Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Dec 1.
Published in final edited form as: Depress Anxiety. 2022 Oct 18;39(12):835–844. doi: 10.1002/da.23290

The Relationship between Physical Activity States and Depressive Symptoms: Using Ambulatory Assessment to Characterize Day-to-Day Associations Among Individuals with and Without Bipolar Spectrum Disorder

Rachel F L Walsh 1, Logan T Smith 1, Madison K Titone 2,3, Tommy H Ng 4, Namni Goel 5, Lauren B Alloy 1
PMCID: PMC9729395  NIHMSID: NIHMS1841824  PMID: 36254832

Abstract

Introduction:

The role of activation in the pathogenesis of bipolar spectrum disorders (BSD) is of growing interest. Physical activity is known to improve mood, but it is unclear whether low activity levels contribute to inter-episode depressive symptoms observed in BSD. This study examined whether sedentary and vigorous activity, as well as the timing of the activity, were differentially associated with next-day depressive symptoms for individuals at low-risk for BSD, high-risk for BSD, and diagnosed with BSD.

Methods:

Young adults (n=111, ages 18–27) from three groups (low BSD risk, high BSD risk, and BSD diagnosis), participated in a 20-day ecological momentary assessment study. Physical activity was measured via wrist actigraphy counts. The percentage of time awake spent in sedentary, light, moderate, and vigorous activity states was calculated, as was percentage of morning hours and evening hours in each activity state. Multilevel models examined whether BSD risk group moderated associations between sedentary and vigorous activity and depressive symptoms, which were assessed three times daily.

Results:

There were no between-group differences in time spent in each activity state, nor were there main effects of sedentary or vigorous activity on depression. Increased time spent engaging in vigorous activity was associated with a greater reduction in subsequent depressive symptoms for the BSD group. An increase in evening, but not morning, vigorous activity was significantly associated with a reduction in subsequent depressive symptoms for the BSD group after controlling for chronotype.

Conclusions:

Interventions targeting physical activity may effectively help regulate inter-episode mood disturbances in BSD.

Keywords: bipolar spectrum disorder, physical activity, actigraphy, ecological momentary assessment

Introduction

Bipolar spectrum disorders (BSDs) are mood disorders characterized by periods of elevated or expansive mood, periods of depressed mood, accompanied by sleep and circadian disturbances, increased goal-directed activity, impulsivity, and cognitive changes. Historically, mood disturbances were regarded as the defining feature of BSDs as evidenced by psychiatrist Kraepelin coining the term “manic-depressive psychosis” over a century ago (Hoff, 2015). However, burgeoning research suggests that motor activity and energy disturbances are just as important as mood dysregulation in the nosology of BSDs (Cheniaux et al., 2014). This change is reflected in the updated DSM-5, in which the “increased activity/energy” criterion was promoted from a secondary symptom to a primary “criterion A” symptom necessary for meeting diagnostic criteria for a hypomanic or manic episode (American Psychological Association, 2013).

Substantial empirical research supports this amendment. A recent meta-analysis examining 56 studies on activation (a term describing objective activity changes and subjective energy changes) and bipolar disorder concluded that activation is distinct from mood disturbances and that individuals with BSDs have lower average activity levels (Scott et al., 2017). Clinically, this finding has the potential to increase diagnostic specificity because energy/activity changes are more observable than internal mood changes, and thus, symptom reporting may be more reliable (Machado-Vieira et al., 2017).

Individuals with BSDs appear to have disrupted physical activity patterns even when euthymic. Meta-analytic evidence found that individuals with BSDs are active around 3.5 hours each day and spend approximately 10 hours a day being sedentary (Vancampfort et al., 2016). In a study by Janney and colleagues (2014), adult outpatients with BSDs who were largely euthymic were sedentary for 78% of the day (over 13 hours) and did not reach the 150 minute/week activity guideline for healthy living set by the Department of Health and Human Services (Department of Health and Human Services, 2018). These metrics reflect a significant deviation from the activity patterns exhibited by demographically-matched controls; individuals with BSDs were more sedentary and less active than both psychiatric and healthy controls (Janney et al., 2014).

Physical activity levels among individuals with serious mental illness has received much attention primarily due to mortality implications (Kupfer, 2005). Indeed, individuals with BSDs have increased rates of cardiovascular disease, metabolic syndrome, and obesity (Weiner et al., 2011), conditions which are potentially modifiable with physical activity changes. Physical activity levels also have been linked to overall clinical improvement among those with mood disorders (Hearing et al., 2016). Findings from a systematic review revealed that for individuals with BSDs, exercise is associated with a reduction in depressive symptoms, decreased psychiatric comorbidity, and increased quality of life (Melo, Daher, et al., 2016).

There is consistent evidence that physical activity has positive effects on mood and that fluctuations in physical activity correlate with mood fluctuations (Giacobbi et al., 2005; Kalmbach et al., 2018). However, it is unclear whether the ameliorating effects are greater for those with BSDs compared to controls. Further, many studies assessed physical activity using self-report measures rather than objective indicators such as accelerometers and actigraphs, which are more accurate (Dyrstad et al., 2014; Vancampfort et al., 2013). Additionally, many of the existing studies either are cross-sectional or have a long follow-up period, which prevents a fine-grained examination of activity and mood dynamics. In the only study that used actigraphy and ecological momentary assessment (EMA) to examine dynamic associations between mood and activity within a sample that included individuals with BSDs, an increase in motor activity was associated with a reduction in sad mood, and this association was stronger for those with bipolar I disorder (Merikangas et al., 2019).

