Computation of Contrast Adaptation and Normalization

Computation of Contrast Adaptation and Normalization

Sensory neurons encode stimulus intensity in their spike rate and adjust their sensitivity by adaptation. In the visual cortex, adaptation is crucial because the dynamic range of individual neurons is far narrower than the range of contrasts encountered in natural scenes. The dynamic range is shaped by normalization, which has been modeled by using the averaged local excitation to drive recurrent divisive inhibition.

In this project, we used two-photon calcium imaging of local neuronal populations in cat primary visual cortex and found a surprisingly wide range of CRFs, which effectively extends their collective dynamic range. We observed that as the stimulus contrast increases, members of the network do not just increase their firing in unison, but also new players are recruited to join the active pool of neurons. Therefore, we argue that in addition to adaptation and normalization,
recruitment provides a third mechanism of coding contrast.

Many mechanisms for cortical adaptation have been proposed, for example: fatigue of the thalamocortical synapse, cell intrinsic mechanisms, shunting inhibition, hyperpolarization, or other network mechanisms. Although adaptation has been studied for the past several decades, none of the proposed mechanisms account for the data fully. Many mechanisms of adaptation proposed involve inhibition. Little attention, however, has been paid to the roles of individual subtypes of inhibitory cells during sustained stimulation.

To examine neuron-type specific adaptation, we combined two-photon microscopy and immunohistochemistry. We found two subpopulations of inhibitory cells paradoxically increase their activity during adaptation. These neurons are thus likely to be the long sought-after inhibitory components that provide adaptation.

© 2017 Institut für Neuroinformatik