Behavioral Signal Processing

Behavioral Signal Processing methods provide the basis for modeling human behavior with high resolution. Computational methods developed for Behavioral Signal Processing applications can provide deeper understanding of behavioral phenomena, especially useful for neuropsychiatry. I am developing methods to support dynamic analysis of clinical interviews, online data collection, and neuropsychiatric batteries. These methods are enabling understanding of mechanisms which underlie several disorders including dementia, schizophrenia, trauma, and depression. My work features analysis on the Framingham Heart Study Cognitive Aging Cohort, DCAPS Distress Corpus, and Mental Health America.

Behavioral and Physiological Synchrony contribute to our understanding several prosocial behaviors, including shared attention in conversations, and cooperation during collaborative tasks. I study methods surrounding the integration of multiple data modalities (e.g. heart rate, head and body pose, vocalizations) to provide robust markers of synchrony within dyadic interactions.