The goals of this study were to use EMA and actigraphy to 1) characterize sedentary behavior and activity patterns among individuals at low-risk for BSD, high-risk for BSD, or diagnosed with a BSD, 2) examine associations between sedentary and vigorous activity and subsequent depressive symptoms, and 3) evaluate whether the timing of sedentary and vigorous activity differentially impacts next-day depressive symptoms and symptom variability. As set forth by the reward hypersensitivity model of BSDs, individuals with elevated reward responsiveness were considered to be at heightened risk for developing a BSD (Alloy & Abramson, 2010). This conceptualization is supported by empirical work showing that high scores on self-report questionnaires and behavioral tasks of reward responsiveness predicted first onset of BSD prospectively (Alloy, Bender, et al., 2012). We hypothesized that individuals with BSD would have more sedentary behavior and less vigorous activity than the other two groups. We also expected an increase in vigorous activity would be associated with a reduction in mean levels and variability of depressive symptoms, and that this reduction would be greater for those with BSDs.

Methods

Participants

The participants in this study were a subset of individuals enrolled in Project TEAM (Teen Emotion and Motivation Study; Alloy et al., 2012), a longitudinal study of BSDs in adolescents. During the screening phase, nearly 10,000 adolescents and young adults completed two self-report measures of trait reward sensitivity: the Behavioral Activation Scale (BAS) of the Behavioral Inhibition System/Behavioral Activation System Scales (BIS/BAS; Carver & White, 1994) and the Sensitivity to Reward (SR) subscale of the Sensitivity to Punishment Sensitivity to Reward Questionnaire (SPSRQ; Torrubia et al., 2001). In keeping with prior research and theory, individuals with moderate reward sensitivity were considered to be at low risk for developing a BSD, whereas those with high reward sensitivity were considered to be at high risk. Individuals with either moderate reward sensitivity (40th–60th percentile on both measures) or high reward sensitivity (85th–99th percentile on both measures) completed a diagnostic interview (the expanded Schedule for Affective Disorders and Schizophrenia – Lifetime [exp-SADS-L]); (Alloy et al., 2008; Endicott & Spitzer, 1978) to screen out individuals with a lifetime history of BSD, hypomanic episodes, or psychosis. See Alloy et al. (2012) for further information about Project TEAM.

Several years (M=2.85, SD=2.48) after their initial enrollment in Project TEAM, participants were invited to complete an EMA study on rhythms, sleep, mood, and reward. For inclusion in the present analyses, participants must have completed both the actigraphy and EMA components (n= 115). Four participants were excluded: three participants were excluded because they were missing either all actigraphy data due to actigraph malfunction, all EMA data due to software malfunction, or their body mass index information was missing; one participant was excluded because they were a shift worker. The final sample included 111 individuals. Descriptive statistics are in Table 1.

Table 1.

Descriptive Statistics

Bipolar Spectrum Disorder (BSD) Risk Group
Entire Sample (n=111) Low-risk (MRew) n=31 High-risk (HRew) n=52 Bipolar Spectrum Disorder (HRew+BSD) n=28 Chi-Square/ANOVA
Sex (% female) 60.36 67.74 53.85 64.29 X2(2, 111) =1.95, p=.38
Age in years (M, SD) 22.01 (2.19) 22.31 (2.16) 22.05 (2.46) 21.59 (1.66) F(2, 108)=.81, p=.45
Race (% White) 56.48 43.33 62.75 59.26 X2(2, 111) = 2.91, p=.23
Depressive Symptoms (M, SD) 1.31 (.26) 1.24 (.24) 1.32 (.28) 1.37 (.32) F(2, 108)=1.79, p=.17
Percent of Day Sedentary Activity (M, SD) 12.78 (5.33) 12.19 (4.51) 12.97 (5.55) 13.06 (5.85) F(2, 108)=.26, p=.77
Percent of Day Light Activity (M, SD) 33.70 (10.06) 33.70 (7.75) 32.42 (7.91) 36.07 (14.78) F(2, 108)=1.21, p=.30
Percent of Day Moderate Activity (M, SD) 40.12 (12.06) 41.26 (11.51) 41.23 (11.65) 36.81 (13.18) F(2, 108)=1.42, p=.25
Percent of Day Vigorous Activity (M, SD) 13.40 (5.35) 12.86 (4.53) 13.38 (5.02) 14.05 (6.74) F(2, 108)=.36, p=.70
Open in a new tab

Notes: M=mean; SD=standard deviation; MRew=moderate reward; HRew=high reward

Procedure

At study visits for Project TEAM (i.e., baseline and prospective 6-month follow-ups), participants completed a diagnostic interview and the Morningness-Eveningness Questionnaire. During the 20-day study period, participants answered questions about depressive symptoms three times per day. The survey prompts alerted participants randomly within three four-hour blocks such that they responded once in the morning, afternoon, and evening. An actigraph was worn continuously throughout the study period.

Measures

Structured Clinical Diagnostic Interview

At Project TEAM baseline visits, psychiatric diagnoses (DSM-IV-TR and RDC) were determined using the exp-SADS-L. Diagnoses at follow-ups were made using the change interview (exp-SADS-C; Alloy et al., 2008; Endicott & Spitzer, 1978). The exp-SADS has demonstrated high inter-rater reliability (κ>.96; Alloy et al., 2008).

Reward Sensitivity Screening Measures

The Behavioral Inhibition System/Behavioral Approach System (BIS/BAS) Scale.

The BIS/BAS is a 20-item self-report questionnaire that measures sensitivity to rewards (behavioral approach) and punishments (behavioral inhibition). Each of the 20 statements about approach and avoidance tendencies are rated on a four-point Likert scale. The BIS/BAS had good internal consistency and retest reliability during its initial validation (Carver & White, 1994) and demonstrated strong internal consistency (α=.80) in the Project TEAM sample (Alloy, Bender, et al., 2012). There are four scales-one BIS scale and three BAS subscales (reward responsiveness, fun-seeking, and drive), that are summed to create a total BAS score.

Sensitivity to Punishment Sensitivity to Reward Questionnaire (SPSRQ).

Another self-report questionnaire, the SPSRQ, also was used to measure reward sensitivity (Torrubia et al., 2001). The SPSRQ is a 48-item scale that assesses specific rewards and punishments, which form two 24-item subscales (SR and SP). In the screening sample, the internal consistency of the SR scale was α=.76, and the SR and BAS scores were correlated (r=.40). Alloy et al. (2012) found that participants in the high reward sensitivity group were significantly more likely to develop a BSD than participants in the moderate reward sensitivity group.

