Molecular communication is a new paradigm to solve the problems of conventional communication system, such as capacity
or energy consumption. The neural network is one of the molecular communication mechanism valid for higher animals and
considered as a highly advanced information transfer network in terms of capacity, reliability and energy consumption. In this
paper, we provide the capacity of neuro-spike communication system, which is inspired from neural network. The neuro-spike
communication consists of axon propagation, vesicle release and neurotransmitter diffusion. Through modeling and analysis of
the above three parts, the capacity of the neuro-spike communication can be obtained. Numerical results show the trends of the
capacity for main parameters, such as bandwidth, absolute refractory period and distance. The capacity is converged to certain
value even though bandwidth of input signal becomes larger and larger. As a result, it can be seen that refractory period is a key
parameter in neuro-spike communication system