Habituation in the KIII olfactory model using gas sensor arrays

Inspired by the habituation process in the olfactory system, this article presents an approach for analyzing electronic-nose data using Freeman’s KIII neurodynamics model. In order to ensure the additivity of patterns from odor mixtures, input data from a gas sensor array is first processed with a family of discriminant functions that yield an orthogonal binary representation. The process of habituation is then simulated through synaptic depression with a decay term that reduces the strength of mitral and granule connections when the KIII model is excited with a continuous stimulus. As a result, the system is able to mimic the effects of habituation when processing odor mixtures with gas sensor arrays.