紀要論文

スパイキングニューロン回路の多彩な分岐現象の解析

jpn

21879923

https://doi.org/10.15002/00010409

法政大学大学院理工学・工学研究科

55

1

6

2014-03-24

Version of Record

This paper studies filter-induced bifurcation phenomena in simple spiking neuron models. Repeating integrate-and-fire switching between a periodic base signal and a threshold, the neuron model outputs spike-trains. The dynamics depends crucially on the shape of the base signal and we make the base signal by applying two kinds of filters to a square source signal. First, we apply the basic RC low-pass filter to the source signal. As a key parameter (time constant) varies, the shape of the base signal varies and the model can exhibit bifurcation phenomena of various periodic/chaotic spike-trains. Presenting a simple test circuit, typical phenomena are confirmed experimentally. Second, we apply the ideal low-pass filter to the source signal. The ideal filter cannot be realized by analog circuits and is an object in artificial numerical experiments. This filter causes ripple on the base signal (the Gibbs’ phenomenon) and the model can have various co-existence spike-trains.

Article

AA12677220