Mood Symptom Measures

Beck Depression Inventory-II (BDI-II).

The BDI-II is a widely used, psychometrically robust self-report questionnaire that measures depressive symptoms (Beck et al., 1996). Participants answered four BDI-II items three times per day. Each item was rated on a four-point Likert scale from 1–4. Mean levels of daily depressive symptoms were determined by averaging responses from the three daily prompts. For participants who responded to at least two prompts in a given day, the standard deviation was calculated as an indicator of within-day symptom variability.

Chronotype and Sleep

Morningness-Eveningness Questionnaire (MEQ).

The MEQ is a 19-tem self-report questionnaire of chronotype (Horne & Ostberg, 1976). It measures the extent to which an individual prefers engaging in daily activities in the morning or evening. For example, someone with a morning chronotype tends to wake up early, performs better in morning hours, and goes to bed earlier. Chronotype as measured by the MEQ is believed to reflect a trait-based circadian preference and demonstrates test-retest reliability (Neubauer, 1992). Higher scores reflect greater morningness.

Sleep Time and Activity Parameters

During the 20-day EMA period, participants wore an Actiwatch Spectrum (Philips Healthcare, Bend, OR) on their non-dominant wrist to track physical activity and sleep/wake cycles, which allowed for continuous, non-invasive measurement. The Actiware software was used to determine total sleep time and sleep onset latency using one-minute epochs (Titone et al., 2020, 2022). Actigraphy is a valid measure of sleep that correlates with polysomnography and sleep diaries in both healthy sleepers and clinical samples (Cole et al., 1992; Kaplan et al., 2012; Ng et al., 2015).

Average hourly activity count was computed from the one-minute epochs. Each waking hour was categorized as sedentary (<145 counts), light (145–274 counts), moderate (274–597 counts) or vigorous (>597) activity, according to previously-validated cut-off thresholds (Allison et al., 2021; Neil-Sztramko et al., 2017). Waking hours were identified based on the actigraph-derived sleep time and wake time. To account for differences in how many hours each participant was awake, we calculated the percent of a participant’s waking hours that was spent in each of the four states. Morning and evening sedentary and vigorous activity was operationalized as the percentage of waking hours an individual spent engaging in sedentary or vigorous activity between 6 am – 12 noon and 6 pm – 12 midnight, respectively.

Data Analysis

R version 3.6.2 (Team, 2013) was used for all analyses. A multilevel modeling approach was utilized because observation-level data (i.e., EMA responses) were nested within persons, violating the independence assumption necessary for traditional linear regression, and multilevel modeling provides more accurate estimates for hierarchical data structures. All models included a random intercept and fixed slope and were estimated using the lme4 (Bates et al., 2015) and lmerTest packages (Kuznetsova et al., 2017). Models utilized restricted maximum likelihood.

To examine the associations between percentage of the day spent in sedentary and vigorous activity states and subsequent depressive symptoms, several sets of multilevel models were estimated. The percentage of the day spent engaging in vigorous activity and the percentage of the day spent being sedentary were lagged by one day for temporal precedence and included in the models as focal predictors. One model examined the impact of sedentary and vigorous activity on average levels of next-day depression symptoms and another model examined the impact of sedentary and vigorous activity on the variability of next-day depression symptoms. Analyses did not include the percentage of time spent engaging in light or moderate activity as to not overfit the models, and because we were most interested in the two activity extremes (e.g., sedentary and vigorous).

The predictors (e.g., percentage of time spent in sedentary and vigorous activity) were person-mean centered to reflect within-person variability. The predictors and outcome variables were z standardized. Models also controlled for age, sex, BSD risk group, previous-day depression symptoms, and body mass index (BMI). The reference group for BSD risk status was “low-risk.”

In addition to investigating main effects, in step 2 of model estimation, we also evaluated whether BSD risk group moderated the relationship between percentage of time in each activity state and subsequent depressive symptoms by introducing cross-level interactions. For all significant interactions, simple slopes were calculated using the reghelper (v.1.0.2; Hughes, 2021) package. Given the BSD risk group variable has three levels, all models were re-estimated a second time to obtain all unique pairwise comparisons. These models were identical to primary models, but “high-risk” was the reference category instead of “low-risk”.

In exploratory analyses, we examined whether the timing of sedentary or vigorous activity impacted next-day depressive symptoms. Given that circadian dysregulation is implicated in the pathophysiology of BSD, we focused on the time-of-day extremes (i.e., morning and evening activity), where we expected group differences were most likely to emerge. Step 1 modeled the main effects of morning sedentary and vigorous activity, as well as evening sedentary and vigorous activity. Step 2 contained the moderator analyses, specifically four cross-level interactions between BSD risk group and morning/evening sedentary/vigorous activity. These analyses controlled for age, sex, BMI, prior day depressive symptoms, and chronotype (MEQ score).

Given that four participants in the BSD group were in a DSM mood episode at the time of the EMA study (two had depressive episodes, one had a mixed episode, and one had a hypomanic episode), we repeated the analyses with these four participants removed as a sensitivity analysis.

Results

Descriptive statistics are presented in Table 1. The average EMA survey completion rate was 86.58%, indicating high compliance, and there were no significant differences in compliance rates by BSD risk group (X2(2,114)=0.29, p=.75) or by time of day (X2(2,114)=1.27, p=.28). Overall, participants spent most of their time engaging in light (33.760%) and moderate (40.103%) activity. On average, participants were sedentary 12.748% of their waking hours and engaged in vigorous activity 13.389% of the time. There were no group differences in the percentage of the day spent in any of the four activity states (all ps>.05).

