you were a member of a group.Perhaps you had a best friend with whom you felt

you could share anything.... After each sugges-tion, participants completed a set of question-naires designed to measure sociability, optimism,fear of negative evaluation, anger, anxiety, socialskills, shyness, self esteem, mood, impact ofevents, loneliness, and perceived social support.

As would be expected if the manipulations were

effective, participants scored higher on lonelinessand lower on social support measures in the lonelythan embedded condition. More interestingly, theparticipants responses to the psychological sur-

veys revealed that they were significantly lower onsocial skills, sociability, optimism, self esteem, andpositivity, and higher on anger, anxiety, negativity,and fear of negative evaluation in the lonely thanembedded condition. Importantly, the partici-pants in the lonely and embedded condition

scored comparably on the subscales intrusion.and avoidance of the impact of events scale

just as had the participants in our cross-sectionalstudy. Thus, the same differences and lack ofdifferences we observed between our lonelyand embedded groups were replicated within sub-

jects when they were induced to feel very lonelyor socially embedded. Together, these results sup-port the hypothesis that how individuals construeand feel about themselves in terms of thosearound them can affect the surface manifesta-tions of states and traits and, presumably, theirphysiology.

9. Conclusion

The current research underscores the centralityof social relationships to human existence andexperience. For instance, lonely individuals weremore anxious, angry, and negative, and less posi-tive, optimistic, comfortable, and secure than em-bedded individuals, a pattern of findings that was


10/12


associated with feelings of loneliness whether ex-amined between-subjects or within-subjects. At-tentional control appeared comparable in lonelyand embedded individuals until voluntary atten-

tional control conflicted with automatic attentio-nal processes, at which point lonely individualsshowed an attentional deficit. Interestingly, lonelyand embedded individuals were found to have

comparable social capital e.g. attractiveness, SES,.intelligence and social contacts, but lonely indi-

viduals, in contrast to embedded individuals, madeless use of social capital, expected negative out-comes, were less likely to reach out or to seekhelp from others, and were more likely to thinkthey were already doing all they could do in their

relationships. Loneliness was associated with arange of altered physiological processes, as well,including muted autonomic activation, elevatedactivation of the HPA axis at least in the subsetof individuals who were chronically lonely, anddisregulated sleep. These data converge to sug-gest that the unfulfilled need to belong is associ-ated with complex yearnings for intimacy yet feel-ings of insecurity and mistrust, anger combined

with anger suppression, punishing feelings ofisolation coupled with a fear of negative evalua-tion by others, emotional dysphoria and with-drawal rather than active coping attempts. Giventhis psychological complex, it is no surprise whydaily encounters yield fewer uplifting events andmore frequent hassles for lonely than embeddedindividuals; across a lifetime, the allostatic load .McEwen and Stellar, 1993 of daily challengesand stressors should be greater, as well.

The social world has tended to be ignored inpsychophysiology and in the neurosciences be-cause it falls outside the physical boundaries ofthe body. Yet our genetic makeup compels us to

be social animals, and social contact and relation-ships are among the most powerful stimuli in theenvironment. Isolation from the social world isknown to be as significant a risk factor as smokingand high blood pressure, but the absence of anobvious physical mechanism has made rigorousscientific analysis difficult. The current researchsuggests that social neuroscience is not an oxy-moron, that instead it represents a viable, multi-level integrative approach for understanding the

mind and brain. It is also consistent with thenotion that the social world in large part defines

who we are, how we appraise and relate to eventsin our lives, and how our biology develops, re-

sponds, and ages.A multi-level integrative approach spanning bi-

ological and social levels of organization does not .imply dualism Cacioppo and Berntson, 1992 .

Multiple levels of analysis do not imply indepen- .dent non-reductionistic determinants. There is a

single set of determinants that may be conceptu-alized at different levels of organization. Hormo-nal effects, for instance, are mediated by theirchemistry but their actions on mating are notoptimally organized by chemicalrreceptor inter-

actions alone. The early American psychologist .William James 1890r1950 was among the firstto articulate the notion that neurophysiologicalprocesses underlie psychological phenomena.James further argued that developmental, envi-ronmental, and sociocultural factors influence theneurophysiological processes underlying psycho-logical and social phenomena. Although theseinfluences could be studied as neurophysiologicaltransactions, James recognized that unnecessarydiseconomies and conundrums would result ifpsychological phenomena were described only asneurophysiological events. Social and psychologi-cal constructs such as loneliness and social em-beddedness, both of which undoubtedly have neu-rophysiological implementations, are simpler tomeasure and are more reliably measured usingself-report instruments than some physiological

