肌张力障碍基底节环路的长时程动态神经解码和深部脑刺激作用建模研究

Dystonia is a movement disorder characterized primarily by abnormal muscle contractions. Due to the complexity of symptomatic phenotypes, there are several problems exist in clinical treatment with deep brain stimulation. The clinical outcomes could be varied for individual case and difficult to predict. It also has a generally long time of onset. Currently, the disease related function of the basal ganglia circuit remains unclear. There are few descriptions to quantify the symptomatic phenotypes of dystonia and there are no specific biomarkers reflect the dynamic changes of the symptoms. Thus, it is incapable of dynamically and adaptively adjusting stimulation parameters based on the effects of stimulation on symptoms and neural activities. Consequently, the optimization of parameter setting and the improvement of clinical outcome for the individuals would be limited. .In this project, we will analysis the neural activities recorded from the subthalamic nucleus and globus pallidus of patients with dystonia, and the electromyography signal simultaneously recorded from the symptom involved muscles. We will establish a multiple-feature neural decoding method to analyze the dynamic changes of the neural activities in the subthalamic-pallidal circuit. To quantitatively identify different symptomatic phenotypes, we will extract patterned features from the electromyography of the symptom involved muscles. Combined the multiple features of neural activities and the electromyography patterns, we would then explore the pathophysiological mechanism that how different phenotypes of symptom generate. Additionally, we will use the neural activities and the electromyography recorded during deep brain stimulation with varied parameters to study the modulations of stimulation. We will establish a Bayes model to describe the relationship among neural activities, symptomatic phenotypes and the stimulation parameters. The results and consequences of this project will be a basis for developing an adaptive deep brain stimulation based on the pathophysiological mechanism of disease.

肌张力障碍是一种肌肉异常收缩的运动障碍疾病。在深部脑刺激治疗中，由于疾病症状十分复杂，存在治疗效果个体差异大、整体起效时间长、治疗效果难预测等问题。目前对疾病相关的基底节环路功能不够明确，缺乏对症状表型的量化方法和反映症状变化的神经活动生物标记，因而无法根据刺激对症状和神经活动的作用，动态自适应地调整刺激参数，限制了深部脑刺激的个体参数优化和治疗效果提升。.本项目对长时程同步记录的肌张力障碍基底节环路的局部场电位信号和症状累及部位的肌电信号进行分析。采用基于动态多维特征的基底节环路神经解码方法，结合基于肌电模式特征的症状表型量化方法，探究不同症状表型发生的神经病理生理学机制；进而通过研究不同参数的刺激对神经活动和肌电模式的调控作用，建立基底节环路神经活动、症状表型、刺激参数之间的动态关系模型。为开发基于神经病理生理学机制的，自适应参数调节的深部脑刺激技术提供模型方法和理论基础。

肌张力障碍是一种由不自主的、持续或间歇性的肌肉收缩引起重复运动或姿势扭曲的神经系统疾病，本项目针对肌张力障碍疾病症状复杂和动态变化的特点，研究基底节核团的动态神经活动与症状特征之间的关系，完成肌张力障碍基底节局部环路神经活动多维特征动态分析、闭环神经调控算法和技术研究等工作。.具体的，研究肌张力障碍不同肌电表征的症状下，基底节苍白球多节律神经振荡的幅度平衡特征与症状严重程度的关系。针对基底神经节的丘脑底核与苍白球局部环路的动态相干谱、互信息、动态变异性特征解码。研究闭环神经调控刺激伪迹消除技术，提出基于非均匀采样的动态刺激伪迹去噪算法，可实现实时条件下对变参数刺激伪迹的消除。设计采用多节律动态神经解码的闭环策略，能够同时针对两种或以上的神经节律多状态进行针对刺激，并将方法进行了动物实验和两例临床预实验。开发用于神经调控远程长期管理的数字疗法系统软件，开展基于智能手环监测的运动症状判别研究及临床应用。.已发表文章3篇，2篇文章在投；获得“日内瓦国际发明展”银奖一项；申请发明专利3项、软件著作权1项；参加线上、线下科普宣传活动2次；合作主办国际会议3次；主办“脑刺激神经调控转化应用”系列线上讲座15期；将项目研究内容所涉及到的神经调控远程管理方案进行了应用转化，参加全国、地区、校级创新创业大赛并获得6次奖项。该项目针对临床问题出发，将为未来研发基于神经病理生理学机制的自适应深部脑刺激调控技术提供数据与算法支撑，并为发展面向真实世界的远程、长期、居家脑健康数字化，提供了研究和应用新思路。