Intermittent locomotor dynamics and its transitions in bipolar disorder

The scale-invariant and intermittent dynamics of human behavior are currently attracting great scientific interest. We have recently reported the universal laws of behavioral organization, especially those describing the complexity of behavioral dynamics, shared by humans and wild-type mice. We also demonstrated the alterations of the law of resting period distributions shared among patients with major depressive disorders, mice with deficiency in a circadian clock gene (Period 2), and patients with schizophrenia. These findings indicate that the statistical laws of behavioral organization provide an objective biobehavioral measures for psychiatric disorders, and the study of behavioral organization could provide further insight into the pathophysiological mechanism. Based on these studies, we recently started ultra long-term continuous recording of locomotor activity and self-reported symptoms for bipolar patients to capture transitions of pathological states (i.e. switching processes between depressive and manic/hypomanic phases), and then succeeded in obtaining a simultaneous recording of dynamical changes in mood and locomotor activity from one patient during a pathological phase transition. To our knowledge, this is the first successful case for such a long-term, high frequency and high resolution recording enabling the study of behavioral dynamics in transitions. In this paper, we show characteristic phenomena in the period preceding the transition from depressive to hypomanic phase, suggestive of the existence of an early warning sign (referred to as critical slowing down) before such a transition. In addition, we demonstrate the significant correlation between mood scores and scaling exponents of resting period distributions. These findings suggest a possibility for quantitative evaluation and/or prediction of pathological states and their transitions in bipolar disorder.