Motor-Unit Continuity in Wearable Electromyography During Resistance Exercise: Detection, Identity, and Mechanical Timing

Abstract

Movement changes the electrical signatures that high-density surface electromyography uses to identify motor units. An interruption in a discharge train therefore requires separate assessment of recording quality, event detection, and unit identity before it can support a physiological interpretation. In controlled ankle contractions, applying separation filters across movement velocities identified more than 92% of the units detected at the highest velocity in both lower-velocity recordings. In biceps simulations with known firing times, Kalman-based separation achieved 95.6% mean sensitivity and 96.5% mean precision during 2-s flexion–extension cycles. These findings demonstrate that detected unit number depends on the analysis as well as on neural activity. An interval record specifies channel usability, accepted discharges, and identity decisions; its interpretation distinguishes verified missed events, cessation conditional on detection sensitivity, ambiguous identity, and unavailable observation. Probability identities define the effects of selective inclusion on unit counts and discharge-rate means. A participant-weighted comparison retains the same eligible units across mechanically comparable intervals. Accurate exercise interpretation requires verified timing, unit correspondence, and explicit sampling denominators; acceptance of a later train cannot establish activity during an intervening unobserved interval.