ePoster

Understanding sensory-motor integration in the PIVC through ensemble analysis

Adam Charles, Lex Gomez, Noga Mudrik, Kathleen Cullen
COSYNE 2025(2025)
Montreal, Canada

Conference

COSYNE 2025

Montreal, Canada

Resources

Authors & Affiliations

Adam Charles, Lex Gomez, Noga Mudrik, Kathleen Cullen

Abstract

Understanding how the brain encodes movements is vital for motor control. Research suggests that this involves comparing proprioceptive inputs from the body with an internal model of expected movement. When these signals match, vestibular signals---which would otherwise stabilize the head against movements---are suppressed, allowing voluntary movement to occur. This mechanism implies that integrating proprioceptive and vestibular inputs is crucial for effective coordination. While previous studies indicate the involvement of multiple brain areas, with the PIVC at the forefront, the mechanisms by which it integrates these signals remain unclear. Here, we simultaneously recorded the activity of dozens of PIVC neurons from rhesus monkeys using Neuropixels during various movements designed to test sensitivity to vestibular vs. proprioceptive inputs. However, understanding how PIVC integrates sensory-motor signals from these recordings requires careful consideration of dimensionality, noise, and trial variability, emphasizing the need for tailored population-level analyses. We build on [1] to explore PIVC integration through neural ensembles, analyzing activity while accounting for trial variability and composition adjustments across movements. We further distinguished movement-style specificity from direction-encoding ensembles, identifying functional PIVC ensembles that encode specific movements, with some adjusting to movement styles and others varying by movement direction. We further discovered that certain ensembles capture similarities in different movement conditions that reflect the usage of a shared system (e.g., proprioception). We observed greater ensemble trace similarities within the same condition trials compared to different conditions, with some ensembles generalizing across movements for background processing. The magnitude of per-ensemble cross-condition traces revealed patterns that capture similarities in system usage, e.g., one ensemble reflects active proprioception while another shows reduced activity for only-head movements, regardless of whether they are passive or active.

Unique ID: cosyne-25/understanding-sensory-motor-integration-2d8b6244