Neural field model of receptive field restructuring in primary visual cortex.

Area of Science:

• Computational Neuroscience
• Visual System Physiology
• Neural Field Modeling

Background:

• Receptive fields (RFs) in the visual cortex alter size based on individual state, impacting visual resolution during drowsiness.
• Previous hypotheses suggested state-dependent lateral geniculate nucleus (LGN) firing patterns cause RF restructuring, but mechanisms remain untested.

Purpose of the Study:

• To analytically describe changes in cortical receptive field properties using a neural field approach.
• To investigate the underlying mechanisms responsible for state-dependent receptive field size changes in the visual cortex.

Main Methods:

• Developed a neural field model to analytically describe spatiotemporal receptive field properties in feedforward and recurrent networks.
• Derived expressions for receptive fields and analyzed membrane potential profiles.
• Fitted analytically derived point-spread functions to experimental data to differentiate between feedforward and recurrent mechanisms.

Main Results:

• The model accurately reproduced receptive field restructuring effects and predicted nonlinear increases in visual latency with stimulus distance.
• In feedforward networks, the spatial membrane potential profile width remained constant despite changing RF sizes.
• Recurrent network models showed RF and membrane potential profile widths could increase over time, dependent on connection strengths.

Conclusions:

• The observed receptive field sharpening in the visual cortex is primarily driven by input from the lateral geniculate nucleus (LGN).
• Recurrent intracortical connections are less likely to be the main cause of these state-dependent receptive field changes.
• Model parameter estimates align with existing literature, supporting the feedforward-driven hypothesis for RF sharpening.