A fundamental law underlying predictive remapping

Abstract

ABSTRACTPredictive remapping (R) — the ability of cells in retinotopic brain structures to transiently exhibit spatiotemporal shifts beyond the spatial extent of their classical anatomical receptive fields — has been proposed as a primary mechanism that stabilizes an organism’s percept of the visual world around the time of a saccadic eye movement. Despite the well-documented effects ofR, a biologically plausible mathematical abstraction that specifies a fundamental law and the functional architecture that actively mediates this ubiquitous phenomenon does not exist. I introduce the Newtonian model ofR, where each modular component ofRmanifests as three temporally overlapping forces - a centripetal, convergentand translational force, that perturb retinotopic cells from their equilibrium extent. The resultant and transient influences of these forcesgives rise to a neuronal force field that governs the spatiotemporal dynamics ofR. This neuronal force field fundamentally obeys an inverse-distance law, akin to Newton’s law of universal gravitation [1] and activates retinotopic elastic fields (elφs). I posit that elφs are transient functional structures that are self-generated by a visual system during active vision and approximate the sloppiness (or degrees of spatial freedom) within which receptive fields are allowed to shift while ensuring that retinotopic organization does not collapse. The predictions of the proposed general model are borne out by the spatiotemporal changes in sensitivity to probe stimuli in human subjects around the time of a saccadic eye movement and qualitatively match neural signatures associated with predictive shifts in the receptive fields of cells in premotor and higher-order retinotopic brain structures.