Conjunctive tuning and cortical geometry shape predictive visual remapping.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Visual Perception

Background:

• Perceptual continuity across eye movements (saccades) relies on pre-saccadic receptive field (RF) remapping.
• This remapping is driven by corollary discharge (CD) signals, but the mechanism for accurate remapping in non-uniformly represented visual space is unclear.

Purpose of the Study:

• To develop and validate a computational model explaining accurate receptive field remapping during saccades.
• To investigate how the brain's non-uniform visual representation impacts remapping accuracy.
• To reconcile visual acuity and continuity demands within the visual cortex.

Main Methods:

• Developed a recurrent neural network model integrating CD signals for conjunctive tuning to location and saccade direction.
• Incorporated pre-saccadic suppression to stabilize remapping and counteract direction selectivity distortions.
• Recorded neural activity in macaque V2 during cued saccades to validate model predictions.

Main Results:

• The model successfully preserved cell-cell RF relationships and constrained population dynamics.
• Neural recordings confirmed saccade direction selectivity and preserved RF relationships in V2 neurons.
• Identified eccentricity-dependent remapping errors arising from non-uniform cortical representation, a prediction supported by empirical data.
• Demonstrated that countervailing distortions in cortical representation can minimize remapping errors.

Conclusions:

• The study presents a novel model for accurate pre-saccadic remapping, explaining how visual continuity is maintained.
• Reveals previously unknown properties of visual neurons and their role in visual perception.
• Reconciles the trade-offs between visual acuity and continuity in the visual cortex.