Electromyographic Study of Motor Learning for a Voice Production Task

Abstract

This study's broad objective was to examine the effectiveness of surface electromyographic (EMG) biofeedback for motor learning in the voice production domain. The specific objective was to examine whether concurrent or terminal biofeedback would facilitate learning for a relaxed laryngeal musculature task during spoken reading. Twenty-two healthy adult speakers were randomly assigned to 1 of 2 groups. One group received real-time EMG waveform displays of muscle activation from bilateral thyrohyoid sites during reading trials (concurrent feedback group). The other group received static terminal EMG waveform displays about activation levels for the same sites on completion of successive trials (terminal feedback group). All participants were instructed to minimize EMG amplitudes from the thyrohyoid sites during phonation in an oral reading task. Signals were also collected from control, orofacial sites, but participants received neither instructions nor feedback for those sites. The pooled data (2 feedback groups x 2 electrode sites) showed that, overall, muscle activation levels did decrease across baseline, training, and no-feedback test phases. However, no clear evidence was seen of reliable changes in the targeted laryngeal muscle activation levels across the phases, for either the concurrent or the terminal feedback groups. Paradoxically, and entirely unanticipated, reliable decreases were seen in muscle activation for the orofacial, no-feedback control sites. Those decreases were equivalent across concurrent and terminal feedback groups. The unanticipated findings indicate that the provision of biofeedback for a target muscle group facilitated incidental learning in another, untargeted muscle group. Discussion focuses on the possible role of locus of attention in motor learning. Building on literature from other domains, the hypothesis is advanced that attention to muscular contractile force during training trials may suppress intentional learning for attended target sites but may benefit incidental learning for nearby, unattended sites.