Limits of coincidence detector neurons as decoders of Polychronous Neuronal Groups firing completely

摘要

This work analyses Polychronous Neuronal Groups (PNGs) and the possibilities of encoding and decoding information from these signals. We considered the use of single neuron coincidence detectors as decoders of PNGs and determined the total number of symbols these detectors could handle. Simulations using the Izhikevich's neuron model yielded several constraints of these systems, from which we obtained that such system has a maximum processing power of 146 bit/s, when considering symbols generated by groups of 5 neurons.