Fractal dynamics in spontaneous motor activity associate with effective antidepressant chronotherapeutic response, cerebellar volumes, and white matter microstructure in bipolar disorder

Aim: Bipolar disorder (BD) is characterized by circadian dysregulation, altered sleep–wake behaviors, and heterogeneous antidepressant response. Chronotherapeutic interventions such as total sleep deprivation combined with light therapy (TSD + LT) produce rapid antidepressant effects, but predictors of outcome are lacking. We investigated whether fractal dynamics of motor activity, indexed by the short-term scaling exponent α₁, are associated with antidepressant response, sleep–wake regulation, and structural brain measures and may serve as a biomarker of treatment responsiveness. Methods: Sixty-eight BD inpatients underwent three cycles of TSD + LT. Depressive symptoms were assessed with the Hamilton Depression Rating Scale (HDRS), and treatment response was defined as a post-treatment HDRS score <8. α₁ was derived from actigraphy using detrended fluctuation analysis, with values close to 1 indicating optimal fractal regulation. α₁ was computed before and after TSD + LT, and Δα₁ represented treatment-related change. Sleep–wake and circadian parameters were examined. A subsample underwent multimodal magnetic resonance imaging (MRI). Results: TSD + LT induced clinical improvement in 65.3% of patients. Responders showed a significant reduction in α₁ toward values close to 1, whereas nonresponders did not. Δα₁ correlated with symptom improvement (P = 0.035) and was independently associated with treatment response (P = 0.039). Baseline α₁ was associated with sleep–wake parameters, cerebellar gray matter volume, and white-matter microstructure, including cerebellar peduncles and thalamo-cortical tracts. Circadian rhythmicity increased in responders and decreased in nonresponders. Conclusion: Fractal motor activity tracks clinical changes during chronotherapeutic treatment in BD and may represent a candidate marker of treatment-related behavioral regulation. Its association with cerebellar and white-matter measures suggests multimodal relevance, warranting validation in larger longitudinal studies.