Retinotopic Mechanics derived using classical physics

TL;DR

This paper introduces retinotopic mechanics, a new computational framework suggesting that receptive fields actively shift via neurobiologically inspired force fields during active sensing, enhancing understanding of spatial constancy in vision.

Contribution

It proposes a novel dynamic model where receptive fields are influenced by elastic force fields, extending static receptive field descriptions to account for active sensing and spatial constancy.

Findings

Receptive fields can shift beyond classical extents during active sensing.

Elastic force fields are proposed to predictively move receptive fields.

The model explains neural mechanisms supporting spatial constancy during eye movements.

Abstract

The concept of a cell$'$s receptive field is a bedrock in systems neuroscience, and the classical static description of the receptive field has had enormous success in explaining the fundamental mechanisms underlying visual processing. Borne out by the spatio-temporal dynamics of visual sensitivity to probe stimuli in primates, I build on top of this static account with the introduction of a new computational field of research, retinotopic mechanics. At its core, retinotopic mechanics assumes that during active sensing receptive fields are not static but can shift beyond their classical extent. Specifically, the canonical computations and the neural architecture that supports these computations are inherently mediated by a neurobiologically inspired force field (e.g.,$R_s\propto \sim 1 /\Delta M$). For example, when the retina is displaced because of a saccadic eye movement from one point in space to another, cells across retinotopic brain areas are tasked with discounting the retinal disruptions such active surveillance inherently introduces. This neural phenomenon is known as spatial constancy. Using retinotopic mechanics, I propose that to achieve spatial constancy or any active visually mediated task, retinotopic cells, namely their receptive fields, are constrained by eccentricity dependent elastic fields. I propose that elastic fields are self-generated by the visual system and allow receptive fields the ability to predictively shift beyond their classical extent to future post-saccadic location such that neural sensitivity which would otherwise support intermediate eccentric locations likely to contain retinal disruptions is transiently blunted.