Differentiating between the frequency characteristics of different healthy muscles via cluster analysis

The authors attempt to address the question of whether there exists a subtle characteristic difference between different healthy EMG muscle patterns. To answer this question, EMG data were collected from four muscles, namely, the deltoid, the biceps, the interosseus dorsalis, and the frontalis muscles. EMG measurements from surface electrodes were recorded from seven subjects at different percentage of the MVC of each muscle of each subject. Analyzing the different EMG records using the power spectrum revealed visible differences in the spectra characteristic between three of the four muscles investigated. Cluster analysis was then utilized to group the EMG data records into four categories. The records that fall in one of those categories are supposed to have similar characteristics from the power spectrum point of view. Results obtained indicate that cluster analysis based on the periodogram feature is capable of differentiating between the characteristics of the biceps, deltoid and the interosseus dorsalis muscles