Muscle activation patterns in gait with cerebral palsy

患有神经性步态障碍的人，如脑瘫(CP)，表现出与典型发育受试者不同的步态模式。使用3D步态分析、视频和临床检查以及最近的人工智能工具，已经进行了许多尝试来对CP患者的步态模式进行分类。已经开展了几项研究，以开发能够区分典型步态和病理性步态的计算技术。这些研究的主要问题是，他们试图仅使用数学和计算技术来改善和解释肌电信号。很少有研究真正监测肌电与其他机械和临床步态参数之间的关系。然而，在研究人员和临床医生中，肌电数据一直被认为是一个常规的二级信息源，可以与其他参数一起使用，对治疗决策的影响可能较小。我们的假设是，在整个步态周期中确定和提取肌电特征，然后在治疗过程中同时分析和识别EMG数据的模式和其他步态参数，可以为临床医生提供与该治疗相关的有用和关键的建议。因此，我们在这个项目中的目标是开发类似于全球步态指数的全球EMG评估方法，用于表征神经运动参与的程度，以及一个面向特征的分析框架，用于将EMG数据与关节活动范围、肌肉力量、痉挛以及步态特征的临床测量相结合。这最终可能有助于在CP患者中建立以规则为导向的骨科治疗决策树。访问大型数据库是该项目成功的关键部分。我们的档案包括1993-2020年来自1250多名双侧痉挛CP患者和300多名单侧CP患者的数据，其中我们总共有超过2550次肌电检查。我们的工作流程计划包括6个工作包(WP)。WP1(数据编制)将为本项目的目的组织和定制数据。WP2(主观评估)将与临床专家合作评估肌电信号。WP3(肌电模式识别)将在对表型进行分类的横断面研究中评估肌电数据。在纵向研究中，将监测肌电随年龄和骨科干预的变化(WP4)。在WP5中，将制定一个肌电综合指数。WP6(治疗决策树)将根据临床文献和肌电数据制定治疗路径，以改进CP治疗的决策。

People suffering from neurological gait disorders such as cerebral palsy (CP) demonstrate gait patterns differing from those of typically developing subjects. Many attempts have been undertaken to classify gait patterns of patients with CP using 3D gait analysis, video and clinical exams and recently also with artificial intelligence tools. Several studies have been performed for developing computational techniques to enable the classification between typical and pathological gait. The main issue with these studies is that they attempt to improve and interpret EMG signals using mathematical and computing techniques only. There are few studies that actually have monitored the relationship between EMG and other mechanical and clinical gait parameters. However, the EMG data, among researchers and clinicians has been considered as a regular secondary information source that could be used along with the other parameters with potentially having less influence on treatment decision making. Our hypothesis is that determining and extracting EMG features throughout the gait cycle and then analyzing and recognizing the pattern of the EMG data in parallel with the other gait parameters during a treatment process can provide useful and critical recommendations for the clinicians in relation to that treatment. Our aim in this project is therefore to develop global measures for EMG assessment similar to global gait indexes for characterizing a degree of neurologic motor involvement as well as a feature-oriented analysis framework for setting EMG data in context with clinical measures of joint ranges of motion, muscle strength, and spasticity as well as with gait features. This ultimately may help in establishing rule-oriented orthopedic treatment decision trees in patients with CP. Accessing to a large database is an essential part of succeeding with this project. Our archive consists of data from the time 1993-2020 derived from more than 1250 patients with bilateral spastic CP and more than 300 patients with unilateral CP of which we have in total more than 2550 EMG exams available. Our work flow program includes 6 work-packages (WP). WP1 (Data preparation) will organize and customize the data for the purpose of this project. WP2 (Subjective assessment) will assess EMG signals in collaboration with clinical experts. WP3 (EMG pattern recognition) will evaluate the EMG data in a cross-sectional study for classifying phenotypes. In longitudinal studies, EMG changes with age and orthopedic intervention will be monitored (WP4). In WP5 a summary EMG index will be developed. WP6 (Treatment decision trees) will formulate treatment paths according to the clinical documents and the EMG data for improved decision making in the management of CP.