Journal of Affective Disorders 135 (2011) 384–388

Contents lists available at ScienceDirect

Journal of Affective Disorders

j ourna l homepage: www.e lsev ie r.com/ locate / j ad

Brief report

Social rhythms and vulnerability to bipolar disorder

Ben Bullock a,⁎, Fiona Judd b, Greg Murray c

a Australian Catholic University, Melbourne, Australiab University of Melbourne, Melbourne, Australiac Swinburne University of Technology, Melbourne, Australia

a r t i c l e i n f o

⁎ Corresponding author at: School of PsychologyUniversity, 115 Victoria Parade, Fitzroy, Victoria 3065,9953 3127; fax: +61 3 9953 3205.

E-mail address: ben.bullock@acu.edu.au (B. Bulloc

0165-0327/$ – see front matter © 2011 Elsevier B.V.doi:10.1016/j.jad.2011.06.006

a b s t r a c t

Article history:Received 17 February 2011Received in revised form 3 June 2011Accepted 3 June 2011Available online 25 June 2011

Background: Reduced social rhythmicity is a commonly reported feature of bipolar disorderwhich may extend to non-clinical populations at risk of bipolar disorder. The aim of the currentstudy was to investigate social rhythms across three groups of participants; a clinical group ofbipolar disorder outpatients, and two non-clinical groups with high- and low-vulnerability tobipolar disorder, respectively. It was expected that reduced social rhythmicity woulddifferentiate the clinical group from the low-vulnerability group, but not the high vulnerabilitygroup.

Methods: Non-clinical participants were selected on the basis of scores derived from theGeneral Behaviour Inventory and allocated to groups of high (n=36) and low (n=36) traitvulnerability to bipolar disorder. The clinical group (n=15) were volunteers recruited from anoutpatient clinic. Participants completed a self-report social rhythmicity measure daily forseven consecutive days.

Results: One-way analysis of covariance (age) showed a significant overall effect for group, F(2,83)=4.67, pb .05. Post hoc comparisons revealed significant differences in social rhythmsbetween the two nonclinical groups only.

Limitations: The cross-sectional design of the study limits the strength of conclusions that canbe drawn.

Conclusions: The hypothesis was only partially supported. Consistent with expectations, thenon-clinical group with higher vulnerability to bipolar disorder recorded lower socialrhythmicity than the non-clinical group with lower vulnerability to bipolar disorder. Theclinical group however, did not differ in social rhythmicity from the lower vulnerability group.The findings may have consequences for the way in which vulnerability to bipolar disorder ismanaged.

© 2011 Elsevier B.V. All rights reserved.

Keywords:Bipolar disorderVulnerabilitySocial rhythms

1. Introduction

Social rhythmicity refers to the regularity with whichcommon social and lifestyle activities are conducted on adaily basis. Such activities include the timing of breakfast,starting work, having lunch, and going to bed at night. The

, Australian CatholicAustralia. Tel.: +61 3

k).

All rights reserved.

timing of social activities is thought to be partly driven by thebody's internal biological clock, the circadian system(Grandin et al., 2006). Equally however, the circadian systemis entrained by socio-behavioural feedback loops (Dijk andFranken, 2005; Mistlberger et al., 2000). In humans particu-larly, social zeitgebers (time referencing activities) areimportant behavioural entraining factors for circadian controlof the sleep/wake rhythm (Ehlers et al., 1988).

Reduced social rhythmicity is a commonly reportedfeature of bipolar disorder (BD). Ashman et al. (1999) forexample, showed that BD outpatients had significantly lower

385B. Bullock et al. / Journal of Affective Disorders 135 (2011) 384–388

regularity in their day-time activities than a matched controlgroup. Similarly, Shen et al. (2008a) showed that studentswith a lifetime history of cyclothymic disorder or BD-IIreported significantly fewer regular activities than studentswith no history of mood disorder. In the same study, reducedlifestyle regularity was found to be a significant predictor ofshorter time to mood episode onset.

The mechanism by which irregular social rhythms in BDare generated is thought to involve dysregulation of thecircadian system. An increasing array of clinical and pre-clinical evidence suggests that an unstable circadian systemmay be a key component of the core diathesis to the disorder(see Murray and Harvey, 2010). Many psychotherapies for BD,most prominently Interpersonal and Social Rhythm Therapy(IPSRT; Frank et al., 2000), feature a strong emphasis on thecircadian instability hypothesis in maintenance treatment ofBD by promoting regulation of sleep/wake patterns and dailyroutines. The therapy has proven to be effective in stabilisingsocial rhythms (Frank et al., 1997) and preventing relapse inBD outpatients (Frank et al., 2005). The effectiveness of thetherapy is attributed to redressing circadian instability viabetter adherence to social zeitgebers.

