The 'Recognizing Cell' Model for Simulation of Feed- Forward Interaction in the Cognitive System: Numerical Experiment and Psychophysical Data

摘要

The study represents a research into the influence of the feed-forward pre-tuning on the dynamical modes of models of the homogeneous neural-inspired systems when identifying fundamental features in informational signal. On the basis of calculations and received data we put forward a hypothesis about the mechanisms of influence of the sensory signal context on the absolute thresholds of the perception of primitive features. We introduce psychophysical measures which verify the introduced model and we have experimentally determined typical characteristic periods in feed-forward connection dynamics which allow to tune the mathematical model for reproduction of individual modes of singling out a feature.