Temporal pattern learning in a spiking neuron chain

A simple neural network demonstrates the ability to learn Morse code-like temporal patterns, presented via a microphone. A neuronal model, which provides facility for 'spiking' neurons, is used as the basis for this network. The network consists mainly of a chain of neurons, connected so as to fire in sequence. After a period or unsupervised learning, the chain is capable of repeating a temporal pattern upon which it has been trained. The system is particularly robust with respect to noise ill the input pattern. Suggestions are made as to how this system could be extended, by replacing the simple chain structure with complex trees and interconnected 'blobs' of neurons. The example of birdsong as a biological system of temporal pattern learning is used as inspiration for further work in this area.