Zoom link: https://mit.zoom.us/j/91598940526

Speaker: Antoine de Comite

Affiliation: Seethapathi Motor Control Group

Title: Unifying predictive and reactive control of stability across species

Abstract: How do animals maintain stability in the face of errors? While predictive “feedforward” control that leverages passive mechanical stability is undeniably important, the existence of reactive “feedback” control across species remains largely unexplored. This talk puts forth a unified control structure combining predictive and reactive error-corrective mechanisms, with locomotion as our model task. We present data-driven evidence supporting this hypothesized control structure in humans, mice, and flies, revealing how the control strategies vary with neuromechanical embodiment. Further, we demonstrate the potential of this unified control structure to identify otherwise subtle phenotypes in mouse models of human neurodevelopmental disorders, specifically autism spectrum disorder (ASD). Our data-driven framework reveals a double dissociation in computationally distinct control modules for predictive versus reactive control depending on which brain regions are impacted in the genetic models. Ultimately, identifying shared control principles across species advances comparative neuroscience and helps develop generalizable diagnostics.