Accurate Estimation of Motor-unit Action Potential Waveforms and Discharge Patterns

McGill KC, Lateva ZC. Accurate estimation of motor-unit action potential waveforms and discharge patterns. 11th Intl Congr Electromyogr Clin Neurophysiol, Prague, Czech Republic, 1999.

We have developed an interactive computer program for extracting complete motor-unit (MU) discharge patterns and high-fidelity motor-unit-action-potential (MUAP) waveforms for multiple MUs from moderate EMG interference patterns. The program uses an automatic template-matching algorithm to identify as many MUAPs as possible. Then the user can review the results to correct mistakes and manually identify any remaining MUAPs. The program employs a new algorithm that efficiently resolves superimpositions involving an unlimited number of templates, and an interference-cancellation method for averaging MUAP waveforms accurately. We have used the program to analyze hundreds of EMG signals recorded using concentric, monopolar, and fine-wire electrodes in various muscles in contractions up to 40% MVC. The program is proving to be a valuable tool for a wide range of experimental investigations of MU firing behavior and MUAP morphology. We have been able to fully decompose signals containing as many as 16 simultaneously active MUAPs, allowing study of discharge-rate modulation and discharge interactions (common drive, synchronization). Individual spike times of occurrence are measured with high resolution, allowing measurement of jitter between MUAP components. It is possible to ascertain that all active MUs with fibers within a certain vicinity of the electrode (i.e., whose spike amplitudes exceed a certain threshold) have been identified, making possible objective studies of MU recruitment. The MUAP averages typically have residual noise levels less than 2 uV, allowing investigation of small morphological features including onset, terminal wave, and slow afterwave. The averaging window can be as long as 60 ms, allowing investigation of late MUAP components.