Variation of muscle coactivation patterns in chronic stroke during robot-assisted elbow training

Objective: To investigate the variation of muscle coactivation patterns during the course of robot-assisted rehabilitation on elbow flexion and extension for chronic stroke. Design: A detailed electromyographic analysis was conducted on muscle activation levels and muscle coactivation patterns, represented by a cocontraction index of a muscle pair, for the muscles of biceps brachii, triceps brachii, anterior deltoid, and posterior deltoid, during training of elbow extension and flexion, actively assisted by a robot, from 0 degrees to 90 degrees by tracking a target moving at a speed of 10 degrees/s on the screen. Setting: Rehabilitation center research laboratory. Participants: Seven hemiplegic chronic stroke patients received elbow training. Interventions: Each subject received 20 sessions (1.5 hours/session) of the elbow training on his/her paretic side at an intensity of 3 to 5 times a week for a training period of 7 consecutive weeks. Main Outcome Measures: Muscle cocontraction index, muscle activation level, and Modified Ashworth Scale (MAS), Fugl-Meyer Assessment (FMA), and Motor Status Scale (MSS) scores. Results: The electromyographic activation levels of the biceps brachii, triceps brachii, and anterior deltoid of each subject decreased during the training. The overall electromyographic activation levels of the biceps and triceps, which, summarizing the performance of all subjects, decreased significantly in the middle sessions (from the 8th to 12th sessions) of the training (P <.05), associated with the significant decrease (P <.05) in the MAS score. The overall electromyographic activation level of the anterior deltoid also decreased significantly from the 8th to 20th sessions (P <.05). Significant decreases in the cocontractions of all muscle pairs were observed in all subjects and also in the overall cocontraction index (P <.05). The cocontraction between the biceps and triceps significantly decreased when the overall electromyographic levels of the 2 muscles were stable from the 10th to 20th sessions (P <.05). Significant improvements (P <.05) on the FMA and MSS score were also found by the pre- and postassessments. Conclusions: In the 20-session robot-assisted training, the excessive muscle activations reduced mainly in the first half of the training course, which could be related to the learning process of the tracking skill and also to the reduction in muscle spasticity. The muscle coordination for achieving elbow tracking improved significantly in the latter sessions of the training, represented as decreased cocontraction indexes between the muscle pairs.