Qiu Shuang, Feng Jing, Xu Jiapeng, Xu Rui, Zhao Xin, Zhou Peng, Qi Hongzhi, Zhang Lixin, Ming Dong
IEEE Trans Neural Syst Rehabil Eng. 2017 Jan;25(1):59-67. doi: 10.1109/TNSRE.2016.2556687.
Neuromuscular electrical stimulation (NMES) that stimulates skeletal muscles to induce contractions has been widely applied to restore functions of paralyzed muscles. However, the architectural changes of stimulated muscles induced by NMES are still not well understood. The present study applies sonomyography (SMG) to evaluate muscle architecture under NMES-induced and voluntary movements. The quadriceps muscles of seven healthy subjects were tested for eight cycles during an extension exercise of the knee joint with/without NMES, and SMG and the knee joint angle were recorded during the process of knee extension. A least squares support vector machine (LS-SVM) LS-SVM model was developed and trained using the data sets of six cycles collected under NMES, while the remaining data was used to test. Muscle thickness changes were extracted from ultrasound images and compared between NMES-induced and voluntary contractions, and LS-SVM was used to model a relationship between dynamical knee joint angles and SMG signals. Muscle thickness showed to be significantly correlated with joint angle in NMES-induced contractions, and a significant negative correlation was observed between Vastus intermedius (VI) thickness and rectus femoris (RF) thickness. In addition, there was a significant difference between voluntary and NMES-induced contractions . The LS-SVM model based on RF thickness and knee joint angle provided superior performance compared with the model based on VI thickness and knee joint angle or total thickness and knee joint angle. This suggests that a strong relation exists between the RF thickness and knee joint angle. These results provided direct evidence for the potential application of RF thickness in optimizing NMES system as well as measuring muscle state under NMES.
神经肌肉电刺激(NMES)通过刺激骨骼肌诱发收缩,已被广泛应用于恢复瘫痪肌肉的功能。然而,NMES诱发的受刺激肌肉的结构变化仍未得到充分理解。本研究应用超声肌肉成像(SMG)来评估NMES诱发运动和自主运动下的肌肉结构。在膝关节伸展运动过程中,对7名健康受试者的股四头肌在有/无NMES的情况下进行了8个周期的测试,并在膝关节伸展过程中记录了SMG和膝关节角度。使用在NMES下收集的6个周期的数据集开发并训练了最小二乘支持向量机(LS-SVM)模型,而其余数据用于测试。从超声图像中提取肌肉厚度变化,并比较NMES诱发收缩和自主收缩之间的差异,并用LS-SVM对动态膝关节角度与SMG信号之间的关系进行建模。在NMES诱发的收缩中,肌肉厚度与关节角度显著相关,股中间肌(VI)厚度与股直肌(RF)厚度之间观察到显著的负相关。此外,自主收缩和NMES诱发的收缩之间存在显著差异。与基于VI厚度和膝关节角度或总厚度和膝关节角度的模型相比,基于RF厚度和膝关节角度的LS-SVM模型具有更好的性能。这表明RF厚度与膝关节角度之间存在密切关系。这些结果为RF厚度在优化NMES系统以及测量NMES下的肌肉状态方面的潜在应用提供了直接证据。