DERI Seminar with Andrea Benucci
When: Thursday, June 27, 2024, 11:00 AM - 12:00 AM
Where: Zoom
Speaker: Prof Andrea Benucci, School of Biological and Behavioural Sciences, QMUL
Zoom link: https://qmul-ac-uk.zoom.us/j/81148100921
Title: Motor-related signals support localization invariance for stable visual perception
Our ability to perceive a stable visual world in the presence of continuous movements of the body, head, and eyes has puzzled researchers in the neuroscience field for a long time. We reformulated this problem in the context of hierarchical convolutional neural networks (CNN) whose architectures have been inspired by the hierarchical signal processing in the mammalian visual system. We examined perceptual stability as an optimization process in networks trained to accurately classify images in the presence of self-generated movements (simulated saccades). Motor-related activations multiplexed with visual inputs along overlapping convolutional layers and carrying information about self-generated movements, aided classification invariance by making the classification faster to learn and more robust relative to input noise. Classification invariance was reflected in activity manifolds associated with image categories emerging in late CNN layers and with network units acquiring movement-associated activity modulations as observed experimentally during saccadic eye movements. Our findings provide a computational framework that unifies a multitude of biological observations on perceptual stability under optimality principles for image classification in artificial neural networks.