Estimation of Motor Nerve Conduction Velocity Distribution: A Continuous Approach

Estimation of motor nerve conduction velocity distribution (NCVD) from the compound muscle action potential (CMAP) is a challenging and long-studied problem in nerve conduction study. In the present study, we have explored a new approach to determine the motor NCVD from the corresponding CMAP in a non-invasive manner using a continuous approach. In our study, we have taken the diphasic sinusoidal function and also the Hermite polynomial function to simulate the motor unit action potential (MUAP). We have experimented the efficacy of different polynomial functions of different degrees and also the gaussian and double gaussian distributions etc. to model the motor NCVD. Then, using the forward approach of nerve conduction, we have synthetically created the CMAP function. The continuous function of motor NCVD in our modeling helps us exploit the gradient optimization technique to solve the inverse problem of nerve conduction, i.e., estimation of motor NCVD from CMAP minimizing the least-square error. Our estimated results conformed well to the synthetically created CMAP dataset. The proposed technique is non-invasive and offers a way to estimate the motor NCVD from the corresponding CMAP's in a continuous approach which would be a useful tool for detecting the peripheral neuropathies.