Modeling Mood and Emotional Patterns from Speech in Bipolar Disorder-Presentation Slides

Presentation Slides

Pdf Summary

This document is a presentation summary for a webinar titled "Modeling Mood and Emotional Patterns from Speech in Bipolar Disorder", which was part of the Clinical Support System for Serious Mental Illness (CSS-SMI) initiative funded by the Substance Abuse and Mental Health Services Administration (SAMHSA) and implemented by the American Psychiatric Association (APA). The webinar discussed the use of speech analysis to identify markers of bipolar disorder and to monitor mood and emotional patterns in individuals with bipolar disorder. The presentation highlighted the importance of speech as a clinical data source and the potential for using speech patterns and emotions as predictors of illness and outcomes. The PRIORI project, a longitudinal study of bipolar disorder, was introduced, which collected and analyzed mood and emotion data from phone calls of individuals with bipolar disorder. The presentation described the annotation process for mood and emotion data and the use of different features and models for emotion recognition. The results showed that convolutional neural networks outperformed feed-forward neural networks in recognizing emotions. The presentation concluded by discussing the potential uses of the PRIORI Emotion Dataset, such as predicting the need for interventions and digital phenotyping based on passive longitudinal data. The document also included information about requesting consultations and upcoming webinars related to evidence-based care and physical health monitoring for individuals with serious mental illness.

Keywords

Modeling Mood and Emotional Patterns from Speech

Bipolar Disorder

Clinical Support System for Serious Mental Illness

Substance Abuse and Mental Health Services Administration

American Psychiatric Association

Speech Analysis

Markers of Bipolar Disorder

Mood and Emotional Patterns

PRIORI Project

Convolutional Neural Networks