Influence of temperature and noise on the propagation of subthreshold signal in feedforward neural network

Abstract

Temperature is an important environmental factor that all creatures depend on. Suitable temperature is essential for maintaining the normal physiological functions of nervous system. Subthreshold signal is generated by weak activities in the nervous system that is difficult to be detected. Based on an improved Hodgkin-Huxley (HH) neuron model considering temperature and noise, the ten-layers pure excitatory feedforward neural network and the ten-layers excitatory-inhibitory (EI) neural network are constructed to study the propagation of subthreshold excitatory postsynaptic current signal. It's found that increasing temperature can restrain the signal propagation, and raise the noises intensity threshold where the failed signal propagation can transform into succeed signal propagation. Under the large noise, the signal propagation in network of different temperatures exhibits different anti-noise capabilities. Moreover, temperature and noise can modulate the spontaneous activity of the EI neural network to transform between the synchronous regular (SR) state and the asynchronous irregular (AI) state. The EI network's spontaneous activity will completely cover subthreshold signal, and block the signal propagation under large noise. The jumping phenomenon in the value of fidelity, which measures the quality of signal propagation, appears in both pure excitatory network and EI network. This paper provides potential value for understanding the regulation of both temperature and noise in information propagation in neural network.