Spatial Filtering by a Two-Dimensional Interconnected Network with Spike Timing Dependent Synaptic Plasticity Depending on Its Temporal Properties

Abstract

In the present study, we investigated the dependence of characteristics of spatial filtering by the two-dimensional interconnected network with spike timing dependent plasticity (STDP) on its temporal properties. STDP has different temporal properties according to the area of a brain. The temporal properties are indicated by the form of learning window. In the previous study, we studied the response of an one-dimensional interconnected network with STDP in spatial processing using computer simulation and found that the one-dimensional network well responded to a particular spatial frequency component of the input. Here, we showed that the two-dimensional interconnected network with STDP provided two types of spatial filtering. One is spatial low-pass filtering using the learning rule of electric fish type. The other is spatial high-pass filtering using the learning rule of hippocampus type.