Dimensional measurement of seasonal variation in mood and behavior

Harnessing Consumer Wearable Digital Biomarkers for Individualized Recognition of Postpartum Depression Using the All of Us Research Program Data Set: Cross-Sectional Study

BACKGROUND

Postpartum depression (PPD) poses a significant maternal health challenge. The current approach to detecting PPD relies on in-person postpartum visits, which contributes to underdiagnosis. Furthermore, recognizing PPD symptoms can be challenging. Therefore, we explored the potential of using digital biomarkers from consumer wearables for PPD recognition.

OBJECTIVE

The main goal of this study was to showcase the viability of using machine learning (ML) and digital biomarkers related to heart rate, physical activity, and energy expenditure derived from consumer-grade wearables for the recognition of PPD.

METHODS

Using the All of Us Research Program Registered Tier v6 data set, we performed computational phenotyping of women with and without PPD following childbirth. Intraindividual ML models were developed using digital biomarkers from Fitbit to discern between prepregnancy, pregnancy, postpartum without depression, and postpartum with depression (ie, PPD diagnosis) periods. Models were built using generalized linear models, random forest, support vector machine, and k-nearest neighbor algorithms and evaluated using the κ statistic and multiclass area under the receiver operating characteristic curve (mAUC) to determine the algorithm with the best performance. The specificity of our individualized ML approach was confirmed in a cohort of women who gave birth and did not experience PPD. Moreover, we assessed the impact of a previous history of depression on model performance. We determined the variable importance for predicting the PPD period using Shapley additive explanations and confirmed the results using a permutation approach. Finally, we compared our individualized ML methodology against a traditional cohort-based ML model for PPD recognition and compared model performance using sensitivity, specificity, precision, recall, and F1-score.

RESULTS

Patient cohorts of women with valid Fitbit data who gave birth included <20 with PPD and 39 without PPD. Our results demonstrated that intraindividual models using digital biomarkers discerned among prepregnancy, pregnancy, postpartum without depression, and postpartum with depression (ie, PPD diagnosis) periods, with random forest (mAUC=0.85; κ=0.80) models outperforming generalized linear models (mAUC=0.82; κ=0.74), support vector machine (mAUC=0.75; κ=0.72), and k-nearest neighbor (mAUC=0.74; κ=0.62). Model performance decreased in women without PPD, illustrating the method's specificity. Previous depression history did not impact the efficacy of the model for PPD recognition. Moreover, we found that the most predictive biomarker of PPD was calories burned during the basal metabolic rate. Finally, individualized models surpassed the performance of a conventional cohort-based model for PPD detection.

CONCLUSIONS

This research establishes consumer wearables as a promising tool for PPD identification and highlights personalized ML approaches, which could transform early disease detection strategies.

Harnessing consumer wearable digital biomarkers for individualized recognition of postpartum depression using the All of Us Research Program dataset

Postpartum depression (PPD), afflicting one in seven women, poses a major challenge in maternal health. Existing approaches to detect PPD heavily depend on in-person postpartum visits, leading to cases of the condition being overlooked and untreated. We explored the potential of consumer wearable-derived digital biomarkers for PPD recognition to address this gap. Our study demonstrated that intra-individual machine learning (ML) models developed using these digital biomarkers can discern between pre-pregnancy, pregnancy, postpartum without depression, and postpartum with depression time periods (i.e., PPD diagnosis). When evaluating variable importance, calories burned from the basal metabolic rate (calories BMR) emerged as the digital biomarker most predictive of PPD. To confirm the specificity of our method, we demonstrated that models developed in women without PPD could not accurately classify the PPD-equivalent phase. Prior depression history did not alter model efficacy for PPD recognition. Furthermore, the individualized models demonstrated superior performance compared to a conventional cohort-based model for the detection of PPD, underscoring the effectiveness of our individualized ML approach. This work establishes consumer wearables as a promising avenue for PPD identification. More importantly, it also emphasizes the utility of individualized ML model methodology, potentially transforming early disease detection strategies.