.processes e.g. cortisol levels .There is growing evidence that multilevel anal-

yses spanning neural and social perspectives canfoster comprehensive accounts of cognition, emo-tion, behavior, and health. This is in part because

the social environment shapes neural structuresand processes and vice versa. Meaney and col-

.leagues Meaney et al., 1993; Liu et al., 1997 , forinstance, have found that variations in maternalcare influence the development of individual dif-ferences in neuroendocrine responses to stress inrats. As adults, the offspring of mothers thatexhibited more licking and grooming of pups dur-ing the first 10 days of life were also characterizedby reduced adrenocorticotropic hormone and cor-


11/12


ticosterone responses to acute stress. As mothers,these rats also tended to lick and groom theirpups. This research raises the possibility thatlonely and socially embedded individuals have

very different experiences early in life that shapetheir HPA functioning as well as their socialbehavior. That is, although the stress of chronicloneliness may affect HPA activation, the recipro-cal influence also appears likely.

Our results also raise a number of interestingquestions, many of which require experimentaland prospective studies to unravel. What doesseem clear, however, are that these questions areamenable to social, cognitive, and biological lev-els of analysis and that the questions are worth

.asking. As E.O. Wilson 1998 noted,

people must belong to a tribe; they yearn to have apurpose larger than themselves. We are obliged by thedeepest drives of the human spirit to make ourselves morethan animated dust, and we must have a story to tell about

.where we came from, and why we are here p. 6 .

Acknowledgements

We wish to thank David Lozano and Dan Lit-vack of the Ohio State University Social Neuro-science Laboratory for their technical assistance,Robert Stickgold from the Harvard UniversityNeurophysiology Lab and Bita Naouriani fromthe Stanford University Department of Psychiatryfor their assistance in this investigation, the staffof the Ohio State University Residence Hallsdirected by Steve Kramer for their help, CarolynCheney of The Ohio State University Medical

Labs for her contributions, and the General Clin-ical Research Center, including Dana Ciccone,Bob Rice, and the nursing staff headed by TeresaSampsel, for the assistance and cooperation. This

work was supported by training grant MH-19728from the National Institutes of Mental Health, aNational Institute of Health grant to the GeneralClinical Research Center, MO1-RR00034, andthe John D. and Catherine T. MacArthur Foun-dation.

References

Baumeister, R.F., Leary, M.R., 1995. The need to belong:desire for interpersonal attachment as a fundamental hu-man motivation. Psychol. Bull. 117, 497529.

Berkman, L.F., Syme, S.L., 1979. Social networks, host resis-tance, and mortality: a nine-year follow-up study ofAlameda County residents. Am. J. Epidemiol. 109, 186204.

Berscheid, E., 1985. Interpersonal attraction. In: Lindzey, G., .Aronson, E. Eds. , The Handbook of Social Psychology.

Random House, New York, NY.

Buysse, D.J., Reynolds, C.F., Monk, T.H., Berman, S.R. et al.,1989. The Pittsburgh Sleep Quality Index: a new instru-ment for psychiatric practice and research. Psychiatry Res.28, 193213.

Burt, R.S., 1986. Strangers, friends, and happiness. GSS Tech-nical Report No. 72. National Opinion Research Center,

University of Chicago, Chicago.Cacioppo, J.T., Berntson, G.G., 1992. Social psychological

contributions to the decade of the brain: the doctrine ofmultilevel analysis. Am. Psychol. 47, 10191028.

Cacioppo, J.T., Malarkey, W.B., Kiecolt-Glaser, J.K., Uchino,B.N., Sgoutas-Emch, S.A., Sheridan, J.F., Berntson, G.G.,Glaser, R., 1995. Cardiac autonomic substrates as a novelapproach to explore heterogeneity in neuroendocrine andimmune responses to brief psychological stressors. Psy-chosom. Med. 57, 154164.

Carter, C.S., Lederhendler, I.I., Kirkpatrick, B., 1997. TheIntegrative Neurobiology of Affiliation. New York Academyof Sciences, New York.

Carver, C.S., Scheier, M.F., Weintraub, J.K., 1989. Assessingcoping strategies: a theoretically based approach. J. Pers.Social Psychol. 56, 267283.

Cohen, S., McKay, G., 1984. Social support, stress, and thebuffering hypothesis: a theoretical analysis. In: Baum, A.,

.Taylor, S.E., Singer, J.E. Eds. , Handbook of Psychologyand Health. Erlbaum, Hillsdale, NJ, pp. 253267.

Eccles, J.S., Lord, S.E., Roeser, R.W., Barber, B.L. et al., 1997.The association of school transitions in early adolescencewith developmental trajectories through high school. In:

.Schulenberg, J., Maggs, J.L., Hurrelmann, K. Eds. , HealthRisks and Developmental Transitions During Adolescence.Cambridge University Press, New York, NY, USA, pp.283320.