First, we examined potential associations between the percentage of the day spent in sedentary or vigorous activity and next-day depressive symptoms, as well as the variability in next-day depressive symptoms. As seen in Table 2, the main effects models (e.g., Step 1) revealed that there were no main effects of sedentary or vigorous activity on next-day depression (all ps>0.05). However, both the high-risk (B=.263, SE=.109, p=.018) and BSD (B=.311, SE=.123, p=.014) participants had significantly greater variability in depressive symptoms than the low-risk group. Findings from the interaction models (e.g., Step 2) revealed that an increase in vigorous activity was significantly associated with a greater reduction in next-day depressive symptoms for the BSD group than both the low-risk (B=−.142, SE=.051, p=.005) and high-risk groups (B=−.132, SE=.044, p=.003), as seen in Figure 1. Follow-up simple slopes analyses revealed a significant association between vigorous activity and depressive symptoms for the BSD group (B=−.124, SE=.035, p<.001). Further, an increase in vigorous activity for the BSD group also was significantly associated with a decrease in next-day depression symptom variability compared to the high-risk group (B=−.121, SE=.054, p=.024). Simple slope analysis indicated that there was a significant association between vigorous activity and next-day depressive symptom variability (B=−.095, SE=.042, p=.025) for the BSD group. None of the interactions examining whether BSD risk group impacted the relationship between sedentary behavior and depression were significant.

Table 2.

Associations between percentage of day spent in sedentary or vigorous activity and next-day depressive symptoms and symptom variability, controlling for prior day’s depressive symptoms.

Mean Depression Symptoms Depression Symptom Variability
Step 1 Step 2 Step 1 Step 2
B SE p B SE p B SE p B SE p
Level 1
Sedentary Activity −.008 .019 .668 −.012 .037 .748 −.006 .023 .795 .009 .046 .850
Sedentary * High-risk (ref: low-risk) −.030 .046 .516 −.038 .056 .502
Sedentary * BSD (ref: low-risk) .049 .051 .332 <.001 .062 .995
Sedentary * BSD (ref: high-risk)† .079 .044 .071 .038 .053 .471
Vigorous Activity −.029 .018 .116 .018 .037 .623 −.016 .022 .464 −.002 .044 .965
Vigorous * High-risk (ref: low-risk) −.010 .046 .831 .028 .055 .609
Vigorous * BSD (ref: low-risk) −.142 .051 .005 −.093 .061 .129
Vigorous * BSD (ref: high-risk)† −.132 .044 .003 −.121 .054 .024
Previous day’s depressive symptoms .151 .026 <.001 .152 .026 <.001 .124 .029 <.001 .124 .029 <.001
Level 2
Group
 High-risk (ref: low-risk) .185 .140 .192 .185 .140 .191 .263 .109 .018 .263 .109 .018
 BSD (ref: low-risk) .271 .160 .094 .273 .160 .091 .311 .124 .014 .311 .124 .014
 BSD (ref: high-risk)† .086 .144 .550 .088 .144 .542 .049 .111 .666 .048 .111 .667
Observations 111 111 111 111
Participants 1464 1464 1386 1386
ICC .392 .394 .187 .187
Residual Variance .506 .502 .696 .695
Open in a new tab

Notes: ref: = reference; BSD = bipolar spectrum disorders; Sedentary and vigorous activity are centered on each participant’s mean; All analyses control for sex, age, and BMI; Step 1 models only contain main effects; Interaction components were added in Step 2; Sedentary activity, vigorous activity, and depressive symptoms were z-standardized; Significant findings are in bold;

†

These estimates were ascertained from separate models, identical to the primary models, except the reference group for BSD risk group was changed from low-risk to high-risk to ascertain all pairwise comparisons.

Figure 1.

Figure 1.

Open in a new tab

Vigorous Activity and Depressed Mood by BSD Risk Group

Note: BSD = Bipolar Spectrum Disorder.

Results from analyses examining the relationship between the timing of the sedentary and vigorous activity are presented in Table 3. None of the moderation analyses examining the interactive effects of group and morning vigorous activity reached significance. However, BSD risk group moderated the association between evening vigorous activity and depressive symptoms: an increase in evening vigorous activity was significantly associated with a reduction in subsequent mean depressive symptoms for the BSD group relative to the low-risk (B=−.145, SE=.066, p=.002) and high-risk groups (B=−.183, SE=.058, p=.002; see Figure 2). As indicated by simple slopes, a one standard deviation increase in evening vigorous activity was associated with a .135 (SD=.045) reduction in next-day depressive symptoms for the BSD group (p=.003). There also was a significant interaction between evening vigorous activity and group on depressive symptom variability for the BSD group compared to the high-risk group (B=−.153, SE=.070, p=.029), but the simple slopes were not significant. None of the analyses involving the timing of sedentary activity reached significance, and thus, the findings are not shown in Table 3. The pattern of findings remained the same when the analyses were conducted without the four BSD participants who were in episode at the time of the EMA study.

Table 3.

Associations between vigorous activity in the morning and evening and next-day mood symptoms, controlling for prior-day’s depressive symptoms

Mean Depressive Symptoms Depression Symptom Variability
Step 1 Step 2 Step 1 Step 2
B SE p B SE p B SE p B SE p
Level 1
AM Vigorous Activity −.019 .023 .408 .033 .044 .450 −.034 .028 .223 .008 .053 .874
AM Vigorous * High-risk (ref: low-risk) −.090 .055 .100 −.099 .065 .131
AM Vigorous * BSD (ref: low-risk) −.064 .069 .354 .032 .085 .709
AM Vigorous * BSD (ref: high-risk)† .026 .062 .674 .130 .078 .094
PM Vigorous Activity −.015 .024 .542 .010 .048 .831 .016 .029 .590 .008 .057 .893
PM Vigorous * High-risk (ref: low-risk) .038 .060 .526 .068 .072 .344
PM Vigorous * BSD (ref: low-risk) −.145 .066 .029 −.085 .080 .286
PM Vigorous * BSD (ref: high-risk)† −.182 .058 .002 −.153 .070 .029
Previous day’s depressive symptoms .181 .028 <.001 .179 .028 <.001 .141 .032 <.001 .142 .032 <.001
Level 2
Group
 High-risk (ref: low-risk) .225 .144 .124 .224 .048 .831 .310 .113 .008 .311 .113 .007
 BSD (ref: low-risk) .297 .165 .077 .304 .145 .126 .346 .130 .009 .353 .129 .008
 BSD (ref: high-risk)† .071 .151 .639 .080 .151 .599 .035 .118 .764 .043 .117 .718
Observations 103 103 103 103
Participants 1218 1218 1162 1162
ICC .385 .387 .179 .177
Residual Variance .515 .511 .709 .709
Open in a new tab