Seasonality and personality: a prospective investigation of Five Factor Model correlates of mood seasonality

The aim of the present study was to explore winter pattern seasonality of mood by investigating its Five Factor Model (FFM) correlates. The study was an advance on existing research in that seasonality was measured not as a retrospective self‐description of mood variation, but as a prospective pattern of current mood states in winter and summer (across two years). Based on contemporary theorizing about the structure and function of mood, Positive Affect (PA) was the mood construct selected for measurement. A seasonality score was calculated for each participant as the difference between summer mood levels and winter mood levels. The NEO‐FFI was administered at all four waves. Three hundred and three respondents from a random community sample in Melbourne, Australia, provided complete data. Regression analyses found that the adaptive trait Openness to Experience (O) was specifically associated with the tendency towards lowered mood in winter relative to summer. The vulnerability trait Neuroticism (N) was not a directional predictor of winter pattern seasonality, but was the sole predictor of absolute seasonality score. Findings are discussed in terms of two affective processes that may be involved in mood variation across the seasonal time frame: adaptive environmental sensitivity and endogenous mood variability. Copyright © 2002 John Wiley & Sons, Ltd.

Out of the dark, into the light: The impact of social exclusion on judgments of darkness and brightness

Based on theories of grounded cognition, we assumed that the experience of social exclusion is grounded in a concept of darkness. Specifically, we hypothesized that social exclusion causes perceptual judgments of darkness and a preference for brightness as a compensatory response. To investigate these hypotheses, we conducted four studies using different manipulations and measurements. In Studies 1a and 1b, excluded participants judged a picturized room as darker and drew more attention to its brightest part than included participants. In Study 2, excluded participants judged a surface as darker and decided for brighter clothing than included participants. In Study 3, excluded participants judged their lab room as darker and expressed a higher preference for brightness than included participants. Providing consistent support for our hypotheses, these findings confirm the idea that the experience of social exclusion is grounded in multiple ways that share a common representational system.

Moderators of Cognitive Therapy and Bright Light Therapy Effects on Depressive Symptoms in Patients with Breast Cancer

BACKGROUND

Cognitive therapy (CT) and bright light therapy (BLT) have been found to be effective to treat depressive symptoms in breast cancer patients. No study has investigated the baseline patients' characteristics that are associated with better outcomes with CT vs. BLT in this population. This study aimed to assess, in breast cancer patients, the moderating role of eight clinical variables on the effects of CT and BLT on depressive symptoms.

METHODS

This is a secondary analysis of a randomized controlled trial conducted in 59 women who received an 8-week CT or BLT and completed questionnaires evaluating depression and possible moderating variables.

RESULTS

Patients benefited more from BLT when they had no prior history of major depressive disorder, higher depression scores on the Hospital Anxiety and Depression Scale (HADS-D) at baseline, a greater initial preference for BLT, and when they received BLT during spring or summer. Patients benefited more from CT when they had a lower initial preference for receiving CT, higher depression scores on the HADS-D, and seasonal depressive symptoms.

CONCLUSIONS

Although replication is needed, findings of this study suggest the existence of different profiles of patients more likely to benefit from CT and BLT.

TRIAL REGISTRATION

NCT01637103 https://clinicaltrials.gov/ct2/show/NCT01637103.

Seasonal variation of spontaneous blink rate and beta EEG activity

Seasonal variations of the photoperiod have been shown to regulate biological and behavioral functions, with also effects on clinical symptom and course of several psychiatric conditions. Although melatonin is considered the principal signal used to transmit informations about the light and dark cycle, a dopamine (DA) role in regulating seasonal changes has been suggested. Few studies have addressed a seasonal pattern of dopamine, and human studies have been conducted on inter-subject differences, comparing measures obtained during fall-winter with those of spring-summer. We studied within-subject seasonal changes of blink rate (BR), a indirect marker of central DA activity, in 26 normal subjects (15 females and 11 males, mean age: 24.7 ± 4.0) during winter, spring, summer and fall. Occipital EEG activity and subjective measures of vigilance and mood were also assessed to account for variations on arousal and fatigue. A significant seasonal effect was found for BR, with higher rate in summer, and for EEG beta activity, with higher activity in spring and summer. Subjective fatigue was found higher in winter. According to our data, it is possible that higher BR and increased EEG beta activity result by an arousal activation sustained by dopamine systems during the months with a long photoperiod.

