P28: Analysis of Surface Electromyography (sEMG) Signals for Motion Signatures in Strength Training Exercises via Stretchable Wearable Electronic Sensors

Abstract

Background: The characterization of commonly used strength training exercises has a range of implications for analysis of techniques displaying proper form and injury prevention in both sports and clinical rehabilitation settings. This study examines the ability of wearable sensors to effectively determine the degree of activation of different muscle groups using surface electromyography (sEMG) signals. The interpretation of these signals across muscle groups within an exercise allow for the characterization of their motion signatures. Wearable sensors with embedded electrodes (BioStamps RC, MC10 Inc.) were applied to two defined muscle groups for three exercises: a triceps push-up (PU), a biceps curl (BC) and an unweighted squat (US). Methods: Seven healthy volunteers performed three consecutive exercises in the following order; PU, BC, and US. Each exercise was performed for five repetitions with BioStamps applied to the distal and proximal ends of the subjects’ dominant side using adhesive bandages. For each of the defined exercises the investigator played a timed voice recording in which the subjects were instructed to perform the series of repetitions with consistent timing. Sensors automatically upload de-identified data to a tablet device during use. The following parameters are reported from the stamp: electromyographic data (of the muscle causing abduction and adduction) and acceleration (of the distal and proximal ends of the limb which is being abducted/adducted). The data was uploaded to Python and analyzed using a rolling average. P-values were then determined using a one-tailed t-test. Results: Each of the three strength exercises had two identifiable motion signatures, one from each muscle group, derived from the sEMG data which shows the peak muscle activation during the 5 repetitions of an exercise (Figure 1, E). Placement on the BioStamps on the TB and BB recorded activation in the PU and BC (Figure 1, A & B). There is a difference between the TB and BB activation in the PU (p=0.358) but no difference was found between their activation in the BC (p=0.06). This finding was limited by the small sample size (n=7), which would be improved in a future study. Placement of the BioStamps on the BF and RF was used in the unweighted squat (Figure 1, C & D). There is a difference in the RF and BF activation in the US (p=0.179). Discussion: By nature of the antagonistic motion of the triceps brachii (TB) and biceps brachii (BB) in a push-up, activation levels measured by sEMG recordings are expected to be similar. The TB acts as one of the primary movers, while the BB acts as an antagonist and stabilizes the movement to allow for a controlled push-up. Similarly, in a squat similar levels of activation were expected in the rectus femoris (RF) and the biceps femoris (BF) as the RF functions as a primary mover and the BF as a stabilizer. Activation of both the primary mover and stabilizing muscles is important for proper form and function.Figure 1. A & B. Placement of BioStamps on the triceps brachii and biceps brachii, muscles used in the biceps curl (p=0.06) and triceps push-up (p=0.358). C & D. Placement of BioStamps on the rectus femoris and the biceps femoris, muscles used in the unweighted squat (p=0.179). E. Unweighted squat sEMG graph showing muscle activation (mV) of rectus femoris over time (s).