Notes: ref: = reference; BSD = bipolar spectrum disorders; Sedentary and vigorous activity are centered on each participant’s mean; “AM” refers to 6am-12 noon, “PM” refers to 6 pm-12 midnight period; All analyses control for sex, age, chronotype, and BMI; Step 1 models only contain main effects; Interaction components were added in Step 2; Sedentary activity, vigorous activity, and depressive symptoms were z-standardized; Significant findings are in bold;

†

These estimates were ascertained from separate models, identical to the primary models, except the reference group for BSD risk group was changed from low-risk to high-risk to ascertain all pairwise comparisons.

Figure 2.

Figure 2.

Open in a new tab

Evening Vigorous Activity and Next-Day Depressive Symptoms

Note: BSD = Bipolar Spectrum Disorder.

To begin exploring the possibility that the moderating role of BSD diagnosis in the relationship between evening vigorous activity and subsequent depressive symptoms is related to sleep disturbance, follow-up analyses examined whether greater evening vigorous activity was predictive of total sleep time or sleep onset latency, and whether this relationship differed by BSD risk group. Evening vigorous activity was not related to total sleep time, but an increase in evening vigorous activity was associated with increased between-person sleep onset latency (B=2.199, SE=.822, p=.009). Analyses examining group differences found this effect was greater for individuals with BSD relative to the high-risk group (B=3.760, SE=1.85, p=.044), but not the low-risk group (B=2.723, SE=2.355, p=.250). A table of these results is located in supplementary Table 1.

Discussion

Using EMA and actigraphy, this study expanded research on the role of activation in BSD and examined the interplay between physical activity states and depressive symptoms among young adults diagnosed and at low- and high-risk for BSDs over a 20-day period. On average, there were no group differences in the percentage of waking hours spent in sedentary, light, moderate, and vigorous activity states. However, an increase in vigorous activity was associated with a significantly greater decrease in next-day depressive symptoms for individuals with BSD. Follow-up analyses examining whether the reduction in subsequent depressive symptoms was driven by the timing of the activity suggested that vigorous activity in the evening may be driving these findings. This naturalistic study of physical activity and mood helps characterize inter-episode mood changes among individuals with BSDs and complements existing literature on the effectiveness of exercise interventions for BSD.

There were no group differences in percentage of time spent in each activity state (sedentary, light, moderate, and vigorous). This was unexpected considering a prior meta-analysis found that individuals with bipolar disorder spent significantly more time sedentary and less time engaging in moderate and vigorous activity (Vancampfort et al., 2017). Of note, the average age of individuals in the aforementioned meta-analysis was 43 years-old and the current study’s sample had a mean age of 22 years-old. One potential explanation is that the shift towards more sedentary activity is a consequence of the length of time living with BSD. Indeed, in a 20-year prospective study of bipolar disorder, as duration of illness increased, so did the proportion of time spent depressed, as did symptom chronicity (Coryell et al., 2009). Additionally, physical activity is known to decrease throughout the lifespan, so the increased physical activity characteristic of young adulthood may be a protective factor, obscuring group differences that become more pronounced later in life (Hallal et al., 2012).

As hypothesized, BSD diagnosis moderated the relationship between vigorous activity and next-day depressive symptoms, although there were no main effects of vigorous activity on depressive symptoms. Shown in Figure 1, an increase in vigorous activity was associated with a significantly greater reduction in depressive symptoms for those with BSD. Prior work found that adults with BSD exhibit greater cross-domain reactivity: that is, there are stronger bidirectional associations among mood, energy, and activity for those with BSD relative to controls (Merikangas et al., 2019). Cross domain reactivity typically is viewed negatively, especially as it pertains to emotion reactivity, in which an event serves as a catalyst for heightened emotional, behavioral, or physiological sequelae, often of greater magnitude than expected (Henry et al., 2012; McLaughlin et al., 2010). Our findings suggest that the elevated cross-domain reactivity characteristic of bipolar disorder could be beneficial in terms of diversifying intervention targets. Indeed, it has been speculated that interventions designed to impact multiple domains may be of increased importance in treating BSD (Merikangas et al., 2019).

In exploratory analyses investigating whether the timing of vigorous activity had differential effects on depressive symptoms, it appears evening, but not morning, vigorous activity was driving our findings. Indeed, as seen in Figure 2, after controlling for chronotype, there was a significant interaction such that evening vigorous activity was associated with a greater reduction in depressive symptoms and symptom variability for the BSD group relative to both the high-risk and low-risk groups. It is interesting to note that, traditionally, late-night exercise has been discouraged due to negative effects on total sleep time and sleep onset latency (American Academy of Sleep Medicine, 2001), and sleep deprivation and chronic sleep problems are a known risk factor for depression and mood lability (Franzen & Buysse, 2008). Considering sleep disturbance is a core feature of BSD and can serve as an indicator of an emerging mood episode (Hirata et al., 2007), it is important to disentangle the associations between exercise, sleep, and mood symptoms/episodes, specifically for individuals with BSDs, because the dynamic associations may differ from the general population.

The supplemental analyses probing the relationship between evening vigorous activity and sleep outcomes found that people with greater evening vigorous activity tended to have increased sleep onset latency, and this association was stronger for individuals with BSD relative to the high-risk group. These findings suggest that delayed sleep onset latency is one possible mechanism that may explain how vigorous activity late in the day contributes to mood, specifically for those with BSD, although there are likely other contributing factors. Additional research is needed to sufficiently test this possibility.