As well as being strongly associated with the clinicalmanifestation of BD, reduced social rhythmicity has also beendemonstrated in some populations with increased risk for BD.Meyer and Maier (2006) showed that students with hypo-manic personality traits (higher vulnerability to BD) exhib-ited lower regularity of daily activities than both a group ofstudents with a ‘rigid’ personality (higher vulnerability tomajor depressive disorder) and a group of controls (low onboth hypomanic and rigid personality). Shen et al. (2008b)however, reported mixed findings in their sample of studentsreporting cyclothymic traits. Increased lifestyle regularitywas associated with reduced day-to-day variability insymptoms, but was not consistently associated with hypo-manic or depressive symptomatology.

The aim of the current study was to investigate socialrhythmicity amongst three groups of participants — a clinicalgroup of outpatients diagnosed with BD-I, and two non-clinical groups with high- and low-vulnerability to BD,respectively. Social rhythmicity was not expected to differ-entiate between the BD outpatient group and the high-vulnerability group, but both groups were expected to havelower social rhythmicity than the low-vulnerability group.

2. Method

2.1. Participants

Non-clinical participants were selected from a pool of 484respondents on the basis of scores derived from the GeneralBehaviour Inventory (GBI; Depue et al., 1989). The 60 highestand 60 lowest scoring participants on the GBI were invited toparticipate in the study proper in order to represent high andlow trait vulnerability to BD, respectively. Acceptances werereceived from 72 participants who were allocated to groupsof high (High-GBI; n=36) and low (Low-GBI; n=36) traitvulnerability to BD. The clinical group (BD; n=15) werevolunteers recruited from a regional hospital outpatientservice. They were a clinically stable group, all in long-termcase-managed care, and taking mood-stabilising medication

(lithium, sodium valproate) amongst other drug treatments.The primary diagnosis for all clinical participants was BD-I,with no significant Axis-I comorbidity (schizoaffective disor-der and substance abuse disorder). Clinical diagnoses wereconfirmed via case file review by an experienced psychiatrist(author FJ) and completion of the Composite InternationalDiagnostic Inventory (CIDI-Auto; Robins et al., 1988). TheCIDI-Auto also confirmed that all participants werewell at thetime of the study with no reported evidence of depressiveand/or hypomanic diagnostic indicators.

2.2. Materials

For all three groups, hypomanic and depressive tendencieswere measured on the GBI, a 73-item self-report inventory ofBD-related experiences that incorporates elements of intensity,duration, and frequency of symptomatology. Total GBI score,and vulnerability to BD, are derived from a linear combinationof scores on two subscales, depression and hypomania. The GBIhas demonstrated adequate sensitivity to identify a full range ofboth syndromal and subsyndromal affective intensities, in bothclinical and non-clinical populations (e.g., Depue and Klein,1988; Reichart et al., 2004).

Social rhythmicitywas assessedusing thebrief Social RhythmMetric (SRM-5; Monk et al., 2002). The SRM-5 is a daily self-report log of the timing of five key activities throughout thewakingday. Threemorningactivities– (i)Getout of bed, (ii) Firstcontact with another person, (iii) Start work/housework/volunteer activities – and two evening activities – (i) Havedinner, (ii) Go to bed – form the basis of the instrument.Participants record the approximate time they commenced eachactivity on a pro-forma table. SRM-5 scores range from 1 to 7with higher scores indicating greater regularity. The SRM-5 is acondensed version of the original 17-item instrumentwhich hasdemonstrated reliability and validity in both clinical and non-clinical populations (Monk et al., 1990). However, in clinicalpopulations especially, the length and complexity of the 17-iteminstrument may be burdensome for some participants. Theshorter version has been shown to correlate highly with theoriginal version in both clinical and non-clinical populations, andhas also demonstrated sufficient sensitivity and specificity indetecting irregularity of social rhythms (Monk et al., 2002).

The CIDI-Auto was used to confirm BD diagnoses of theclinical group. It is a computerised diagnostic instrument thathas been shown to provide valid and reliable diagnosis ofmood disorders based on DSM-IV and ICD-10 criteria (Komitiet al., 2001; Peters and Andrews, 1995).

