Activity Recognition for Behavioral Activation in Depression with Artificial Intelligence

Abstract

Behavioral Activation is a method in Cognitive Behavioral Therapy which uses behavior to influence the emotional condition of the person. Behavioral Activation aids in engaging activities to activate a positive emotional state and overcome the depression. In this paper, we showcase a new method to recognize the right activity for behavioral activation by detecting emotion, sentiment and understanding the context and interests of the person during counseling. We showcase a multi-modal method to recognize activity for behavioral activation through speech and text modalities using artificial intelligence. Projected model attained an accuracy of 83% for emotion recognition, 81% for sentiment detection and 82% for identifying the right activity for behavioral activation.