Heartbeat detection from high-density EMG electrodes on the upper arm at different EMG intensity levels using Zephlet

Area of Science:

• Biomedical Engineering
• Cardiovascular Technology
• Signal Processing

Background:

• Cardiac arrhythmias cause preventable deaths globally.
• Wearable devices offer continuous electrocardiogram (ECG) monitoring for early detection.
• Current research focuses on body surface ECG, but upper arm monitoring is less explored.

Purpose of the Study:

• To develop and validate a heartbeat detection method using ECG signals from the upper arm.
• To identify optimal electrode placements on the upper arm for various conditions.
• To assess the performance of the developed method under different muscle contraction levels.

Main Methods:

• Optimal electrode locations on the upper arm were determined for rest and elbow flexion (EF) conditions.
• A synthesized ECG signal was generated using selected electrodes and generalized weights.
• Feature extraction was performed using zero-phase wavelet (Zephlet), followed by heartbeat detection with a multiagent detection scheme (MDS).

Main Results:

• The F1-score for heartbeat detection reached 0.94 ± 0.16 at rest and decreased with increasing muscle contraction levels.
• Performance varied with muscle contraction, with F1-scores ranging from 0.94 (Rest) to 0.67 ± 0.31 (C3).
• Reducing the detection tolerance from 50 ms to 20 ms impacted the F1-score, yielding 0.81 ± 0.20 for Rest and 0.44 ± 0.26 for C3.

Conclusions:

• Precise and consistent electrode localization on the upper arm was achieved for varying muscle contraction levels.
• An accurate heartbeat detection method was developed and validated for upper arm ECG signals.
• The findings contribute to advancing wearable technology for cardiac monitoring, particularly in non-ideal conditions.