2.3. Procedure

Participants completed the SRM-5 for seven consecutivedays in a naturalistic setting. They were instructed tocomplete the form at the end of each day just prior to goingto bed. The forms were collected from participants at the endof the 7-day data collection period.

3. Results

Demographic data for the three groups are presented inTable 1. Also included in Table 1 are the mean GBI scores foreach group as well as the percentage of participants in each

Table 1Demographic data and GBI scores for the Low-GBI, High-GBI, and BD groups

Group n Age Female%

GBI score Cut-off%

M SD M SD

Low-GBI 36 21.03a 2.31 75 10.36a 6.17 0High-GBI 36 22.28a 2.93 67 107.97b 26.42 50BD 15 46.93b 12.43 53 112.67b 52.85 53

Note. GBI score = total score from the GBI using the Likert-scale scoringmethod (see Murray et al., 2007); cut-off = the percentage of participants ineach group scoring above recommended cut-off scores for identification olifetime BD diagnosis (see Depue et al., 1989). Different subscripts indicatesignificant group differences at pb .001.

SRM

5.00

4.50

4.00

3.50

3.00

2.50

2.00

4.23

Low-GBI

Fig. 1. Profile plot of SRM scores adjusted

386 B. Bullock et al. / Journal of Affective Disorders 135 (2011) 384–388

.

f

group scoring above the recommended cut-offs for identifi-cation of lifetime BD diagnoses (see Depue et al., 1989).

Table 1 shows that significant differences in age betweenthe three groups were apparent, Brown–Forsythe F (2,15)=58.53, pb .001. Games–Howell post hoc tests for unequalvariances revealed that the BD group was significantly olderthan the Low-GBI (pb .001) and High-GBI (pb .001) groups.Table 1 also shows that, as intended, significant differences inGBI score were apparent, Brown–Forsythe F (2,19)=75.47,pb .001. Games–Howell post hoc tests for unequal variancesrevealed that the Low-GBI group scored significantly lowerthan the High-GBI (pb .001) and BD (pb .001) groups. Therewas no significant difference in mean GBI score between theHigh-GBI and BD groups.

G

Hi

for age

Eighty-four out of 87 participants (96.6%) provided com-plete 7-day SRM records, whilst three participants failed torecord data on 1 day each. Mean SRM scores for the High-GBI(M=3.18, SD=1.21) and Low-GBI (M=3.93, SD=1.14)groups were higher than mean SRM data reported by Meyerand Maier (2006) in their non-clinical sample (M=2.19,SD=.75 and M=2.47, SD=.81, for their ‘bipolar risk’ andcontrol groups, respectively).Mean SRM score for the BD groupin the current study (M=4.27, SD=1.55)was also higher thanmean SRM data reported by Ashman et al. (1999) in theirclinical sample (M=2.67, SD=.54).

One-way analysis of covariance (ANCOVA) was used toassess group differences in SRM scores. Age was used as thecovariate due to the apparent stabilisation of social rhythmswith advancing age (Monk et al., 1997). From a statisticalperspective, age correlated positively and significantly withSRM scores for the sample (r=.29, pb .01). The ANCOVAshowed a significant overall effect for group with a moderateeffect size, F (2,83)=4.67, pb .05, partial η2=.10. Fig. 1displays the profile plot of mean SRM scores for each groupadjusted for age. Error bars represent the upper and lower95% confidence intervals of the adjusted means.

The profile plot shows that the BD group recorded thelowestmean SRM score adjusted for age, followed by the High-GBI group. The Low-GBI group recorded the highest mean SRMscore adjusted for age. The linear polynomial trend from lowestto highest putative vulnerability to BD approached significance,contrast estimate=.87, p=.09, however post hoc comparisonsof the adjusted means revealed significant differences in SRM

3.01

3.40

roup

BDgh-GBI

for the Low-GBI, High-GBI, and BD groups.

387B. Bullock et al. / Journal of Affective Disorders 135 (2011) 384–388

score between the Low-GBI and High-GBI groups only (pb .02,Bonferroni adjusted for three comparisons).

4. Discussion

The aim of the current study was to investigate socialrhythmicity amongst three groups of participants withvarying degrees of vulnerability to BD. Based on the circadianinstability hypothesis of BD, and using a model of traitvulnerability to the disorder, it was expected that a clinicalgroup of outpatients with BD-I and a non-clinical group ofstudents with high levels of BD-related temperament wouldshow lower social rhythmicity than a group of students withlow levels of BD-related temperament. The hypothesis wasonly partially supported. Although there were no significantdifferences in social rhythmicity between the High-GBI andBD groups (as predicted), significant differences between theBD group (highest putative vulnerability to BD) and the Low-GBI group (lowest putative vulnerability to BD) were notfound. There was however, a significant difference in socialrhythmicity between the High-GBI and Low-GBI groups.

