EPSP simulator

[villemt]

Modern artificial neural networks are often derivants of the McCulloch-Pitts model and don't have much common with actual biological neural cells functionality. For example neural models send high-precision values from one cell to another, while actual biological cells send only individual impulses or impulse-groups. Another problematic concept is backpropagation, which is required for artificial networks learning. In biological cells backpropagation can occur within one cell, but its accuracy is not enough for backpropagation-defined parameter tweaking. So as part of my univ. studies, I wrote a model/simulator for actual biological cells EPSP (excitatory post-synaptic potential) activations, which learns, detects and repeats variable time-step sequences, often referred as coincidences (demos-menu).