DeepCoMechTome

Using deep learning to understand computations in neural circuits with Connectome-constrainedMechanistic Models

01/06/2023 to 31/05/2028

Advances in experimental techniques yield detailed wiring diagrams of neural circuits in model-systems such as the Drosophila melanogaster. How can we leverage these complex connectomes,together with targeted recordings and perturbations of neural activity, to understand how neu-ronal populations perform computations underlying behavior? Achieving a mechanistic under-standing will require models that are consistent with connectomes and biophysical mechanisms,while also being capable of performing behaviorally relevant computations. Current models fail toaddress this need: Mechanistic models satisfy anatomical and biophysical constraints by design,but we lack methods for optimizing them to perform tasks. Conversely, deep learning modelscan be optimized to perform challenging tasks, but fall short on mechanistic interpretability.To address this challenge, we will provide a machine learning framework that unifies mechanisticmodeling and deep learning, and will make it possible to algorithmically identify models thatlink biophysical mechanisms, neural data, and behavior. We will use our approach to study twokey neural computations in D. melanogaster. We will build large-scale mechanistic models of theoptic lobe and motor control circuits which are constrained by connectomes and physiologicalmeasurements, and optimize them to solve specific computational tasks: Extracting behaviorallyrelevant information from the visual input, and coordinating leg movements to achieve robustlocomotion. Our methodology for building, interpreting and updating these ‘deep mechanisticmodels’ will be applicable to a wide range of neural circuits and behaviors. It will serve asa powerful hypothesis generator for predicting neural tuning and optimizing experimental per-turbations, and will yield unprecedented insights into how connectivity shapes efficient neuralcomputations in biological and artificial networks.