On-line real-time digital computer separation of extracellular neuroelectric signals.

Abstract

An on-line real-time spike recognition and separation system for up to six different spikes in multiple unit extracellular records has been developed using the Micro-LINC 300 computer. Assuming biphasic potentials, the computer extracts four unambiguous variables, two in amplitude and two in time. One variable of a selected pair is plotted against the other on the computer CRT. Points representing the same class of signal will fall into clusters. Cluster boundaries for spike separation are set visually by the investigator. Output is in the form of pulses on different relay lines. The system permits waveforms to overlap up to 15 percent. Examples of system performance with real signals are presented. A comparison is made with other spike recognition systems. A simple system is presented for generating artificial test signals and a technique described for testing any spike recognition system.