SIAM News Blog

Automatic Behavioral Analysis for Computational Psychiatry at Home

By Guillermo Sapiro

Guillermo Sapiro, Duke University.
According to the Centers for Disease Control and Prevention, autism affects one in 68 children, with higher incidences among males and siblings of children on the Autism spectrum. Autism can be diagnosed in children as young as 18 months, but the current average diagnosis age in the United States is about five years old. Waiting time to see an autism specialist in leading U.S. hospitals can be up to 18 months, and about 70 percent of counties in the state of North Carolina lack a specialist. South Saharan Africa has an estimated half billion children, served by only about 50 specialists. These are just a few examples illustrating the challenges of developmental and mental health screening, diagnosis, and monitoring, where the introduction of algorithms and technology is lagging far behind other medical disciplines.

Despite major advances in magnetic resonance imaging and genetics, behavioral observation is still the gold standard for monitoring developmental disorders and mental health. There is thus a need to develop mathematical and computational tools, which can operate on standard hardware such as smartphones, to automatically analyze behaviors and make diagnosing technology widely available to the general population. This calls for a unique teaming of psychiatrists, mathematicians, engineers, and developers to create the right stimuli to evoke appropriate behaviors that can be properly understood by mathematics and algorithms, all integrated into mobile devices for stimuli presentation, sensing, and analysis. The necessary tools range from emotion analysis, understanding, and gaze tracking to audio interpretation, pressure sensitivity, and more standard bio-signals, such as heart rate. The integration of all sensing modalities available in today’s ubiquitous mobile devices can thus enable personalized and at-home screening, diagnosis, and tracking of developmental and mental health disorders, opening the door to new developments with great societal impacts.

Guillermo Sapiro’s research team is developing apps to improve mental health monitoring in pediatric clinics. Sapiro will present his team’s advances at the 2018 SIAM Annual Meeting.

During the 2018 SIAM Annual Meeting, I will present our interdisciplinary team’s demonstrated advances in the development of apps—currently being used in pediatric clinics and downloadable from iTunes—for mental health monitoring. Our work has produced the largest existing dataset of child behavior recorded in natural environments.

Guillermo Sapiro is the Edmund T. Pratt, Jr. School Professor of Electrical and Computer Engineering at Duke University.
blog comments powered by Disqus