The difference in social rhythmicity between the two non-clinical groups supports the findings of Meyer and Maier(2006) and Shen et al. (2008b). The current findings thereforeadd to an accumulating body of evidence showing reducedsocial rhythmicity amongst non-clinical populations atgreater risk of BD.

The clinical significance of the current findings is tem-pered somewhat by the lack of group differences between theBD group and the Low-GBI group. Although a linear trendshowed social rhythmicity decreased as a function ofincreasing levels of vulnerability to BD, differences betweenthe groups were not statistically significant. The relativelylarge amount of variation in SRM scores for the BD groupappears to explain much of this unexpected finding (seeFig. 1). Between-subject differences in age of onset, numberof previous episodes, and history of rapid cycling may explainsome of the variation in SRM scores for this group (see, forexample, Malkoff-Schwartz et al., 2000).

The study had several limitations. The cross-sectional designlimits the strength of conclusions that canbedrawn.Monitoringsocial rhythms prospectively amongst a non-clinical sample athigh risk of BDwould provide for a more rigorous investigationof the study's aims. A greater number of participants in the BDgroup tomatch the participant numbers of the two non-clinicalgroups would have been useful for the purposes of statisticalanalysis. More detailed background information on the clinicalgroup (e.g., age of onset, number of previous episodes, andhistory of rapid cycling) may have also aided interpretation ofstatistical outcomes. Finally, social rhythms were monitoredover 7 days. The validity and reliability of the SRM-5 instrumenthas not been evaluated over this brief time frame.

In conclusion, the investigation of social rhythms in threegroups of participants with varying degrees of vulnerability toBD found that non-clinical participants with higher vulnerabil-ity to BDhad significantly lower social rhythmicity thana groupof non-clinical participants with lower vulnerability to BD. Thisfindingwas consistentwith expectationsbasedon thecircadianinstability hypothesis of BD. Inconsistent with this hypothesishowever, was the finding that the non-clinical group withlower vulnerability to BDdid not differ in social rhythmicity to a

group of BD outpatients. Encouraging for future studies was atrend towards a negative linear relationship between decreas-ing social rhythmicity and increasing levels of vulnerability toBD. The outcomesmayhave important implications for thewaythat risk for BD is managed.

Role of funding sourceThe study was partly funded by a research development grant from

Swinburne University of Technology. The funding agency had no further rolein the study design, in the collection, analysis, and interpretation of data, inthe writing of the report, and in the decision to submit the paper forpublication.

Conflict of interestThe authors have no commercial associations that might pose a conflict

of interest in connection with this manuscript.

References

Ashman, S.B., Monk, T.H., Kupfer, D.J., Clark, C.H., Myers, F.S., Frank, E.,Leibenluft, E., 1999. Relationship between social rhythms and mood inpatients with rapid cycling bipolar disorder. Psychiatry Research 86, 1–8.

Depue, R.A., Klein, D.N., 1988. Identification of unipolar and bipolaraffective conditions in nonclinical and clinical populations by theGeneral Behavior Inventory. In: Dunner, D.L., Gershon, E.S., Barrett, J.E.(Eds.), Relatives at Risk for Mental Disorder. Raven Press, Ltd, NewYork, pp. 179–204.

Depue, R.A., Krauss, S., Spoont, M.R., Arbisi, P., 1989. General behaviourinventory identification of unipolar and bipolar affective conditions in anonclinical university population. Journal of Abnormal Psychology 98,117–126.

Dijk, D.J., Franken, P., 2005. Interaction of sleep homeostasis and circadianrhythmicity: dependent or independent systems? In: Kryger, M.H., Roth,T., Dement, W.C. (Eds.), Principles and Practice of Sleep Medicine.Elsevier/Saunders, Philadelphia, PA, pp. 418–434.

Ehlers, C.L., Frank, E., Kupfer, D.J., 1988. Social zeitgebers and biologicalrhythms: a unified approach to understanding the etiology of depres-sion. Archives of General Psychiatry 45, 948–952.

