Modeling nonlinear errors in surface electromyography due to baseline noise: a new methodology.

摘要

In a recent issue Law et al. (2011) found that it is customary for scientists to subtract the baseline noise in electromyographic (EMG) recordings under the assumption that noise adds linearly to the EMG signal. They showed that a linear subtraction of the noise from the measured EMG results in an underestimation of the true EMG amplitude; the error increased with decreasing signal-to-noise ratio (SNR), resulting in a substantial underestimation of the EMG signal at low SNRs. The authors also present a non-linear model comprising three parameters empirically derived from their dataset, which is superior to the linear subtraction of the noise.