Research report
Seasonality and circadian phase delay: prospective evidence that winter lowering of mood is associated with a shift towards Eveningness

https://doi.org/10.1016/S0165-0327(02)00059-9Get rights and content

Abstract

Background: Lewy’s phase shift hypothesis (PSH) of seasonal affective disorder (SAD) asserts that winter depression is caused by an abnormal delay of circadian phase in winter. The aim of this study was to further investigate the PSH by testing the novel hypothesis that winter pattern seasonality of mood in the general population is associated with a phase delay of circadian rhythms in winter. Methods: Using a random community sample of 244 adults in Melbourne, seasonality of mood was measured prospectively as the within-subject difference between self-reports of behavioural engagement in summer and winter across 3 years. The Horne–Ostberg Morningness–Eveningness questionnaire (MEQ) was used as a self-report estimate of circadian phase. Phase delay in winter was measured as the within-subject difference between summer and winter estimates of circadian phase, with relative shifts towards Eveningness being interpreted as phase delay. Results: As expected, a significant association was found between winter pattern seasonality of mood and within-subject phase delay in winter (r=0.17, P<0.01). Limitations: Circadian phase was operationalised on the MEQ. While a number of studies have shown the MEQ to correlate with physiological measures of circadian phase, the findings await replication using a more direct measure of the circadian system. Conclusions: A positive association was found between lowered mood in winter and winter phase delay amongst a random community sample. This provisional finding adds support to Lewy’s PSH by demonstrating that phase delay may be causally important not just in clinical cases of SAD, but across the continuum of mood seasonality.

Introduction

Chronobiological models have always been prominent in theorising about seasonal affective disorder (SAD: Rosenthal et al., 1984). Amongst these models, Lewy’s phase-shift hypothesis (PSH) has stimulated the most research and is a major stimulus for circadian research into depression generally (Van Dongen et al., 1997). Lewy’s proposition is that the depressions seen in SAD are a result of a phase delay of the endogenous circadian oscillator relative to the sleep–wake cycle (Lewy et al., 1989). Bright light administered in the morning ought therefore to be therapeutic because it corrects this abnormal delay by phase advancing the endogenous oscillator (according to the human phase response curve, Pittendrigh, 1989). Bright light administered in the evening should not be therapeutic, because it ought to promote an exacerbating phase delay. A critical test of the PSH, then, is whether morning treatment is superior to evening treatment.

The empirical status of the PSH has varied over the past decade (for a review, see Terman, 1998). The hypothesis remains viable, however, and two large-n studies have reported the therapeutic superiority of morning over evening light, as predicted by the PSH (Lewy et al., 1998, Terman et al., 1998). In a recent commentary, Avery (1998) concluded that bright light is the first line treatment for SAD, it is most effective when administered in the morning and that these findings are consistent with Lewy’s hypothesis.

Although SAD is described in DSM as a variant of categorical mood disorder, it is typically understood as the extreme end of a continuum of normative seasonal variation in mood and behaviour (Gordon et al., 1999, Harmatz et al., 2000). As expressed by Hardin and colleagues, “since winter-seasonal affective disorder (winter-SAD) was first described, it has become apparent that the tendency for people to experience seasonal changes in mood and behaviour (seasonality) is not confined to severely affected individuals, but appears to vary in degree across the normal population” (Hardin et al., 1991, p.75). While it has been suggested that the tendency towards seasonal variation in mood may have more empirical support than categorical SAD (Bauer and Dunner, 1993), the construct of seasonality is still poorly defined and understood (Murray, 2001, Murray et al., 2000).

Blehar and Lewy (1990) have suggested that the PSH might extend to non-clinical seasonal variation in mood, with the degree of seasonality being associated with the degree of wintertime phase delay. These authors note that the tendency towards a phase delay in winter can be understood either normatively (between-subjects) or ipsatively (within-subjects). That is, the critical predictor of winter pattern seasonality might be (1) the extent of delay relative to other people in winter, and/or (2) the extent of delay relative to one’s own phase in summer. The present study provides an initial test of this two-part hypothesis.

There is growing evidence that an estimate of circadian phase can be derived from self-report. A number of studies have found that self-report of diurnal type (Morningness–Eveningness, or M–E) is associated with physiological rhythms which have a marked circadian component (Froberg, 1977, Horne and Ostberg, 1977, Lacoste and Wetterberg, 1993, Wilson, 1990). Gibertini et al. (1999), for example, found a significant association between M–E type and the time of peak melatonin (χ2(4)=12.7, P=0.01). Similarly, Baehr et al. (2000) found that M–E score was a significant predictor (r=−0.52, n=182, P<0.05) of the phase of the circadian temperature rhythm (measured as ambulatory core body temperature over 6 days).