Frank, E., Hlastala, S., Ritenour, A., Houck, P., Ming Tu, X., Monk, T.H.,Mallinger, A.G., Kupfer, D.J., 1997. Inducing lifestyle regularity inrecovering bipolar disorder patients: results from the maintenancetherapies in bipolar disorder protocol. Biological Psychiatry 41,1165–1173.

Frank, E., Swartz, H.A., Kupfer, D.J., 2000. Interpersonal and social rhythmtherapy: managing the chaos of bipolar disorder. Biological Psychiatry48, 593–604.

Frank, E., Kupfer, D.J., Thase, M.E., Mallinger, A.G., Swartz, H.A., Fagiolini, A.M.,Grochocinski, V., Houck, P., Scott, J., Thompson, W., Monk, T.H., 2005.Two-year outcomes for interpersonal and social rhythm therapy inindividuals with bipolar I disorder. Archives of General Psychiatry 62,996–1004.

Grandin, L.D., Alloy, L.B., Abramson, L.Y., 2006. The social zeitgeber theory,circadian rhythms, and mood disorders: review and evaluation. ClinicalPsychology Review 26, 679–694.

Komiti, A., Jackson, H., Judd, F.K., Cockram, A.M., Kyrios, M., Yeatman, R.,Murray, G., Hordern, C., Wainwright, K., Allen, N.B., Singh, B., 2001. Acomparison of the Composite International Diagnostic Interview (CIDI-Auto) with clinical assessment in diagnosing anxiety and mooddisorders. The Australian and New Zealand Journal of Psychiatry 35,224–230.

Malkoff-Schwartz, S., Frank, E., Anderson, B.P., Hlastala, S.A., Luther, J.F.,Sherrill, J.T., Houck, P.R., Kupfer, D.J., 2000. Social rhythm disruption andstressful life events in the onset of bipolar and unipolar episodes.Psychological Medicine 30, 1005–1016.

Meyer, T.D., Maier, S., 2006. Is there evidence for social rhythm instability inpeople at risk for affective disorders? Psychiatry Research 30, 103–114.

Mistlberger, R.E., Antle, M.C., Glass, J.D., Miller, J.D., 2000. Behavioral andserotonergic regulation of circadian rhythms. Biological Rhythm Re-search 31, 240–283.

Monk, T.H., Reynolds, C.F., Kupfer, D.J., Hoch, C.C., Carrier, J., Houck, P.R., 1997.Differences over the life span in daily life-style regularity. ChronobiologyInternational 14, 295–306.

Monk, T.H., Flaherty, J.F., Frank, E., Hoskinson, K., Kupfer, D.J., 1990. The socialrhythm metric: an instrument to quantify the daily rhythms of life. TheJournal of Nervous and Mental Disease 178, 120–126.

388 B. Bullock et al. / Journal of Affective Disorders 135 (2011) 384–388

Monk, T.H., Frank, E., Potts, J.M., Kupfer, D.J., 2002. A simple way to measurelifestyle regularity. Journal of Sleep Research 11, 183–190.

Murray, G., Harvey, A., 2010. Circadian rhythms and sleep in bipolar disorder.Bipolar Disorders 12, 459–472.

Peters, L., Andrews, G., 1995. Procedural validity of the computerized versionof the Composite International Diagnostic Interview (CIDI-Auto) in theanxiety disorders. Psychological Medicine 25, 1269–1280.

Reichart, C.G., Van Der Ende, J., Wals, M., Hillegers, M.H.J., Ormel, J., Nolen,W.A.,Verhulst, F.C., 2004. The use of the GBI in a population of adolescentoffspring of parents with a bipolar disorder. Journal of Affective Disorders80, 263–267.

Robins, L.N., Wing, J., Wittchen, H.U., Helzer, J.E., Babor, T.F., Burke, J., Farmer,A., Jablenski, A., Pickens, R., Regier, D.A., 1988. The CompositeInternational Diagnostic Interview. An epidemiologic Instrument suit-able for use in conjunction with different diagnostic systems and indifferent cultures. Archives of General Psychiatry 45, 1069–1077.

Shen, G.H.C., Alloy, L.B., Abramson, L.Y., Sylvia, L.G., 2008a. Social rhythmregularity and the onset of affective episodes in bipolar spectrumindividuals. Bipolar Disorders 10, 520–529.

Shen, G.H.C., Sylvia, L.G., Alloy, L.B., Barrett, F., Kohner, M., Iacoviello, B., Mills,A., 2008b. Lifestyle regularity and cyclothymic symptomatology. Journalof Clinical Psychology 64, 482–500.