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

Video Transcription

Video Summary

The video is a presentation by Dr. Melvin McGinnis on the topic of modeling mood and emotional patterns from speech in bipolar disorder. Dr. McGinnis is the Thomas B. and Nancy F. John Woodworth Professor of Bipolar Disorder and Depression and the Director of the Heinz C. Prechter Bipolar Research Program. The presentation is part of the SMI Advisor webinar series, which is an initiative devoted to helping clinicians implement evidence-based care for those living with serious mental illness.<br /><br />Dr. McGinnis discusses the use of speech as a proxy measure for internal emotional mood and affective states in individuals with bipolar disorder. He explains research that has been done to extract acoustic features from speech and use machine learning algorithms to categorize the speech as either manic or depressive. He also discusses the correlation between emotional measures (activation and valence) and mood severity, as well as the potential use of speech analysis for predicting and intervening in mood episodes. Dr. McGinnis presents findings from studies using convolutional neural networks to analyze the acoustic patterns of speech and identify anomalies that could suggest the need for intervention.<br /><br />The presentation highlights the unique emotional dataset that has been created for this research and acknowledges the support of various institutions and participant collaborators. Dr. McGinnis concludes by stating the potential uses of this research, including predictive and prognostic interventions for mood episodes, digital phenotyping, and measuring mood instability.

Keywords

Dr. Melvin McGinnis

Modeling mood

Emotional patterns

Speech analysis

Bipolar disorder

Machine learning algorithms

Convolutional neural networks

Digital phenotyping