Physical Activity Levels and Well-Being in Older Adults

The objective of this study was to identify the interconnectedness of different intensity levels of physical activity and psychological (life satisfaction and positive affect) and physical (physical health) well-being. Participants were from the National Study of Midlife in the United States with assessments in 2004 and aged 25 to 74 living in the United States were included in the analyses. We conducted bivariate correlations to examine significant relationships among the study variables. In addition, after multicollinearity among the independent variable was checked, a series of hierarchical regression analyses with physical health, positive affect, and life satisfaction as criterion variables were conducted. The results showed that light physical activities were positively associated with physical health and life satisfaction in summer, whereas light physical activities and all dependent variables were positively correlated in winter. Furthermore, engaging in moderate physical activities was positively related only with physical health. Meanwhile, vigorous physical activities were not associated with life satisfaction, physical health, and positive affect in summer and winter.

Sunshine-exposure variation of human striatal dopamine D(2)/D(3) receptor availability in healthy volunteers

BACKGROUND

In addition to the serotonergic system, the central dopaminergic system has been reported to be correlated with seasonality. The aim of this study was to explore the difference in striatal dopamine D(2)/D(3) receptor availability between healthy volunteers who had a high-sunshine exposure and those who had a low exposure.

METHODS

Sixty-eight participants were enrolled, and those in the upper and lower quartiles in terms of sunshine exposure were categorized into high- (n = 17) and low-sunshine-exposure (n = 18) subgroups. Single photon emission computed tomography with [(123)I] iodo-benzamide was used to measure striatal dopamine D(2)/D(3) receptor availability.

RESULTS

Striatal dopamine D(2)/D(3) receptor availability was significantly greater in the subjects with high-sunshine exposure than in those with low-sunshine exposure (F = 7.97, p = 0.01) after controlling for age, sex, and smoking status.

LIMITATIONS

Different subjects were examined at different time points in our study. In addition, the sex and tobacco use distributions differed between groups.

CONCLUSION

The central dopaminergic system may play a role in the neurobiological characteristics of sunshine-exposure variation.

Circadian photoreception: ageing and the eye's important role in systemic health

AIM

To analyse how age-related losses in crystalline lens transmittance and pupillary area affect circadian photoreception and compare the circadian performance of phakic and pseudophakic individuals of the same age.

METHODS

The spectral sensitivity of circadian photoreception peaks in the blue part of the spectrum at approximately 460 nm. Photosensitive retinal ganglion cells send unconscious information about environmental illumination to non-visual brain centres including the human body's master biological clock in the suprachiasmatic nuclei. This information permits human physiology to be optimised and aligned with geophysical day-night cycles using neural and hormonal messengers including melatonin. Age-related transmittance spectra of crystalline lenses and photopic pupil diameter are used with the spectral sensitivity of melatonin suppression and the transmittance spectra of intraocular lenses (IOLs) to analyse how ageing and IOL chromophores affect circadian photoreception.

RESULTS

Ageing increases crystalline lens light absorption and decreases pupil area resulting in progressive loss of circadian photoreception. A 10-year-old child has circadian photoreception 10-fold greater than a 95-year-old phakic adult. A 45-year-old adult retains only half the circadian photoreception of early youth. Pseudophakia improves circadian photoreception at all ages, particularly with UV-only blocking IOLs which transmit blue wavelengths optimal for non-visual photoreception.

CONCLUSIONS

Non-visual retinal ganglion photoreceptor responses to bright, properly timed light exposures help assure effective circadian photoentrainment and optimal diurnal physiological processes. Circadian photoreception can persist in visually blind individuals if retinal ganglion cell photoreceptors and their suprachiasmatic connections are intact. Retinal illumination decreases with ageing due to pupillary miosis and reduced crystalline lens light transmission especially of short wavelengths. Inadequate environmental light and/or ganglion photoreception can cause circadian disruption, increasing the risk of insomnia, depression, numerous systemic disorders and possibly early mortality. Artificial lighting is dimmer and less blue-weighted than natural daylight, contributing to age-related losses in unconscious circadian photoreception. Optimal intraocular lens design should consider the spectral requirements of both conscious and unconscious retinal photoreception.

