A database of upper limb surface electromyogram signals from demographically diverse individuals

Area of Science:

• Biomedical Engineering
• Human-Computer Interaction
• Signal Processing

Background:

• Surface electromyogram (EMG) signals are promising for noninvasive upper limb neuromuscular interfaces.
• Understanding physiological and anatomical influences on EMG is crucial for reliable gesture decoding.
• Existing research lacks comprehensive datasets addressing demographic variability in EMG signals.

Purpose of the Study:

• To present a novel dataset of upper limb EMG signals and physiological measures from a demographically diverse adult population.
• To facilitate research on how factors like age and body mass index affect EMG signal distribution.
• To establish a benchmark for developing fair and unbiased gesture decoding algorithms.

Main Methods:

• Collected EMG signals and physiological data (e.g., skin hydration, elasticity) from 91 diverse adults.
• Included participants across a wide age range (18-92 years) and BMI categories (healthy, overweight, obese).
• Validated data quality using contemporary gesture decoding methodologies.

Main Results:

• A comprehensive dataset of upper limb EMG and associated physiological measures was successfully compiled.
• The dataset captures significant demographic diversity, including age and BMI variations.
• Initial validation confirms the data's suitability for studying demographic confounds in EMG.

Conclusions:

• The presented dataset is a valuable resource for investigating demographic influences on EMG-based interfaces.
• It serves as a critical benchmark for advancing fair and unbiased hand gesture decoding algorithms.
• This work paves the way for more robust and inclusive human-computer interaction technologies.