Data for Comparison of Audio, Visual, and AudioVisual Biofeedback for Electromyographic Control

Sampling rate is 40 Hz from RMS EMG features calculated online using 50-sample (50 ms) windows with 25 ms overlap. For each participant 8 recordings are provided:</p><ol><li>Contraction and relaxation cycles during visual biofeedback for muscle relaxation, ("P#<i>_</i>R_VB.csv"),</li><li>Contraction and relaxation cycles during audio biofeedback for muscle relaxation, ("P#<i>_</i>R_AB.csv"),</li><li>Contraction and relaxation cycles during audio+visual biofeedback for muscle relaxation, ("P#<i>_</i>R_AVB.csv"),</li><li>Contraction and relaxation cycles during visual biofeedback for muscle contraction, ("P#<i>_</i>C_VB.csv"),</li><li>Contraction and relaxation cycles during audio biofeedback for muscle contraction, ("P#<i>_</i>C_AB.csv"),</li><li>Contraction and relaxation cycles during audio+visual biofeedback for muscle contraction, ("P#<i>_</i>C_AVB.csv"),</li><li>Pre-trial threshold recording for muscle relaxation, ("P#_R_T.csv"),</li><li>Pre-trial threshold recording for muscle contraction, ("P#_C<i>_</i>T.csv").</li></ol><p dir="ltr">EMG recording was carried out using a Myoware Version 2 differential amplifier front end (Advancer Technologies) with 200× gain and 494.8 Hz low-pass filter.

Additional signal conditioning was achieved with a 92 Hz second-order high-pass Sallen-Key filter, 10× gain second-stage amplifier, and mid-supply reference potential for ADC input range, all using MCP606 operational amplifiers (Microchip). Amplified and filtered EMG was sampled at 1 kHz using an Arduino Uno R3 microcontroller with a 10-bit ADC. Digitised raw EMG recordings were converted to RMS as described above.

All data is supplied in raw ADC values in .csv column vectors, to convert to mV multiply by the conversion factor of 2.2195.</p><p dir="ltr">This participant study was approved by UCL Research Ethics Committee: 12363/005.</p>