The Horne–Ostberg Morningness–Eveningness Questionnaire (MEQ: Horne and Ostberg, 1976, Horne and Ostberg, 1977) was designed as a trait measure of diurnal type. However, consistent with the data above, a number of previous studies have interpreted shifts towards Eveningness on the MEQ as indicating a phase delay (Drennan et al., 1991, Elmore et al., 1993, Kloppel, 1982, Lacoste and Wirz-Justice, 1989).

Lewy’s phase delay hypothesis of winter SAD can be extended to make predictions about the chronobiological correlates of mood seasonality in the general population. Specifically, it can be postulated that winter lowering of mood in the normal population might correlate with either a normative circadian phase delay in winter, or an ipsative phase delay in winter relative to summer. This proposition is important in two ways. First, it offers a test of an important extension of the PSH into the normal population. Secondly, it provides a cross-validation of the construct of seasonality itself, which is yet to receive adequate attention. Growing evidence that circadian phase can be estimated from self-reports of Morningness–Eveningness suggests a methodology for an initial test of the proposition. The current study used a longitudinal survey design to collect mood data from a normal community sample across winter and summer for 3 years.1 In the final two waves of the design, participants’ current circadian phase in summer and winter was estimated using the MEQ. Normative phase delay was measured as the MEQ score in winter, and ipsative phase delay was measured as the within-subject difference between summer and winter MEQ scores.

Section snippets

Sample

The study was conducted in Melbourne, the most southerly metropolis on mainland Australia. The climate of Melbourne is temperate (Reid, 1995), and the photoperiod ranges from 09:32 h at the winter solstice (21 June) to 14:47 h at the summer solstice (22 December). The winter photoperiod, and the seasonal difference in photoperiod, are therefore similar to that seen in Montgomery County, Washington, DC, the locale of the seminal epidemiological study of SAD (Kasper et al., 1989).

Subjects were

Response rates and representativeness of the sample

One thousand and eighty questionnaires were originally mailed, to which 562 individuals (52.0%) responded at the first wave. Of the 562 individuals who returned completed questionnaires at wave 1, 335 (59.6%) completed all six waves.

Comparison with census data for the district found that individuals around the age of 50 were somewhat over-represented in the present sample, while individuals in their mid-20s were relatively under-represented. The distribution of gender in the sample (55.3%

Discussion

The study’s two-part hypothesis was supported. Prospective winter pattern seasonality (within-subject decrease in BE levels in winter) showed a small but significant positive correlation with normatively elevated Eveningness in winter and a within-subject shift towards Eveningness in winter relative to summer. These findings applied beyond the year in which MEQ measurements were taken, suggesting that the relationship was not an artefact of participants’ completing MEQ and mood reports at the

Summary

Longitudinal survey methodology was used to test the hypothesis that normal population seasonal variation of mood is associated with a circadian phase delay in winter. Consistent with Lewy’s PSH for SAD, small but significant correlations between seasonal mood change and seasonal phase change were found amongst 244 respondents in a random community sample. These findings constitute support for an extension of the PSH into the normal population, and add support to the validity of the seasonality

References (45)

  • M.R. Spoont et al.

    Dimensional measurement of seasonal variation in mood and behavior

    Psychiatry Res.

    (1991)
  • I. Tankova et al.

    Circadian typology and individual differences: A review

    J. Pers. Ind. Diff.

    (1994)
  • G.D. Wilson

    Personality, time of day, and arousal

    J. Pers. Ind. Diff.

    (1990)
  • D.H. Avery

    A turning point for seasonal affective disorder and light therapy research?

    Arch. Gen. Psychiatry

    (1998)
  • E.K. Baehr et al.

    Individual differences in the phase and amplitude of the human circadian temperature rhythm: with an emphasis on morningness-eveningness

    J. Sleep Res.

    (2000)
  • M.C. Blehar et al.

    Seasonal mood disorders: consensus and controversy

    Psychopharmacol. Bull.

    (1990)
  • L.A. Clark et al.

    Theoretical and empirical issues in differentiating depression from anxiety

  • R.A. Depue et al.

    Dopamine functioning in the behavioral facilitation system and seasonal variation in behavior: normal population and clinical studies

  • R.A. Depue et al.

    A two-dimensional threshold model of seasonal bipolar affective disorder

  • D.A. Dillman
  • S.K. Elmore et al.

    Body temperature and diurnal type in women with seasonal affective disorder

    Health Care Women Int.

    (1993)
  • M.G. Harmatz et al.

    Seasonal variation of depression and other moods: a longitudinal approach

    J. Biol. Rhythms

    (2000)
  • Cited by (0)

    View full text