A new digital signal processing technique for the estimation of motor unit action potential templates

Accurate estimation of actual motor unit action potential template parameters from noisy measurements is quite important for quantitative electromyograhic (EMG) analysis. This is especially true in case of clinical applications. In This paper presents a new method for the accurate estimation of motor unit action potential (MUAP) templates, contaminated by instrumentation and biological noise, and by the interference of MUAPs of neighboring motor units. This method is based on three processes: Interfering MUAPs are detected by a sensitive detection algorithm. The sample points of these MAUPs are rejected and excluded from the estimation process. An α-trimmed mean is then calculated for each sample point of all MUAPs within a motor unit action potential train (MUAPT). To further smooth the baseline of the template while preserving its shape, a simple wavelet based de-noising technique is applied to the resulting template. The proposed method is compared to traditional estimation techniques such as mean, median, mode, σ-trimmed mean and Bayesian estimation. Using the isolated MUAPs within the 46 trains of concentric needle-detected MUAPs (25–447 per train) obtained using EMG signal decomposition, template baseline RMS levels and their spike shape fidelities were measured. Overall, this newly developed method compared to traditional estimation techniques provided the smoothest baselines and an acceptable spike fidelity factor.