Ernst, J.M., Cacioppo, J.T., 1999. Lonely hearts: Psychologicalperspectives on loneliness. Appl. Preventive Psychol. 8,122.

Gardner, W.L., Gabriel, S., Diekman, A.B., in press. Interper-sonal processes. In: Cacioppo, J.T., Tassinary, L.G., Bernt-

.son, G.G. Eds. , Handbook of Psychophysiology. Cam-bridge University Press, New York.

Gardner, W.L., Pickett, C., Brewer, M.B., in review. Socialexclusion and selective memory: How the need to belongaffects memory for social information.

Glaser, R., Kiecolt-Glaser, J.K., Bonneau, R.H., Malarkey, W.et al., 1992. Stress-induced modulation of the immune


12/12


response to recombinant hepatitis B vaccine. Psychoso- .matic Medicine 54 1 , 2229.

Goldin-Meadow, S., Mylander, C., 1983. Gestural communica-tion in deaf children: noneffect of parental input on lan-guage development. Science 221, 372374.

Goldin-Meadow, S., Mylander, C., 1984. Gestural communica-tion in deaf children: the effects and noneffects of parentalinput on early language development. Monogr. Soc. Res.Child Dev. 49, 1121.

Gupta, V., Korte, C., 1994. The effects of a confidant and apeer group on the well-being of single elders. Int. J. AgingHum. Dev. 39, 293302.

House, J.S., Landis, K.R., Umberson, D., 1988. Social relation-ships and health. Science 241, 540545.

House, J.S., Robbins, C., Metzner, H.L., 1982. The associationof social relationships and activities with mortality:prospective evidence from the Tecumseh CommunityHealth Study. Am. J. Epidemiol. 116, 123140.

Hugdahl, K., 1995. Dichotic listening

probing temporallobe functional integrity. In: Davidson, R.J., Hugdahl, K. .Eds. , Brain Asymmetry. MIT Press, Cambridge, MA, pp.123156.

James, W., 1890r1950. Principles of Psychology. Dover, NewYork.

Kiecolt-Glaser, J.K., Garner, W., Speicher, C.E., Penn, G.,Glaser, R., 1984. Psychosocial modifiers of immunocompe-tence in medical students. Psychosom. Med. 46, 714.

Kirschbaum, C., Hellhammer, D.H., 1994. Salivary cortisol inpsychoneuroendocrine research: recent developments andapplications. Psychoneuroendocrinology 19, 313333.

Liu, D., Diorio, J., Tannenbaum, B., Caldji, C., Francis, D.,Freedman, A., Sharma, S., Pearson, D., Plotsky, P.M.,

Meaney, M.J., 1997. Maternal care, hippocampal glucocor-ticoid receptors, and hypothalamicpituitaryadrenal re-sponses to stress. Science 277, 16591662.

Meaney, M.J., Bhatnagar, S., Diorio, J., Larocque, S., Francis,D., ODonnell, D., Shanks, N., Sharma, S., Smythe, J., Viau,

V., 1993. Cellular Molecular Neurobiol. 13, 321

347.McEwen, B.S., Stellar, E., 1993. Stress and the individual:

mechanisms leading to disease. Arch. Internal Med. 153,20932101.

Munck, A., Guyre, P.M., Holbrook, N.J., 1984. Physiologicalfunction of glucocorticoids in stress and their relation topharmacological actions. Endocr. Rev. 5, 2544.

Perkins, H.W., 1991. Religious commitment, Yuppie values,and well-being in post collegiate life. Rev. Religious Res.32, 244251.

Russell, D., Peplau, L.A., Cutrona, C.E., 1980. The RevisedUCLA Loneliness Scale: concurrent and discriminant valid-ity evidence. J. Pers. Social Psychol. 39, 472480.

Saab, P.G., Matthews, K.A., Stoney, C.M., McDonald, R.J.,1989. Premenopausal and postmenopausal women differ intheir cardiovascular and neuroendocrine responses to be-havioral stressors. Psychophysiology 26, 270280.

Uchino, B.N., Cacioppo, J.T., Kiecolt-Glaser, J.K., 1996. Therelationship between social support and physiologicalprocess: A review with emphasis on underlying mechanismsand implications for health. Psychol. Bull. 119, 488531.

Uvnas-Mosberg, K., 1997. Physiological and endocrine effectsof social contact. Ann. N.Y. Acad. Sci. 807, 146163.

Van Cauter, E., 1990. Diurnal and ultradian rhythms in hu-man endocrine function: a minireview. Hormones Res. 34,4553.

Wilson, E.O., 1998. Consilience. Alfred A. Knopf, New York.

Aging and Loneliness

Documents

Transcript of Aging and Loneliness