Surface EMG crosstalk between knee extensor muscles: experimental and model results

Surface electromyographic (EMG) crosstalk between vastus lateralis, vastus medialis, and rectus femoris muscles was evaluated by selective electrical stimulation of one muscle and recording from the stimulated and another muscle with linear surface arrays of eight electrodes. The ratio between the amplitude of the signals recorded over nonstimulated and stimulated muscles and their correlation coefficient were used as indices to quantify crosstalk. Single-differential and double-differential detection systems were used with interelectrode distances in the range 10-40 mm. The multichannel EMG signals clearly showed that crosstalk is largely due to nonpropagating potentials that correspond in time to the end of the propagation of the action potentials generated by the stimulated muscle. The crosstalk signal increased with increasing interelectrode distance and was statistically higher for single- than for double-differential recordings. The correlation-based indices of crosstalk were poorly correlated with the amplitude-based indices. Moreover, the characteristic spectral frequencies of the signals detected over the nonstimulated muscles were statistically higher than those from the stimulated muscles. A mathematical model of signal generation was used to explain the experimental findings. This study clarifies many controversial findings of past investigations and creates the basis for crosstalk interpretation, simulation, and reduction.