Principal Component Regression and Partial Least Squares Regression Evaluation of Electrode Performance for Non-Invasive Multimodal Measurement of Arm Muscle Activity

Surface electromyography (sEMG) is the de-facto biopotential medical instrumentation solution for non-invasive detection of limb motion EMG activation signals, which in turn provides important information regarding intentional and reactive muscle activity. However, sEMG delivers information of electrical activity mainly due to superficial muscles over a wide sensing field. Additional physiological processes may be elucidated by combining sEMG data with other measurements, such as Electrical Impedance Tomography (EIT) imaging, to assess muscular activity due to increased vascularization. In any case, the choice of the electrode material and shape influences the quality of measured data; "wet" electrodes (Ag/AgCl) are typically used for non-invasive sEMG measurements, whereas "dry" electrodes (i.e. stainless steel, copper-nickel) are preferred for electrical impedance measurements. Here, the authors explore the use of different types of electrode materials to measure sEMG upper limb motion signals, in comparison with Ag/AgCl electrodes, to determine a suitable electrode array that can accommodate both sEMG and EIT measurements. sEMG data was obtained from 10 healthy volunteers and processed using Short Time Fourier Transform (STFT) and Principal Component Spectral Analysis (PCSA) for Principal Component Regression (PCR) and Partial Least Squares Regression (PLSR). The results suggest that steel electrodes perform closer to commonly used Ag/AgCl electrodes, offering the potential for a developing a multimodal sEMG/EIT wearable measurement system for upper limb motion evaluation.