VDR gene variants associate with cognitive function and depressive symptoms in old age

Vitamin D has been recently implicated in brain function. Our objective was to test whether genetic variance in the vitamin D receptor (VDR) gene is associated with cognitive functioning and depressive symptoms in old age. The study was carried out in the prospective population-based Leiden 85-plus Study. All 563 participants of the study were genotyped for Cdx-2, FokI, BsmI, ApaI and TaqI polymorphisms in the VDR gene. Our data revealed an overall worse performance on tests measuring cognitive functioning for carriers of BsmI (p=0.013) and TaqI (p=0.004) polymorphisms, and of haplotype 2 (BAt) (p=0.004). In contrast, carriers of ApaI variant-allele and of haplotype 1 (baT) had better cognitive functioning together with less depressive symptoms. These associations could not be explained by differences in calcium levels, and by selective survival, since no associations between the VDR gene variants and calcium levels and mortality were observed. In conclusion, our results show that genetic variance in the VDR gene influences the susceptibility to age-related changes in cognitive functioning and in depressive symptoms.

Vitamin D deficiency is associated with low mood and worse cognitive performance in older adults

BACKGROUND

Vitamin D deficiency is common in older adults and has been implicated in psychiatric and neurologic disorders. This study examined the relationship among vitamin D status, cognitive performance, mood, and physical performance in older adults.

METHODS

A cross-sectional group of 80 participants, 40 with mild Alzheimer disease (AD) and 40 nondemented persons, were selected from a longitudinal study of memory and aging. Cognitive function was assessed using the Short Blessed Test (SBT), Mini-Mental State Exam (MMSE), Clinical Dementia Rating (CDR; a higher Sum of Boxes score indicates greater dementia severity), and a factor score from a neuropsychometric battery; mood was assessed using clinician's diagnosis and the depression symptoms inventory. The Physical Performance Test (PPT) was used to measure functional status. Serum 25-hydroxyvitamin D levels were measured for all participants.

RESULTS

The mean vitamin D level in the total sample was 18.58 ng/mL (standard deviation: 7.59); 58% of the participants had abnormally low vitamin D levels defined as less than 20 ng/mL. After adjusting for age, race, gender, and season of vitamin D determination, vitamin D deficiency was associated with presence of an active mood disorder (odds ratio: 11.69, 95% confidence interval: 2.04-66.86; Wald chi(2) = 7.66, df = 2, p = 0.022). Using the same covariates in a linear regression model, vitamin D deficiency was associated with worse performance on the SBT (F = 5.22, df = [2, 77], p = 0.044) and higher CDR Sum of Box scores (F = 3.20, df = [2, 77], p = 0.047) in the vitamin D-deficient group. There was no difference in performance on the MMSE, PPT, or factor scores between the vitamin D groups.

CONCLUSIONS

In a cross-section of older adults, vitamin D deficiency was associated with low mood and with impairment on two of four measures of cognitive performance.

Mood swings in patients with anxiety disorders compared with normal controls

BACKGROUND

About 70-80% of patients with anxiety syndromes suffer from depression. Mood variability including hypomania, cyclothymia and hyperthymia have been described in 40-50% of patients with depression. There is an emerging literature that such variability could also characterize anxiety disorders. The aim of this study was to visually document and quantify mood variability in patients with anxiety disorders.

METHODS

Twenty patients with anxiety disorders and 22 normal control subjects completed two visual analogue scales (VAS) on depressed mood and high mood, twice per day for 14 days. The Beck Depression Inventory and the Altman Self-Rating Mood Scale were used for concurrent validity.

RESULTS

On the VAS, patients showed higher levels of depressed and high moods, and greater mood variability than the controls. Variability of depressed and high moods was highly correlated.

LIMITATIONS

This was a relatively small sample from a single center. Patients were selected by convenience and were under treatment. The control subjects were not interviewed.

CONCLUSIONS

Subsyndromal mood variability in patients with anxiety disorders can be visually depicted and quantified. The mood variability of patients with anxiety disorders who also complain of mood swings is greater than the mood fluctuations described by normal subjects.

The Seasonal Pattern Assessment Questionnaire as a measure of mood seasonality: a prospective validation study

The aim of this study was to test the validity of the Seasonal Pattern Assessment Questionnaire (SPAQ) as a measure of winter pattern seasonality of mood in the general population. The criterion was four repeated measurements of mood, collected in summer and winter each year for 2 years in a longitudinal questionnaire study. To account for the imperfect reliability of SPAQ reports, the SPAQ was completed on two occasions, and aggregate SPAQ variables created. Three variables from the SPAQ were tested for their correlation with prospectively measured lowering of mood in winter: seasonal pattern type, Global Seasonality Score (GSS) and self-reported problems with seasonal variation. The sampling frame was the electoral roll in a suburban region of Melbourne, Australia. Complete four-wave data were obtained from 304 respondents (53.9% female). Seasonal pattern type (self-report as feeling worst in winter months) was the only SPAQ variable to show a significant association with prospectively measured winter pattern seasonality of mood. It is provisionally concluded that the SPAQ can function as a measure of seasonality of mood in the normal population, but that validity has only been confirmed for the relatively gross variable of seasonal pattern type.