Further, altered circadian rhythmicity is a known feature of BSD (Takaesu, 2018). Individuals with BSD are more likely to have evening chronotypes, delayed sleep phase, and altered melatonin levels (Melo, Garcia, et al., 2016). Prior research on exercise and sleep latency found that, in a non-clinical sample of young adults, chronotype moderated the relationship between exercise and sleep timing such that evening exercise had a more negative impact on sleep for individuals with morning, compared to evening, chronotypes (Glavin et al., 2021). Our findings were significant despite controlling for chronotype, suggesting that the observed effects of evening exercise may exceed those attributed to time-of-day preference.

This study has several notable strengths and a few limitations. Strengths include an objectively-assessed indicator of physical activity and the use of previously-validated cutoff thresholds for activity type (e.g., vigorous exercise; Allison et al., 2021; Neil-Sztramko et al., 2017); however a limitation is that formal exercise was not measured. This study examined a sample of young adults, which is a period of heightened risk for BSD; however, our findings may not be generalizable to older individuals. Depressive symptoms were assessed using a multi-item scale three times per day in vivo, which increases the ecological validity. Nevertheless, the survey structure may have been too infrequent to capture mood variability, a construct that may be better studied using a more extensive sampling schedule. Lastly, to retain power, we did not further separate the BSD group by bipolar subtype or medication status.

Conclusions

This study utilized an intensive actigraphy and EMA protocol to understand dynamic associations between physical activity states and depressive symptoms. We found an increase in vigorous activity, specifically in the evening, was associated with a greater reduction in subsequent depressive symptoms. Interventions targeting physical activity may be effective non-pharmacologic options for regulating inter-episode mood disturbances in BSD.

Supplementary Material

1

Acknowledgements:

This study was supported in part by the National Science Foundation’s Graduate Research Fellowship to Rachel Walsh and by National Institute of Mental Health R01 grants MH077908, MH102310, and MH126911 to Lauren B. Alloy. Namni Goel was supported in part by National Aeronautics and Space Administration (NASA) grants NNX14AN49G and 80NSSC20K0243 and National Institutes of Health grant R01DK117488. Madison Titone’s contributions were supported by the Department of Veterans Affairs Office of Academic Affiliations Advanced Fellowship Program in Mental Illness Research and Treatment, the Medical Research Service of the Veterans Affairs San Diego Health Care System, and the Department of Veterans Affairs Desert Pacific Mental Illness Research, Education, and Clinical Center (MIRECC).

