Algorithms of Classification of Neuron Impulses Based on Dynamic Equations

摘要

A description of computer implementation of an algorithm for classification of neuron impulses (ACNI) using nonlinear dynamic equations is given. ACNI includes an automated learning procedure and allows a real-time classification of impulses of several neurons, recorded by a single microelectrode. Results of algorithm testing on a simulated signal with different noise levels are presented. We also demonstrate applications of ACNI for studying of neuron activity in the subthalamic kernel of brain, which is the goal zone for high-frequency electric stimulation of depth zones of brain in Parkinson's disease.