Modeling the top-down influences on the lateral interactions in the visual cortex

摘要: Attention modulates the amount of excitatory and inhibitory lateral interactions in visual cortex. A recurrent neural network is proposed to account for modulatory influence top-down signals. In model, two types inhibitions are distinguished: dendritic inhibitions. Dendritic inhibition regulates impact that surrounding cells may exert on a target cell via dendrites neurons subpopulation mediating inhibition. increases prevents contextual interactions, while it has no effect when there surround input. Computer simulations showed model able exhibit properties attentional gating. condition focused attention, activity presence stimuli restored level as stimulus presented alone. Moreover, show contrast gain response sensitivity function depending strength