References

  1. Allison KC, Hopkins CM, Ruggieri M, Spaeth AM, Ahima RS, Zhang Z, Taylor DM, & Goel N (2021). Prolonged, Controlled Daytime versus Delayed Eating Impacts Weight and Metabolism. Current Biology, 31, 650–657.e3. 10.1016/j.cub.2020.10.092 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Alloy LB, & Abramson LY (2010). The role of the behavioral approach system (BAS) in bipolar spectrum disorders. Current Directions in Psychological Science, 19, 189–194. 10.1177/0963721410370292 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Alloy LB, Abramson LY, Walshaw PD, Cogswell A, Grandin LD, Hughes ME, Iacoviello BM, Whitehouse WG, Urosevic S, Nusslock R, & Hogan ME (2008). Behavioral Approach System and Behavioral Inhibition System sensitivities and bipolar spectrum disorders: Prospective prediction of bipolar mood episodes. Bipolar Disorders, 10, 310–322. 10.1111/j.1399-5618.2007.00547.x [DOI] [PubMed] [Google Scholar]
  4. Alloy LB, Bender RE, Whitehouse WG, Wagner CA, Liu RT, Grant DA, Jager-Hyman S, Molz A, Choi JY, Harmon-Jones E, & Abramson LY (2012). High behavioral approach system (BAS) sensitivity, reward responsiveness, and goal-striving predict first onset of bipolar spectrum disorders: A prospective behavioral high-risk design. Journal of Abnormal Psychology, 121, 339–351. 10.1037/a0025877 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Alloy LB, Urošević S, Abramson LY, Jager-Hyman S, Nusslock R, Whitehouse WG, & Hogan M (2012). Progression along the bipolar spectrum: A longitudinal study of predictors of conversion from bipolar spectrum conditions to bipolar I and II disorders. Journal of Abnormal Psychology, 121, 16–27. 10.1037/a0023973 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. American Academy of Sleep Medicine. (2001). The international classification of sleep disorders, revised: Diagnostic and coding manual. In American Academy of Sleep Medicine. [Google Scholar]
  7. American Psychological Association. (2013). Diagnostic and Statistical Manual of Mental Disorders. In American Psychological Association (5th ed.). 10.5771/0943-7444-2017-8-668 [DOI] [Google Scholar]
  8. Bates D, Maechler M, Bolker B, & Walker S (2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67, 1–48. 10.18637/jss.v067.i01>. [DOI] [Google Scholar]
  9. Beck AT, Steer RA, & Brown GK (1996). Manual for the Beck depression inventory-II. San Antonio, TX: Psychological Corporation. [Google Scholar]
  10. Carver CS, & White TL (1994). Behavioral Inhibition, Behavioral Activation, and Affective Responses to Impending Reward and Punishment: The BIS/BAS Scales. Journal of Personality and Social Psychology, 67, 319–333. 10.1037/0022-3514.67.2.319 [DOI] [Google Scholar]
  11. Cheniaux E, Filgueiras A, Silva R. D. A. Da, Silveira LAS, Nunes ALS, & Landeira-Fernandez J (2014). Increased energy/activity, not mood changes, is the core feature of mania. Journal of Affective Disorders, 152–154, 256–261. 10.1016/j.jad.2013.09.021 [DOI] [PubMed] [Google Scholar]
  12. Cole RJ, Kripke DF, Gruen W, Mullaney DJ, & Gillin JC (1992). Automatic sleep/wake identification from wrist activity. Sleep, 15, 461–469. 10.1093/sleep/15.5.461 [DOI] [PubMed] [Google Scholar]
  13. Coryell W, Fiedorowicz J, Solomon D, & Endicott J (2009). Age transitions in the course of bipolar I disorder. Psychological Medicine, 39, 1247–1252. 10.1017/S0033291709005534 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Department of Health and Human Services. (2018). Physical Activity Guidelines for Americans: Vol. 2nd editio. [Google Scholar]
  15. Dyrstad SM, Hansen BH, Holme IM, & Anderssen SA (2014). Comparison of self-reported versus accelerometer-measured physical activity. Medicine and Science in Sports and Exercise, 46, 99–106. 10.1249/MSS.0b013e3182a0595f [DOI] [PubMed] [Google Scholar]
  16. Endicott J, & Spitzer RL (1978). A Diagnostic Interview: The Schedule for Affective Disorders and Schizophrenia. Archives of General Psychiatry, 35, 837–844. 10.1001/archpsyc.1978.01770310043002 [DOI] [PubMed] [Google Scholar]
  17. Franzen PL, & Buysse DJ (2008). Sleep disturbances and depression: Risk relationships for subsequent depression and therapeutic implications. Dialogues in Clinical Neuroscience, 10, 473–481. 10.31887/dcns.2008.10.4/plfranzen [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Giacobbi PR, Hausenblas HA, & Frye N (2005). A naturalistic assessment of the relationship between personality, daily life events, leisure-time exercise, and mood. Psychology of Sport and Exercise, 6, 67–81. 10.1016/j.psychsport.2003.10.009 [DOI] [Google Scholar]
  19. Glavin EE, Ceneus M, Chanowitz M, Kantilierakis J, Mendelow E, Mosquera J, & Spaeth AM (2021). Relationships between sleep, exercise timing, and chronotype in young adults. Journal of Health Psychology, 26, 2636–2647. 10.1177/1359105320926530 [DOI] [PubMed] [Google Scholar]
  20. Hallal PC, Andersen LB, Bull FC, Guthold R, Haskell W, Ekelund U, Alkandari JR, Bauman AE, Blair SN, Brownson RC, Craig CL, Goenka S, Heath GW, Inoue S, Kahlmeier S, Katzmarzyk PT, Kohl HW, Lambert EV, Lee IM, … Wells JC (2012). Global physical activity levels: Surveillance progress, pitfalls, and prospects. The Lancet, 380, 247–257. 10.1016/S0140-6736(12)60646-1 [DOI] [PubMed] [Google Scholar]
  21. Hearing CM, Chang WC, Szuhany KL, Deckersbach T, Nierenberg AA, & Sylvia LG (2016). Physical Exercise for Treatment of Mood Disorders: A Critical Review. Current Behavioral Neuroscience Reports, 3, 350–359. 10.1007/s40473-016-0089-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Henry C, Phillips M, Leibenluft E, M’Bailara K, Houenou J, & Leboyer M (2012). Emotional dysfunction as a marker of bipolar disorders. Frontiers in Bioscience - Elite, 4 E, 2722–2730. 10.2741/e578 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hirata FC, Lima MCO, De Bruin VMS, Nóbrega PR, Wenceslau GP, & De Bruin PFC (2007). Depression in medical school: The influence of morningness-eveningness. Chronobiology International, 24, 939–946. 10.1080/07420520701657730 [DOI] [PubMed] [Google Scholar]
  24. Hoff P (2015). The Kraepelinian tradition. Dialogues in Clinical Neuroscience, 17, 31–41. 10.31887/dcns.2015.17.1/phoff [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Horne JA, & Ostberg O (1976). A self assessment questionnaire to determine Morningness Eveningness in human circadian rhythms. International Journal of Chronobiology, 4, 97–110. [PubMed] [Google Scholar]
  26. Hughes J (2021). reghelper: Helper Functions for Regression Analysis (p. R package version 1.0.2). https://cran.r-project.org/package=reghelper
