Simulating the Human Eye Movement System based on the Volterra Models in neurophysiological studies

A method for modeling the human eye movement system (EMS) based on Volterra models in the form of multidimensional transient characteristics has been further developed. These characteristics were obtained from experimental "input-output" EMS studies using innovative eye-tracking technology. The respondent was studied in the morning and evening, which made it possible to identify the variability of EMS characteristics in the before-work (normal) and after-work (fatigue) states. It has been established that for EMS identification in practical neuroscience applications, it is advisable to use the quadratic Volterra polynomial, determined using the least squares method with three step test signals of different amplitudes. Datasets for machine learning were formed in different diagnostic feature spaces: heuristic features and features derived from sampling the first- and second-order transient characteristics. Using computer modeling, feature sets that are robust according to the probability of correct recognition criterion were identified. To assess psychophysiological states, a Bayesian classifier and the SVM method were applied. Figs.: 11. Tabl. 3. Refs.: 31 titles.