Capacity bounds of neuro-spike communication by exploiting temporal modulations

We consider a neuro-spike communication system between two nano-machines, with information conveyed in the time intervals of the input spike train. The main contribution of our paper is modeling of the neuro-spike communication channel by an additive Gamma noise channel model. In this channel, the information is corrupted by Gamma distributed noise. We show that the proposed channel model is efficient for the neuro-spike communication when it exploits temporal modulations to transfer information. Then, we consider the Gamma distributed noise and we derive the upper and lower bounds on the channel capacity. Unlike Additive White Gaussian Noise (AWGN) channels, there is no single quality measure like signal-to-noise ratio for this channel model. Thus, we analyze the channel capacity bounds versus different values of time intervals and the decision threshold of the receiver.