  27. Janney CA, Fagiolini A, Swartz HA, Jakicic JM, Holleman RG, & Richardson CR (2014). Are adults with bipolar disorder active? Objectively measured physical activity and sedentary behavior using accelerometry. Journal of Affective Disorders, 152–154, 498–504. 10.1016/j.jad.2013.09.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kalmbach DA, Fang Y, Arnedt JT, Cochran AL, Deldin PJ, Kaplin AI, & Sen S (2018). Effects of Sleep, Physical Activity, and Shift Work on Daily Mood: a Prospective Mobile Monitoring Study of Medical Interns. Journal of General Internal Medicine, 33, 914–920. 10.1007/s11606-018-4373-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Kaplan KA, Talbot LS, Gruber J, & Harvey AG (2012). Evaluating sleep in bipolar disorder: Comparison between actigraphy, polysomnography, and sleep diary. Bipolar Disorders, 14, 870–879. 10.1111/bdi.12021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Kupfer DJ (2005). The increasing medical burden in bipolar disorder. In Journal of the American Medical Association (Vol. 293, Issue 20, pp. 2528–2530). 10.1001/jama.293.20.2528 [DOI] [PubMed] [Google Scholar]
  31. Kuznetsova A, Brockhoff PB, & Christensen RHB (2017). lmerTest Package: Tests in Linear Mixed Effects Models. Journal of Statistical Software, 82, 1–26. 10.18637/jss.v082.i13 [DOI] [Google Scholar]
  32. Machado-Vieira R, Luckenbaugh DA, Ballard ED, Henter ID, Tohen M, Suppes T, & Zarate CA (2017). Increased activity or energy as a primary criterion for the diagnosis of bipolar mania in DSM-5: Findings from the STEP-BD study. American Journal of Psychiatry, 174, 70–76. 10.1176/appi.ajp.2016.15091132 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. McLaughlin KA, Kubzansky LD, Dunn EC, Waldinger R, Vaillant G, & Koenen KC (2010). Childhood social environment, emotional reactivity to stress, and mood and anxiety disorders across the life course. Depression and Anxiety, 27, 1087–1094. 10.1002/da.20762 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Melo MCA, Daher EDF, Albuquerque SGC, & De Bruin VMS (2016). Exercise in bipolar patients: A systematic review. In Journal of Affective Disorders (Vol. 198, pp. 32–38). 10.1016/j.jad.2016.03.004 [DOI] [PubMed] [Google Scholar]
  35. Melo MCA, Garcia RF, Linhares Neto VB, Sá MB, de Mesquita LMF, de Araújo CFC, & de Bruin VMS (2016). Sleep and circadian alterations in people at risk for bipolar disorder: A systematic review. In Journal of Psychiatric Research. 10.1016/j.jpsychires.2016.09.005 [DOI] [PubMed] [Google Scholar]
  36. Merikangas KR, Swendsen J, Hickie IB, Cui L, Shou H, Merikangas AK, Zhang J, Lamers F, Crainiceanu C, Volkow ND, & Zipunnikov V (2019). Real-time Mobile Monitoring of the Dynamic Associations among Motor Activity, Energy, Mood, and Sleep in Adults with Bipolar Disorder. JAMA Psychiatry, 76, 190–198. 10.1001/jamapsychiatry.2018.3546 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Neil-Sztramko SE, Rafn BS, Gotay CC, & Campbell KL (2017). Determining activity count cut-points for measurement of physical activity using the Actiwatch2 accelerometer. Physiology and Behavior, 173, 95–100. 10.1016/j.physbeh.2017.01.026 [DOI] [PubMed] [Google Scholar]
  38. Neubauer AC (1992). Psychometric comparison of two circadian rhythm questionnaires and their relationship with personality. Personality and Individual Differences, 13, 125–131. 10.1016/0191-8869(92)90035-N [DOI] [Google Scholar]
  39. Ng TH, Chung KF, Ho FYY, Yeung WF, Yung KP, & Lam TH (2015). Sleep-wake disturbance in interepisode bipolar disorder and high-risk individuals: A systematic review and meta-analysis. In Sleep Medicine Reviews (Vol. 20, pp. 46–58). 10.1016/j.smrv.2014.06.006 [DOI] [PubMed] [Google Scholar]
  40. Scott J, Murray G, Henry C, Morken G, Scott E, Angst J, Merikangas KR, & Hickie IB (2017). Activation in bipolar disorders a systematic review. JAMA Psychiatry, 74, 189–196. 10.1001/jamapsychiatry.2016.3459 [DOI] [PubMed] [Google Scholar]
  41. Takaesu Y (2018). Circadian rhythm in bipolar disorder: A review of the literature. In Psychiatry and Clinical Neurosciences (Vol. 72, Issue 9, pp. 673–682). 10.1111/pcn.12688 [DOI] [PubMed] [Google Scholar]
  42. Team, R. C. (2013). R: A language and environment for statistical computing. R Foundation for Statistical Computing. [Google Scholar]
  43. Thomson D, Turner A, Lauder S, Gigler ME, Berk L, Singh AB, Pasco JA, Berk M, & Sylvia L (2015). A brief review of exercise, bipolar disorder, and mechanistic pathways. In Frontiers in Psychology (Vol. 6, Issue MAR). 10.3389/fpsyg.2015.00147 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Titone MK, Goel N, Ng TH, MacMullen LE, & Alloy LB (2022). Impulsivity and sleep and circadian rhythm disturbance predict next-day mood symptoms in a sample at high risk for or with recent-onset bipolar spectrum disorder: An ecological momentary assessment study. Journal of Affective Disorders, 298, 17–25. 10.1016/j.jad.2021.08.155 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Titone MK, McArthur BA, Ng TH, Burke TA, McLaughlin LE, MacMullen LE, Goel N, & Alloy LB (2020). Sex and race influence objective and self-report sleep and circadian measures in emerging adults independently of risk for bipolar spectrum disorder. Scientific Reports, 10, 13731. 10.1038/s41598-020-70750-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Torrubia R, Ávila C, Moltó J, & Caseras X (2001). The Sensitivity to Punishment and Sensitivity to Reward Questionnaire (SPSRQ) as a measure of Gray’s anxiety and impulsivity dimensions. Personality and Individual Differences, 31, 837–862. 10.1016/S0191-8869(00)00183-5 [DOI] [Google Scholar]
  47. Vancampfort D, Correll CU, Probst M, Sienaert P, Wyckaert S, De Herdt A, Knapen J, De Wachter D, & De Hert M (2013). A review of physical activity correlates in patients with bipolar disorder. In Journal of Affective Disorders (Vol. 145, Issue 3, pp. 285–291). 10.1016/j.jad.2012.07.020 [DOI] [PubMed] [Google Scholar]
  48. Vancampfort D, Firth J, Schuch FB, Rosenbaum S, Mugisha J, Hallgren M, Probst M, Ward PB, Gaughran F, De Hert M, Carvalho AF, & Stubbs B (2017). Sedentary behavior and physical activity levels in people with schizophrenia, bipolar disorder and major depressive disorder: a global systematic review and meta-analysis. World Psychiatry, 16, 308–315. 10.1002/wps.20458 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Vancampfort D, Firth J, Schuch F, Rosenbaum S, De Hert M, Mugisha J, Probst M, & Stubbs B (2016). Physical activity and sedentary behavior in people with bipolar disorder: A systematic review and meta-analysis. In Journal of Affective Disorders (Vol. 201, pp. 145–152). 10.1016/j.jad.2016.05.020 [DOI] [PubMed] [Google Scholar]
  50. Weiner M, Warren L, & Fiedorowicz JG (2011). Cardiovascular morbidity and mortality in bipolar disorder. In Annals of Clinical Psychiatry (Vol. 23, Issue 1, pp. 40–47). [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1841824-supplement-1.pdf (144.4KB, pdf)

RESOURCES