Techniques for Processing EMG Signals
Diane K. Port · The Journal of the Acoustical Society of America · 1973
Analysis of electromygraphic activity is potentially a fruitful technique for investigating speech production. However, the EMG signal picked up by bipolar wire electrodes is weak and inherently variable and must be carefully handled. At Haskins Laboratories a computer processing system was developed to analyze simultaneously up to eight EMG channels of data. This system allows for identifying, sampling, and digitally averaging EMG signals for a set of utterances of the same type. Integration of the EMG signals, both in the hardware and software, has been recently revised and its importance to the interpretation of EMG signals will be discussed. The relationship between an averaged EMG signal and the individual utterances in the sample will be discussed in the context of editing and display capabilities. Further, an experiment involving multiple placement of electrodes in both muscles of a bilateral pair has been conducted. Computation of correlation coefficients for utterances across the multiple electrode placements shows that one electrode can provide EMG activity representative of activity in the entire muscle. This is true for both individual utterances and for the average of a set of utterances. [This research was partially supported by the National Institute of Dental Research.]