Skip to main content
Talk

Open Digital Phenotyping for Psychiatry Research

Dr. Omar Costilla Reyes presented his team’s open digital-phenotyping platform for psychiatry research at the 2020 Technology in Psychiatry Summit. The summit was organized by McLean Hospital.

Banner of the 2020 Technology in Psychiatry Summit, organized by McLean Hospital’s Institute for Technology in Psychiatry. Credit: Image courtesy of Dr. Omar Costilla Reyes.
Title slide of the team’s open digital-phenotyping presentation at TIPS 2020; the authors listed are Omar Costilla-Reyes, Kayla Villa, Elazer R. Edelman, and Mercedes Balcells-Camps. Credit: Image courtesy of Dr. Omar Costilla Reyes.
Methods graphic from the team’s December 2023 announcement: self-reports and passively recorded sensor measures are filtered by behavior and compared as networks in an n-of-1 analysis. Related photograph from Individual behavioral insights in schizophrenia: A network analysis and mobile sensing approach. Credit: Image courtesy of Dr. Omar Costilla Reyes.

Presenting at a psychiatry technology summit

The Technology in Psychiatry Summit (TIPS) 2020 was organized by McLean Hospital’s Institute for Technology in Psychiatry at Harvard Medical School and held online. On October 23, 2020, Dr. Omar Costilla Reyes, then a postdoctoral fellow at MIT’s Institute for Medical Engineering and Science (IMES), announced that he would present his research on digital phenotyping and behavioral health there. After the first day, October 28, he reported strong speakers and presentations and invited attendees to the continuation of his team’s presentation on their platform for open digital phenotyping the next day.

The title slide of the presentation, “Building an open-source machine learning pipeline to understand multimodal digital phenotyping sensor data to advance psychiatry research,” lists Omar Costilla-Reyes, Kayla Villa, Elazer R. Edelman, and Mercedes Balcells-Camps of IMES as its authors. Digital phenotyping describes a person’s behavior in everyday life from data that personal devices such as smartphones already record. “Multimodal” means that several kinds of signals are recorded together, and a pipeline is the sequence of steps that turns raw sensor streams into cleaned measurements and then into models. Making the pipeline open source means other researchers can read, reuse, and check each step. That matters in psychiatry, where sensor data are rich but noisy, participants’ lives differ widely, and an open pipeline lets teams design privacy-conscious studies and compare methods transparently.

Related work

The same research line continues in Dr. Costilla Reyes’s later digital mental-health publications, including “Counterfactual explanations and predictive models to enhance clinical decision-making in schizophrenia using digital phenotyping” (2023) and “Individual behavioral insights in schizophrenia: A network analysis and mobile sensing approach” (2024). His interpretable machine-learning studies of smartphone-delivered interventions ask which baseline characteristics predict who improves in digital mental-health treatment. In the same weeks he was also invited to speak on digital phenotyping by the Universidade de São Paulo and the Tecnológico Nacional de México, as described on their own pages. The summit presentation brought that approach to a psychiatry audience: a transparent, reusable way to turn everyday smartphone and sensor data into measurements that clinicians and researchers can examine together.

Topics