TY - JOUR AU - Wang, Ruoxi AU - An, Qi AU - Yang, Ningjia AU - Kogami, Hiroki AU - Yoshida, Kazunori AU - Yamakawa, Hiroshi AU - Hamada, Hiroyuki AU - Shimoda, Shingo AU - Yamasaki, Hiroshi R. AU - Yokoyama, Moeka AU - Alnajjar, Fady AU - Hattori, Noriaki AU - Takahashi, Kouji AU - Fujii, Takanori AU - Otomune, Hironori AU - Miyai, Ichiro AU - Yamashita, Atsushi AU - Asama, Hajime PY - 2022 M3 - Original Research TI - Clarify Sit-to-Stand Muscle Synergy and Tension Changes in Subacute Stroke Rehabilitation by Musculoskeletal Modeling JO - Frontiers in Systems Neuroscience UR - https://www.frontiersin.org/articles/10.3389/fnsys.2022.785143 VL - 16 SN - 1662-5137 N2 - Post-stroke patients exhibit distinct muscle activation electromyography (EMG) features in sit-to-stand (STS) due to motor deficiency. Muscle activation amplitude, related to muscle tension and muscle synergy activation levels, is one of the defining EMG features that reflects post-stroke motor functioning and motor impairment. Although some qualitative findings are available, it is not clear if and how muscle activation amplitude-related biomechanical attributes may quantitatively reflect during subacute stroke rehabilitation. To better enable a longitudinal investigation into a patient's muscle activation changes during rehabilitation or an inter-subject comparison, EMG normalization is usually applied. However, current normalization methods using maximum voluntary contraction (MVC) or within-task peak/mean EMG may not be feasible when MVC cannot be obtained from stroke survivors due to motor paralysis and the subject of comparison is EMG amplitude. Here, focusing on the paretic side, we first propose a novel, joint torque-based normalization method that incorporates musculoskeletal modeling, forward dynamics simulation, and mathematical optimization. Next, upon method validation, we apply it to quantify changes in muscle tension and muscle synergy activation levels in STS motor control units for patients in subacute stroke rehabilitation. The novel method was validated against MVC-normalized EMG data from eight healthy participants, and it retained muscle activation amplitude differences for inter- and intra-subject comparisons. The proposed joint torque-based method was also compared with the common static optimization based on squared muscle activation and showed higher simulation accuracy overall. Serial STS measurements were conducted with four post-stroke patients during their subacute rehabilitation stay (137 ± 22 days) in the hospital. Quantitative results of patients suggest that maximum muscle tension and activation level of muscle synergy temporal patterns may reflect the effectiveness of subacute stroke rehabilitation. A quality comparison between muscle synergies computed with the conventional within-task peak/mean EMG normalization and our proposed method showed that the conventional was prone to activation amplitude overestimation and underestimation. The contributed method and findings help recapitulate and understand the post-stroke motor recovery process, which may facilitate developing more effective rehabilitation strategies for future stroke survivors. ER -