The Promise of Crowd-sourcing and Artificial Intelligence in Psychiatric Neuroscience

Greg Siegle, Ph.D.
Associate Professor, Department of Psychiatry
University of Pittsburgh School of Medicine

Abstract: Neuroscience has increasingly informed our understanding of how to conceptualize, diagnose, and treat mental disorders at the group level. But at the level of individuals, neuroscience has not yet begun to strongly inform treatment. I will provide a snapshot of a few recent studies from our lab suggesting that combining neuroscience and artificial intelligence models can inform treatment decisions for single patients. Given the need to match large amounts of data with idiosyncratic presentations, I will contend that this nascent field could advance particularly rapidly by crowd-sourcing large numbers of scientists and non-scientists to match results from the literature with each patient's imaging data and using artificial intelligence not only at the decision-making stage but for communication of these decisions to patients and providers. I will describe two ongoing projects in which we have begun to use these strategies. The first involves generating natural-language interpretations of a patient's brain reactivity as might be relevant to treatment using a crowd-sourced database. The second involves interpreting clinical electroencephalography data in light of profiles generated using crowd-sourced data from consumer-grade headsets. In conclusion, I will consider directions for future progress, novel ethical considerations raised by this approach, and challenges to dissemination to keep in mind as the next generation of psychiatry is formulated.




















This Web Page Created with PageBreeze Free HTML Editor / Web Hosting