Non-monotonic relationship between correlation and strength of inhibitory feedback

摘要

Correlative activities of neurons are ascribed to inhibitory connections in previous literatures. It is believed that correlative activities of neurons increase with enhancing the strength of the inhibitory connection. However, we find a nonmonotonic relationship between the correlation and the strength of inhibitory feedback. The neural network in the present paper consists of excitatory and inhibitory neurons with global inhibitory feedback. By numerical simulations, we caimiate the relationship among correlation coefficient, firing rate and inhibitorv feedbook.we find that when the stimuli are weak, the correlation increases monotonously with enhancement of the inhibitory feedback. While the stimuli are strong, the correlation first drops with weak inhibition, then rises after the inhibitory feedback gain exceeds a threshold.