Measuring seasonality: psychometric properties of the Seasonal Pattern Assessment Questionnaire and the Inventory for Seasonal Variation

There is a general consensus that seasonal changes in mood, behavior and vegetative functions vary in magnitude across the general population. Studying the dimension of seasonality requires reliable and valid measures. This study examined and compared the psychometric properties of the Seasonal Pattern Assessment Questionnaire (SPAQ) and the Inventory for Seasonal Variation (ISV) in 148 college students; 73 also completed a retest 2 months later. Results indicated good psychometric properties for both scales in terms of score distributions, test-retest reliability, internal consistency, factor structure and item-latent trait relationships using item response theory. Given the long history of its use and some small psychometric and simplicity advantages for the SPAQ, its continued use as a dimensional measure of seasonality is indicated.

A longitudinal investigation of seasonal variation in mood

A prospective panel study was conducted to measure seasonality of mood in a random community sample in Melbourne, Australia (N = 245). Based on research into the structure of human mood, it was predicted that a lowering of mood in winter relative to summer would be observed in positive affect (PA) and behavioral engagement (BE), but not negative affect (NA). These variables were measured across summer and winter for 3 years. Consistent with the majority of research in the Northern Hemisphere, analyses on the entire sample found evidence of a small prospective season effect on the BE scale (explaining 2.1% of variance in BE scores). Also, as expected, no season effect was seen on the NA scale. In the entire sample, the season effect was not significant for PA, but joint factor analysis of the BE, PA, and NA scales confirmed that the season effect seen in the BE scale was largely due to items that were pure measures of PA. Winter pattern seasonality was both reliable across measures and significantly more marked among the subgroup of respondents who self-identified winter pattern of mood on the Seasonal Pattern Assessment Questionnaire.

Summer mood in winter depressives: validation of a structured interview

Two structured interviews, the Hypomania Interview Guide (Including Hyperthymia), for Seasonal Affective Disorder (HIGH-SAD) and its successor, the Hypomania Interview Guide (Including Hyperthymia), Retrospective Assessment Version (HIGH-R), were validated for the assessment of nondepressed spring/summer mood states in patients with DSM-III-R or DSM-IV diagnoses of Recurrent Bipolar disorder (I, II or NOS) or Recurrent Major Depressive Disorder (MDD; unipolar), both with Seasonal Pattern, and in normal control subjects (HIGH-SAD only). The instruments retrospectively rate the frequency and severity of DSM diagnostic criterion features as well as several non-DSM features. Both instruments had high internal consistency. Normal controls had lower total scores than unipolar patients, who had lower scores than bipolar patients. Total score classified 85-91% of patients with seasonal affective disorder (SAD) into the correct unipolar or bipolar group. For boundary mood cases, small subsets of features provided better classification accuracy. Based on total score, MDD patients were divided into three subgroups: euthymes (normal mood), hyperthymes (slightly elevated mood), and high-hyperthymes (scores overlapping with hypomania). With the exception of sharpened thinking, DSM items dominated patient classifications. Distinct clusters of "positive" (pleasant, agreeable) or "negative" (impairing) features described the mood states. The HIGH-R and HIGH-SAD are useful for discriminating and classifying hypomania and mania in bipolar patients, and euthymia and hyperthymia in unipolar patients.

Seasonality, social zeitgebers and mood variability in entrainment of mood. Implications for seasonal affective disorder

BACKGROUND

Seasonal variations in mood (seasonality) appear to be entrained to light, a physical zeitgeber. We hypothesised that people high in seasonality may be responsive to a range of zeitgebers, because of greater mood variability. We investigated whether the moods of people high in seasonality were more strongly entrained to the calendar week, a social zeitgeber, and whether any such effect was dependent on variability of mood.

METHODS

53 participants (14 male, 39 female; overall mean age=30) completed a daily mood report, over 56 consecutive days. Participants also completed the Seasonality Score Index (SSI) of the Seasonal Pattern Assessment Questionnaire. Each participant's time series of daily mood was analysed by spectral analysis to quantify the strength of their weekly mood cycle.

