Meanfield chaos dynamics with changing chaos threshold value

Single neuron dynamics is derived in a coherent interacting mean field with all other neuron dynamics coupled through the Hebbian learning of synaptic weight updates. This exact one-body equation is derived because of the product nature of Hebbian learning rule. It provides the diagnoses needed for the collective interaction of artificial neural networks (ANN). In case of a chaotic neurodynamics generated with the N-shaped sigmoidal function, we can use the meanfield equation to demonstrate any single neuron behavior in the network, e.g. contrast reversal, limited cycles, chaos etc., due to a changing habituation threshold value. In case of negative threshold values, which might be viewed as an external reward to encourage neurons firing with higher rates, an image block quantization compression artifact has been quickly overcome with such a chaotic ANN and the result is quantified with a fractal dimensionality analysis of the image quality