Cross talk in surface electromyograms of human hamstring muscles.

Signals generated from muscles other than the muscle(s) of interest (cross talk) can confound the interpretation of surface electromyograms (EMGs). In this study, the amount of cross talk in surface EMGs of human hamstring muscles was estimated using a protocol in which the quadriceps femoris was electrically stimulated via the femoral nerve. EMGs were recorded from the vastus lateralis and the medial and lateral hamstring muscle groups. The amplitude of the EMG response of the vastus lateralis to electrical stimulation was adjusted to match that of its maximum voluntary effort (MVE) under isometric conditions. Subsequent power density spectrum analysis showed that the median frequencies of the signals generated by electrical stimulation and MVE were not significantly different. In conventional bipolar recordings, cross talk in lateral hamstring EMGs averaged 17.1% MVE and in medial hamstring EMGs 11.3% MVE (average-rectified values). The double differential technique significantly reduced cross talk to 7.6% MVE for the lateral hamstrings, and to 4.2% MVE for the medial hamstrings. The double differential technique appears to be more selective than the bipolar technique when recording EMGs from muscles with highly active neighbors and thus should be used in such situations. Software simulations of the double differential technique also appear to be more selective than the bipolar technique and may be used when the number of amplifiers available is limited.