Towards continuous monitoring of depressive mood from locomotor activity in major depressive disorder

In this study, we measured self-reported depressive mood using Ecological Momentary Assessment and locomotor activity simultaneously in healthy subjects (n=43) and patients with major depressive disorder (MDD; n=14). Multilevel modeling approach showed significant associations of changes between depressive mood scores and intermittency of local locomotor activity, reflecting situations in which the depressive mood becomes worse with both lower activity levels and more frequent episodes of bursts of locomotor activity. Further, the personalized optimization method enhanced the prediction accuracy of the constructed model in MDD patients. These results suggest the possibility for continuous monitoring of depressive mood based on locomotor activity. © 2014, Japan Soc. of Med. Electronics and Biol. Engineering. All rights reserved.