BCI-FES康复训练平台与中风病人运动想象数据分析

发布时间：2018-07-02 20:27

本文选题：脑机接口 + 康复训练　；参考：《上海交通大学》2014年硕士论文

【摘要】：中风患者普遍存在一定程度上的运动功能障碍，，如何增强其肢体肌肉力量一直是康复训练学科的重要问题。然而，传统的康复训练方法效果极其有限。脑机接口（BCIs）是一种能够连接外部设备与人类大脑的技术，使得人类能够运用自己的思维来直接控制外部设备，而不需要肌肉、躯干的参与。因此，脑机接口可以被用于重建中风受损脑区，即利用运动想象与神经反馈增强运动控制网络重建的学习，恢复受损的运动控制功能。本论文首先介绍了基于脑机接口技术的康复训练系统；我们与复旦大学华山医院康复科合作，进行临床试验，采集EEG康复数据，建立数据库，并针对BCI康复数据进行分析，揭示基于运动想象的康复训练过程中的相关康复机理。我们提出了一套康复训练评价体系，并对传统的临床康复训练疗程和基于脑机接口的康复训练疗程进行了比较和评估。我们研究了脑机接口技术对于患者康复训练所起的作用，并通过比较得到了相应的定性和定量分析结果。此外，本论文提出了一种基于高斯混合模型和弱监督学习的二类运动想象分类算法。相比较于传统的公共空间模式算法，该算法能够更好的处理低信噪比的中风病人脑电信号。通过与传统算法的定量比较，我们验证了该算法对于中风病人脑电信号的适用性和有效性。通过对中风病人的脑电信号建立混合模型，我们进一步挖掘了中风病人康复过程中的脑区变化机理。
[Abstract]:Stroke patients generally have a certain degree of motor dysfunction, how to enhance their limb muscle strength has been an important issue in rehabilitation training. However, the effect of traditional rehabilitation training is extremely limited. Brain-Computer Interface (BCIs) is a technology that can connect external devices with human brain, which enables people to use their own thinking to directly control external devices without the participation of muscles and torso. Therefore, brain-computer interface can be used to reconstruct the damaged brain area of stroke, that is, using motor imagination and neural feedback to enhance the learning of motor control network reconstruction and restore the damaged motor control function. This paper first introduces the rehabilitation training system based on brain-computer interface technology. In cooperation with the Department of Rehabilitation of Huashan Hospital, Fudan University, we conduct clinical trials, collect EEG rehabilitation data, set up a database, and analyze BCI rehabilitation data. To reveal the mechanism of rehabilitation training based on sports imagination. We put forward a set of evaluation system of rehabilitation training, and compared and evaluated the traditional course of clinical rehabilitation training and the course of rehabilitation training based on brain-computer interface. We studied the effect of brain-computer interface technology on rehabilitation training of patients, and obtained the corresponding qualitative and quantitative results by comparison. In addition, this paper proposes a classification algorithm based on Gao Si hybrid model and weakly supervised learning. Compared with the traditional common space mode algorithm, the proposed algorithm can deal with the EEG signals of stroke patients with low SNR. Through quantitative comparison with the traditional algorithm, we verify the applicability and effectiveness of the algorithm to stroke patients. By establishing a mixed model of EEG in stroke patients, we further explore the mechanism of brain region change during rehabilitation of stroke patients.
【学位授予单位】：上海交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN911.7;R743.3

【参考文献】