Machine learning and mechanistic modeling for understanding brain in health and disease


Breakthrough technology developments in semi-automated, high-throughput data collection have enabled experimental neuroscientists to acquire more multiscale neural data than ever before. However, the neural origin of the patterns observed in the multiscale, multimodal datasets are often difficult to decipher. There is therefore a critical need for time- and cost-efficient approaches to analyze and interpret the massive datasets to advance understanding of cellular and circuit-level origins of the observed neural dynamics in both health and disease, and to use the insights gained to develop new therapeutics. While machine learning is a powerful technique to integrate multimodal data, classical machine learning techniques often ignore the fundamental laws of physics and may therefore result in non-physical solutions. Multiscale modeling is a successful strategy to integrate multiscale, multiphysics data and unravel mechanisms that explain the emergence of function. However, multiscale modeling alone often fails to efficiently combine large data sets from different sources and different levels of resolution. This workshop aims to highlight research that bridges the disciplines of machine learning and multiscale modeling. Speakers are invited to address open questions, and discuss potential challenges and limitations in several topical areas: differential equations, data-driven approaches, and theory-driven approaches. This multidisciplinary perspective suggests that integrating machine learning and multiscale modeling can provide new insight into disease mechanisms, help identify new targets or treatment strategies, and inform decision making in the benefit of human health.

Video Presentations


William Lytton (SUNY Downstate;

Samuel Neymotin (Nathan Kline Institute;

Workshop Logistics

Date: July 22nd, 2020, Time: 10:30 AM - 4:15 PM (Eastern time), Schedule


Registration free: see