RetiCor · From retina to cortex: modeling the impact of eye movements on visual coding

The project aims to build the first biologically grounded model of visual cortical responses that incorporates both the nonlinear processing of the retina and the effects that eye movements have on cortical activity during natural vision. These movements engage complex biophysical mechanisms in the retina and thalamus that strongly shape cortical responses. Yet, current models of cortical activity largely ignore their impact due to technical difficulties. This limits both mechanistic insight and clinical applicability of cortical models.In the outgoing phase at UW, I will use electrophysiology recordings of V1 and LGN neurons in awake macaques from the Horwitz Lab, paired with simultaneously acquired high-resolution eye-tracking. Eye-position traces will allow me to reconstruct the dynamic retinal input during these experiments. The reconstructed inputs will then be presented to isolated macaque retinae in the Rieke Lab, in order to record matched retinal responses. I will use these matched datasets as input–output constraints for hierarchical computational models linking retinal activity to cortical responses under naturalistic conditions. In the return phase at IdV, I will adapt this framework to the mouse, in collaboration with the Marre and Rochefort labs. Comparing encoding models across mouse and primate will reveal which computations are conserved and which are species-specific.This project will deliver the first models of cortical coding that integrate eye movements with experimentally measured retinal output, advancing our understanding of visual processing under natural conditions. For me as a fellow, it will provide interdisciplinary training in electrophysiology and computational modeling, preparing me to become an independent leader at the interface of theory and experiment. The work has the potential to inform next-generation visual prosthetics and shed light on disorders of oculomotor-sensory integration, ensuring both scientific and societal impact