北京生物医学工程

基于BP神经网络的手功能康复评估研究

Research on hand function rehabilitation evaluation based on BP neural network

作者：谷雯雪王殊轶樊琳林晨琳吕燕飞

单位：上海理工大学医疗器械与食品学院（上海200093）威海市中医院（山东威海 264200）

关键词：手功能康复; 神经网络; 评估; 握力; 肌电

分类号：R318.04

出版年·卷·期（页码）：2020·39·6（622-626）

摘要：

目的针对手功能障碍康复,本文提出了一种基于神经网络的主客观结合的康复评估新方法。方法选取27例偏瘫患者进行抓握测试，记录患者的握力、捏力、肌电数据，采用Fugl-Meyer评定法(Fugl-Meyer assessment, FMA)得到患者的手功能康复评估分数。以握力和肌电等参数为输入，以评估分数为输出，建立基于BP(Back-Propagation)神经网络的手功能康复评估模型，利用获取的24组样本数据对神经网络进行训练，余下的3组样本用于精度验证，并通过聚类分析进一步提高预测精度。结果训练后，模型的平均评估误差率为18.13%，通过调整样本类数，平均评估误差最大降至15.84%。结论评估模型可以较为准确地预测评估分数，BP神经网络可以用于评估手功能障碍的康复程度。

Objective Aiming at the hand disfunction rehabilitation, a combination of the subjective and objective method of rehabilitation assessment based on neural network is proposed. Methods 27 patients with Hemiplegia were selected for the grip test. The gripping force, pinch force and electromyographic data of the patients were recorded. The Fugl-Meyer assessment (FMA) was used to obtain the hand function assessment scores. Using parameters such as grip strength and myoelectricity as input, and evaluation score as output, a hand function rehabilitation evaluation model based on BP (Back-Propagation) neural network was established, and the neural network was trained using the obtained 24 sets of sample data. The remaining 3 sets samples are used for accuracy verification, and cluster analysis further improves prediction accuracy. Results After training, the evaluation error rate of the model was 18.13% . By adjusting the number of sample classes, the maximum evaluation error was reduced to 15.84% . Conclusions The evaluation model can more accurately predict the evaluation score, and the BP neural network can be used to evaluate the degree of rehabilitation of hand dysfunction.

参考文献：

[1] 沈设芬，李瑛，毛云英，等．老年脑卒中患者的康复需求及影响因素分析[J].护士进修杂志，2012，27( 9) : 854－855．
[2] Hu XL, Tong KY, Wei XJ, et al. The effects of post-stroke upper-limb training with an electromyography (EMG)-driven hand robot[J]. Journal of Electromyography and Kinesiology, 2013, 23(5):1065-1074.
[3] 毛英．早期健康教育对脑卒中患者疾病认知程度的影响[J]．中国当代医药,2011，18（21）：178-179.
[4] 贾杰. 脑卒中后手功能康复应评价和治疗并重[J]. 上海医药, 2014, 35(2):6-9.
Jia J. Rehabilitation of hand function after stroke requires attention to both assessment and therapy[J]. Shanghai Medical, 2014, 35(2):6-9.
[5] Malouin F, Pilchard L, Boneau C, et al. Evaluating motor recovery early after stroke: comparison of the Fugl-Meyer assessment and the motor assessment scale[J]. Archives of Physical Medicine & Rehabilitation, 1994, 75(11):1206-1212.
[6] 王新德，朱镛连. 神经病学：神经康复学[M]．北京：人民军医出版社，2001.
[7] 秦茵，毕胜，王福根．脑卒中上肢功能常用评价方法及临床应用[J]．中国康复医学杂志，2004,19（3）：232-235.
[8] 王博超, 张新峰. 基于人工智能和临床诊断的上肢康复评估方法研究进展[J]. 北京生物医学工程, 2018，37（1）：103-108.
Wang BC, Zhang XF. Research progress of assessment methods based on artificial intelligence and clinical diagnosis in upper limb rehabilitation [J]. Beijing Biomedical Engineering, 2018，37（1）：103-108.
[9] 王丽, 张秀峰, 马岩, 等. 脑卒中患者上肢康复机器人及评价方法综述[J]. 北京生物医学工程，2015，34(5): 526-532.
Wang L, Zhang XF, Ma Y, et al. Summary of rehabilitation robot for upper limbs and evaluation methods for stroke patients [J]. Beijing Biomedical Engineering ，2015，34(5): 526-532.
[10] 杨延砚，周谋望，黄东峰．最大握力和捏力检测用于脑卒中患者上肢功能评定的研究 [J]．中国康复医学杂志，2008，23（5）：395-397.
Yang YY, Zhou MW, Huang DF. Standard maximal grip/pinch force test in stroke patients[J]. Chinese Journal of Rehabilitation Medicine, 2008, 23 (5): 395-397.
[11] Lienhard K, Cabasson A, Mest O. sEMG during whole-body vibration contains motion artifacts and reflex activity[J]. Journal of Spots Science medicine, 2015, 14 (1): 54-61.
[12] 袁曾任. 人工神经元网络及其应用[M]. 北京：清华大学出版社，1999：5-14，66-77.
[13] 王桂丽. 应用sEMG评估手功能支具在脑卒中腕手功能障碍的作用研究[D]. 福州：福建中医药大学， 2017.
Wang GL. sEMG assessment of wrist and hand of dysfunction recovery by hand splint  treatment in post-stroke patients：a clinical study [D]. Fuzhou：Fujian University of Traditional Chinese Medicine，2017.
[14] Puh U. Age-related and sex-related differences in hand and pinch grip strength in adults[J]. International Journal of Rehabilitation Research, 2010, 33 (1): 4-11.
[15] 包蕾.表面肌电在康复握力训练中的应用[D]. 2018. （在万方和知网中没查到此文献）
Bao lei. Application of surface electromyography in rehabilitation grip training[D]. 2018.
[16] Kolich M. Predicting automobile seat comfort using a neural network[J]. International Journal of Industrial Ergonomics, 2004, 33(4):285-293.
[17] 史峰，王小川，郁磊，等. MATLAB神经网络30个案例分析[M]. 北京：北京航空航天大学出版社, 2010.
[18] 马佳, 柯艺杰, 苏强,等. 汽车座椅舒适度人工智能评价方法研究[J]. 机械科学与技术, 2011，30(3)：419-422.
Ma J, Ke YJ, Su Q, et al. An automobile seat comfort evaluating method based on artificial intelligence [J]. Mechanical Science and Technology for Aerospace Engineering, 2011，30(3)：419-422.

服务与反馈：

【文章下载】【加入收藏】

提示：您还未登录，请登录！点此登录