Linking bipolar disorders to metabolism through machine learning-based behavioral analysis

In collaboration with the Harvard Brain Initiative Bipolar Seed Grant Program, the goals of this project were to use Motion Sequencing, an unsupervised behavioral discovery platform, to characterize the behavior of two mouse models of bipolar disorder in which behavioral deficits are linked to metabolic changes (hyperthyroid and AKAP11 deleted mice), and to characterize the behavior of patients with bipolar disorder in which endocrine and metabolic parameters are measured.