My Voices, In My Voice: Natural Language Processing and Computational Modeling of Audio Diaries

Auditory Verbal Hallucinations (AVH) occur across a range of recognized mental health conditions, and cause significant impairment and distress in some but not all cases. In this project we will study participants’ real-time “audio diary” descriptions of their experience, which represent a unique and unprecedented resource for the study of AVH covering a broad range of AVH types (e.g. single vs. multiple voice), perceptions (e.g. beneficent vs. malevolent) and emotional responses (e.g. anger vs. fear). We will analyze these data using innovative computational methods to detect disorganized thinking, and the emotional state of participants.

Funded by: National Institute of Mental Health: 3R01MH112641-03S2