Individual-specific change points in circadian rest-activity rhythm and sleep in individuals tapering their antidepressant medication: an actigraphy study

Abstract

Group-level studies showed associations between depressive symptoms and circadian rhythm elements, though whether these associations replicate at the within-person level remains unclear. We investigated whether changes in circadian rhythm elements (namely, rest-activity rhythm, physical activity, and sleep) occur close to depressive symptom transitions and whether there are differences in the amount and direction of circadian rhythm changes in individuals with and without transitions. We used 4 months of actigraphy data from 34 remitted individuals tapering antidepressants (20 with and 14 without depressive symptom transitions) to assess circadian rhythm variables. Within-person kernel change point analyses were used to detect change points (CPs) and their timing in circadian rhythm variables. In 69% of individuals experiencing transitions, CPs were detected near the time of the transition. No-transition participants had an average of 0.64 CPs per individual, which could not be attributed to other known events, compared to those with transitions, who averaged 1 CP per individual. The direction of change varied between individuals, although some variables showed clear patterns in one direction. Results supported the hypothesis that CPs in circadian rhythm occurred more frequently close to transitions in depression. However, a larger sample is needed to understand which circadian rhythm variables change for whom, and more single-subject research to untangle the meaning of the large individual differences.

Figure 1

Example of a cosinor curve and terms used to describe a diurnal rhythm: MESOR, amplitude, and acrophase.

Figure 2

Flowchart of participants with and without a transition in depressive symptoms. Note. For the 56 participants who completed 4 months of monitoring, definitions for the transitions were based on the weekly SCL-90 depression subscale scores, evaluation interview, and self-reported experience of recurrence of depressive symptoms. Of the 50 participants with available actigraphy data, individuals with a transition after the end of actigraphy monitoring (n = 6) and with less than 90 days of observations (n = 10) were excluded as it appeared to be insufficient for the KCP-RS analysis to reliably identify CPs in the data. With approximately 120 days of observations per person and ten variables included in the change point analysis, we expect to accurately detect (moderate and) large changes in circadian rhythm and sleep variables.

Figure 3

Running means for all variables, the transition period, and the Change Point (CP) for participant 1. Note. IS interdaily stability, IV intradaily variability, RA relative amplitude, L5 least active 5 h, MESOR a rhythm-adjusted mean, amplitude the difference between a peak and a mean level of the rhythm, acrophase a moment during the day when the peak of the rhythm occurs, time in bed time duration between going to bed and getting out of bed, sleep efficiency the ratio of total sleep time to time in bed, fragmentation index the amount of movement or restlessness in a sleep period. Grey shaded block represents the week of transition in depressive symptoms; the red dotted line represents the identified CP.