Fig. 1

The overview of the experimental setup and study design. A, KAIST Upper Limb Synergy Investigation System (KULSIS). The isometric force and the electromyography (EMG) signal were acquired and processed by the main computer (PC1). The information for the display (force data for the assessment and EMG data for the training) was transmitted from PC1 to the display computer (PC2) via User Datagram Protocol (UDP). B, The EMG-guided training paradigm. The activation amplitudes of the targeted muscle pair, the brachioradialis (BRD) and biceps brachii (BI), were mapped to the vertical and horizontal displacement of a cursor, respectively. The cursor remained at the right bottom corner of the display when there was no activation of both targeted muscles. Participants were verbally guided to generate an appropriate force to move the cursor within the elongated squared region to reach the blue target zone and remain in the zone for 1 s consecutively to make a successful target match. The area of the square target zone was defined as 30% of the target activation level of each muscle (70% of the maximum voluntary contraction (MVC)). The diagonal line indicates the condition when both targeted muscles are activated together to an equal extent. C, The spatial distribution of 54 normalized force targets for the assessment of the proposed exercise effects. Fx, Fy, and Fz directions represent medial-lateral, backward-forward, and downward-upward directions, respectively. D, Timeline of the six-week training and assessment. The study consisted of 18 sessions of training (three sessions per week) and four assessments (at Weeks 0, 2, 4, and 6). Each assessment included both “Habitual” and “As-Trained” conditions. For each trial of both training and assessment, 3 s of an inter-trial interval, 2 s of a baseline period, and up to 7 s of a target matching window were provided for task completion