Inferring computational principles in neural population recordings of evolutionarily-divergent species through artificial intelligence
Neuroscientists seek to understand how the brain generates cognition and behavior. A critical piece of this goal is fundamental research in model organisms that provide high degrees of experimental access. This spans evolutionarily older species with smaller nervous systems (e.g., worms and fish) to newer species with advanced cognitive abilities (e.g., primates). We hope that by uncovering fundamental principles in such model organisms, we can translate the findings to human patients and improve our ability to diagnose and treat brain disorders.
Such a link, though seemingly implausible, can be a reality. The brain of each individual of a species—be it fish, mouse, or human—specializes during its development into highly specialized features such as types of neurons, or functionally-distinct brain regions. Many of these features are remarkably preserved throughout the animal kingdom, reflecting a common evolutionary origin across species. Despite this presumed evolutionary origin, a clear neural computational signature that unifies across individuals, let alone species, has been elusive. In this proposal, I will apply modern artificial intelligence (AI) techniques to infer such principles in neural recordings from different species across the animal kingdom. This project will be foundational work towards the identification of commonalities in brain function that spans across species. This level of understanding will enhance the direct translation of fundamental research in model organisms to improve our understanding of human health and disease.
Matthew Perich , Université de Montréal
Partners and Donors