Long–Short Ensemble Network for Bipolar Manic-Euthymic State Recognition Based on Wrist-Worn Sensors

Manic episodes of bipolar disorder can lead to uncritical behavior and delusional psychosis, often with destructive consequences for those affected their surroundings. Early detection intervention a manic episode are crucial prevent escalation, hospital admission, premature death. However, people may not recognize that they experiencing symptoms, such as euphoria increased productivity also deter individuals from seeking help. This work proposes perform user-independent, automatic mood-state based on actigraphy electrodermal activity acquired wrist-worn device during mania after recovery (euthymia). article new deep learning-based ensemble method leveraging long (20 h) short (5 min) time intervals discriminate between the mood states. When tested 47 patients, proposed classification scheme achieves an average accuracy 91.59% in euthymic/manic recognition.