ourESRexplains #4

David explains new insights into the network architecture of the visual cortex and their relevance for vision restoration

Vision restoration approaches at the cortical level are the only option left if vision is lost due to an impaired connection between the retina and the visual cortex. Their realization requires a deep understanding of the cortical architecture but to date, not even the primary visual cortex (V1) – the first processing stage in the visual cortex – is sufficiently understood to explain how known cortical circuitry produces certain types of neural activity observed in experiments. Now, a recently published model of layer 4 – the entry point of visual information in V1 – explains a set of seemingly contradicting experimental observations.

In previous experiments, observing moving black and white stripes (so-called “drifting sinusoidal gratings”) modulated the activity of neurons located in layer 4. The modulation was dependent on the location in the visual field from which they were processing information, which is called receptive field of the neuron. Further, the modulation was anti-correlated between excitatory and inhibitory neurons which increase and decrease the activity in connected neurons respectively. When the same visual patterns, which evoked modulation of activity in both neuron types, were presented as flashing stimuli, the modulation was only observed in excitatory neurons while inhibitory neurons were firing independently.

A new study from our group  (Taylor et al. 2021) identified a V1 circuitry scheme that is able to reconcile these seemingly contradicting behaviors of neurons in layer 4 of V1. The model proposes that excitatory neurons preferentially connect to other neurons that respond to similar visual information in the visual field. Whereas, inhibitory neurons preferentially connect to other neurons that respond to opposite visual information in the visual field. Crucially, this functional bias of inhibitory connections is less strong than that of the excitatory ones.

These findings are not only a step towards a consistent model of the visual cortex which explains observed neural activity across a broad range of visual input statistics, but through a better understanding of the underlying V1 circuitry will also guide the development of stimulation protocols for future visual prosthetic systems. This is important, because besides the technological challenge of wireless stimulation hardware which couples to the brain in a long-term reliable way, the development of stimulation patterns which induce meaningful visual perceptions is a fundamental problem that still holds back vision restoration at the cortical level.

Read the full article here

© (2010) PLoS Computational Biology Issue Image | Vol. 6(8) August 2010. PLoS Comput Biol 6(8): ev06.i08. https://doi.org/10.1371/image.pcbi.v06.i08