Upcoming Event: Oden Institute & Dept. of Neuroscience

Computational Methods for Large-Scale Optical Control and Mapping of Neural Circuits

Marcus Triplett, Postdoctoral Research Scientist, Zuckerman Institute, Columbia University

Abstract

Discovering the structure and function of neural circuits is essential for understanding the brain. Crucially, recent technological developments in holographic optogenetics now enable the use of light to manipulate neural circuit activity with near single-cell resolution. This technology has the potential to transform our ability to investigate neural circuits, but has been critically limited by a lack of computational methods that can fully optimize its precision and throughput for neural circuit mapping. In this talk, I will introduce two methods that address this limitation. First, I will describe a novel "model-based" compressed sensing algorithm that enables large-scale mapping of synaptic connectivity using holographic stimulation and intracellular electrophysiology. I will show how this compressed sensing approach yields an order-of-magnitude improvement in mapping throughput compared to previous approaches, and present validation experiments in primary visual cortex. Second, I will introduce an algorithm for optimizing holographic stimulation to elicit specific neural activity patterns using Bayesian nonparametric methods, and present ongoing efforts to implement this technique experimentally to precisely modulate hippocampal activity in awake behaving mice.

Biography

Marcus is a postdoctoral research scientist in the Zuckerman Institute at Columbia University, where he works in the lab of Liam Paninski. He received a PhD in Computational Neuroscience from the University of Queensland, and before that studied Mathematics and Computer Science at the University of Auckland. Marcus' research focuses on the development and application of machine learning methods to discover the structure, dynamics, and function of neural circuits. To this end, his work at Columbia has concentrated on methods for optimally controlling and mapping neural circuits using cutting-edge optogenetic techniques, and is funded under an NIH K99 award. His PhD research was devoted to creating novel statistical methods for revealing low-dimensional structure and dynamics in neural population activity, as well as computational models of neural circuits to study theoretical mechanisms underlying neural development.