RESULTS

Participants with high SSI scores (> or =11) had significantly stronger weekly mood cycles than those with low SSI scores (<11), and significantly greater variability in mood. Covarying for mood variability reduced the difference between high and low SSI groups in mean strength of weekly mood cycle to non-significance.

LIMITATIONS

The time series of moods obtained was relatively short, and moods among high seasonal participants may have been affected by seasonal weather variability.

CONCLUSIONS

People high in seasonality appear to be more responsive to external zeitgebers, and this could be linked to their greater variability in mood. The integration of research on mood variability with research on SAD appears to be warranted.

No changes in mood with the seasons: observations in 3000 chronic pain patients

OBJECTIVE

Seasonal affective disorder (SAD) and the theory of the effect of light on depression have gained some popularity in recent years. Research on epidemiology is largely based on retrospective measures asking explicitly for the experience of seasonal variations. Those measures have a low positive predictive value and do not enable us to distinguish between experience and belief.

METHOD

A consecutive sample of chronic pain patients filled out a depression questionnaire (CES-D) routinely as part of the diagnostic interview on becoming in-patients at a Pain Clinic in Mainz during a 5-year period.

RESULTS

No support for seasonality or light effects was found.

CONCLUSION

The effect of light on depression or seasonality may be smaller than expected in general. SAD as a true disorder is probably rare.

Prevalence of seasonal changes in mood and behavior during the winter months in central Texas

The Seasonal Pattern Assessment Questionnaire (SPAQ) was administered to a sample of students, faculty, and staff at a large state university in central Texas. Estimated prevalence rates of winter seasonal affective disorder (SAD) and subsyndromal winter SAD (S-SAD) derived from their SPAQ responses were considerably higher than expected, given the relatively low latitude of the area. Generally consistent with earlier findings, women experienced greater seasonal change in mood and behavior than men, and older individuals experienced less seasonal change than younger individuals.

Seasonal affective disorder among psychiatric nurses in Aberdeen

The Seasonal Pattern Assessment Questionnaire was posted to psychiatric nurses in Aberdeen, and 443 (73% of eligible subjects) responded. The rate of Seasonal Affective Disorder (SAD) was 2.9% and for subsyndromal SAD (S-SAD) was 9.5%. When rates were age and sex adjusted to the Aberdeen population aged 15 to 64, this yielded a prevalence of 1.9% for SAD and 7.2% for S-SAD. As in other studies, there was a preponderance of young female sufferers, but being married appeared to be protective. Comparisons with epidemiological studies elsewhere lead to the conclusion that population rates of winter depression relate to a combination of genetic susceptibility/selection, latitude of residence and geographical mobility.

Myofascial face pain: seasonal variability in pain intensity and demoralization

Seasonal variability has been detected in a variety of illnesses. The purpose of this study is to examine seasonal variability in pain intensity, demoralization and range of mandibular motion among patients suffering from myofascial face pain. Pain and demoralization for cases (n = 140) and demoralization for controls (n = 133) were measured in each of 10 monthly interviews. Range of motion was measured once only for cases. Their pain intensity (P < 0.009) and demoralization (P < 0.04) were significantly greater in the peak dark months than in the peak light months. There was a non-significant trend (P < 0.07) toward elevated demoralization in the darker months for controls. Range of motion, assessed cross-sectionally, did not demonstrate a seasonal pattern. These data support our prior finding that myofascial face pain and depressed mood are co-morbid and may be maintained by common risk factors.

Behavioral engagement level, variability, and diurnal rhythm as a function of bright light in bipolar II seasonal affective disorder: an exploratory study

Six patients with bipolar II seasonal affective disorder (SAD) and seven normal control subjects rated their moods in winter at six fixed times each week-day during 1 week of dim and 2 weeks of bright light. The scales rated represent the mood dimension specifically associated with depression, a dimension here called behavioral engagement (BE). Compared with controls, depressed SAD subjects (1) showed lower BE levels across all rating times of the day, (2) were more likely to show diurnal variation in BE, (3) displayed more between-day instability in BE diurnal rhythm, and (4) exhibited greater short-term lability (change within 3 hours) in BE. Bright light reduced or eliminated all group differences in BE level and variability.