Duration Estimation of Motor Unit Potentials In Multiscanning-EMG Recordings

In this paper we present an advanced application of a well-established algorithm to detect the duration of the motor unit potentials(MUPs) of a multiscanning-EMG. The established algorithm was initially formulated to calculate the duration of a single MUP recorded from a single point. About 30 MUPs were acquired from 5 healthy subjects while doing a small contraction in the tibialis anterior muscle. A manual “ground truth” of the duration positions (start and end markers) was obtained for each MUP by clinical experts. The algorithm was compared to a widely recognized MUP duration method and with the ground truth. Results from the new method are statistically closer shows much closeness to the ground-truth values than the second method, both in the start and the end marker determination. The mean error tent to be 1.2ms and 4.2ms for the start and the end markers in the correlation method compared to a 3.2ms and 8.2ms in the second method. The p_value of the ttest conducted to be close to 0 also support that the two methods are significantly different and our method has comparatively less error to that of the second algorithm. In summary, our enhanced algorithm not only outperforms traditional methods but also aligns closely with expert assessments that is a promising factor for the assessment of duration in clinical trials.