Single Spiking Neuron Learns Dynamic Stimuli

In this paper we show that simple spiking neural network is able to learn dynamic stimuli by memorizing specific patterns and ignoring others without implementing any recurrent connections. The proposed model consists of one spiking neuron with trainable synaptic connections from neuron-like asynchronous signals. The model was trained in biologically plausible fashion, utilizing exact timings when spikes occur. We expect that future research on investigation of the ability of network stabilization while learning different patterns will lead to processing of complex patterns.