Computational Methods for Precise Holographic Control and Mapping of Neural Circuits

PROJECT ABSTRACT Precisely mapping the structure of synaptic connectivity offers a pathway to understanding the function of neural circuits and how they go awry in neurological disease. Holographic optogenetics has recently emerged as a revolutionary technology capable of optically probing the synaptic structure and function of neural circuits through the stimulation of neurons with near single-cell precision. However, despite its success, this technology is currently greatly underutilized due to a lack of computational methods capable of modeling these complex experiments and interpreting the data they generate. Thus, there is a need for new computational methods to maximize the scientific impact of holographic optogenetics in neuroscience research. In recent work, I have created a machine learning algorithm for inferring monosynaptic connectivity from holographic stimulation of specifically targeted populations of neurons. This algorithm increased the rate of in vitro monosynaptic connectivity mapping by more than an order of magnitude over existing approaches using optogenetic stimulation, demonstrating the transformative effect that computational methods can have in this scientific domain. Building on this advance, I now propose to develop a set of computational methods enabling holographic control and mapping of neural circuits in vivo and at unprecedented detail, scale, and precision. In Aim 1 of this proposal, I will develop a real-time target optimization algorithm to select holographic stimulation parameters that eliminate the unintended activation of non-target neurons when attempting to probe the connectivity or function of neural circuits (K99). In Aim 2, I will extend my earlier connectivity mapping inference approach by using calcium imaging to discriminate the presynaptic origin of postsynaptic responses at extremely fine detail (K99/R00). Finally, in Aim 3, I will develop a computational approach that leverages voltage imaging to all-optically map tens of thousands of potential recurrent connections between hundreds of neurons (R00). Together, these aims provide practical tools enabling high-throughput collection of large-scale maps of synaptic connectivity within individual experiments. Using these tools to probe the structure and function of neural circuits could ultimately shed light on the etiology of neurological diseases. During the K99 phase, this work will be conducted at the Zuckerman Mind Brain Behavior Institute at Columbia University, where I will receive scientific training in computational neuroscience under Dr. Liam Paninski, a leading authority in computational methods for neural data. Additionally, I will receive training in experimental neuroscience from Drs. Hillel Adesnik and Adam Cohen, who are experts in holographic optogenetics and voltage imaging. Their combined scientific expertise and impressive track record of transitioning postdoctoral scientists to faculty positions make them the ideal mentorship team for my goal of becoming an independent group leader working on computational methods for the optical interrogation of neural circuits.