Characterization of EMG time-frequency content during Parkinson walking: a pilot study

Abstract

Surface electromyography (sEMG) is commonly adopted to characterize walking in patients affected by Parkinson’s disease (PD). Timing and morphology of sEMG signal are typically investigated, while poor information on frequency content is available. Thus, the present pilot study was designed to test the hypothesis that continuous wavelet transform (CWT) of sEMG signal is a suitable approach to assess muscle activity during PD-walking task, in both time and frequency domains. To this aim, sEMG signals from 4 leg muscles of 5 patients are acquired during walking and processed to assess CWT-scalogram function. Results show that CWT is able to provide time ranges of muscle-activation over the whole PD population, which matches with what reported in previous studies on PD. The novel contribution of this study consists in achieving a characterization of the frequency content of each one of regions detected in time domain. Although the frequency content does not exceed the typical frequency range between 5 Hz and 450 Hz, different mean frequency contents are observed among muscles and among different activations of the same muscle. In particular, a relevant variability of frequency content is observed for thigh muscles, showing differences up to 180 Hz between stance and swing values. In conclusion, present findings support the use of CWT scalogram for a reliable assessment of muscle activity in time-frequency domain, during walking of PD patients. Outcomes highlight a large inter and intra muscle variability of frequency range, opening a new field of investigation for future studies.