Which actigraphic variables optimally characterize the sleep-wake cycle of individuals with bipolar disorders?

To examine which combination of objectively measured actigraphy parameters best characterizes the sleep-wake cycle of euthymic individuals with bipolar disorder (BD) compared with healthy controls (HC).

Sixty-one BD cases and 61 matched HC undertook 21 consecutive days of actigraphy. Groups were compared using discriminant function analyses (DFA) that explored dimensions derived from mean values of sleep parameters (Model 1); variability of sleep parameters (2); daytime activity (3); and combined sleep and activity parameters (4). Exploratory within-group analyses examined characteristics associated with misclassification.

After controlling for depressive symptoms, the combined model (4) correctly classified 75% cases, while the sleep models (1 and 2) correctly classified 87% controls. The area under the curve favored the combined model (0.86). Age was significantly associated with misclassification among HC, while a diagnosis of BD-II was associated with an increased risk of misclassifications of cases.

Including sleep variability and activity parameters alongside measures of sleep quantity improves the characterization of cases of euthymic BD and helps distinguish them from HC. If replicated, the findings indicate that traditional approaches to actigraphy (examining mean values for the standard set of sleep parameters) may represent a suboptimal approach to understanding sleep-wake cycles in BD.