Guillermo A. Cecchi, Research Staff Member, Computational Neuroscience at IBM, discusses his background and the current state of computational neuroscience research.
Dr. Cecchi has an extensive background in Physics (MSc, University of La Plata, Argentina); in Physics and Biology (PhD, The Rockefeller University); as well as in Imaging in Psychiatry (Postdoctoral Fellow, Cornell University). In this podcast, Cecchi discusses much of his current research. Dr. Cecchi has broken ground in the use of a computational linguistics approach to assess psychiatric conditions. He provides an overview of his work in computational psychiatry and the use of neuroimaging.
Dr. Cecchi’s particular interests are in the varied aspects of theoretical biology, such as brain imaging, molecular computation, Brownian transport, spike reliability in neurons, the song production and representation in diverse songbirds, statistics of natural images as well as visual perception, and more.
The computational neuroscientist discusses their work with adolescents who have a clinical risk for psychosis. As he explains, they conducted baseline interviews then followed patients for several years, to see if it is possible to predict which patients would have dramatic psychotic outbreaks, in an attempt to better understand how to work with patients, predict upcoming problems, and provide better early treatment.
Dr. Cecchi explains how they observe speech and patterns, and interpret signs and signals. The research doctor goes on to discuss parameters they consider in their studies, and as he states, certain factors can be significant. He explains studies that have been successful in the identification of speech markers, and the factors that are involved, looking at baseline, and considering future trials.
Dr. Cecchi cites further examples of their work researching ALS, and how they use data collected from smartphones, in order to observe deterioration, motor activity, and cognitive aspects over time. As Dr. Cecchi states, the goal on some of their work is to create a model that can be brought to the FDA, in order to eventually provide better treatment.