欢迎访问《同济大学学报（医学版）》编辑部！

引用本文：

葛志超,吴佳佳,孙　浩,等.脊髓型颈椎病患者大脑内在活动的改变和基于fMRI的分类研究[J].同济大学学报（医学版）,2021,42(6):785-790. [点击复制]
GE Zhi-chao,WU Jia-jia,SUN Hao,et al.Changes of intrinsic brain activity in patients with cervical spondylotic myelopathy based on fMRI classification[J].同济大学学报（医学版）,2021,42(6):785-790. [点击复制]

【打印本页】【在线阅读全文】【下载PDF全文】【查看/发表评论】【下载PDF阅读器】【关闭】

本文已被：浏览 369次下载 166次	码上扫一扫！
脊髓型颈椎病患者大脑内在活动的改变和基于fMRI的分类研究
葛志超,吴佳佳,孙　浩,赵　庆
0 字体:加大+\|默认\|缩小-
(上海中医药大学附属岳阳中西医结合医院骨伤科，上海　200437;上海中医药大学附属岳阳中西医结合医院康复科，上海　200437)

摘要:

目的　分析脊髓型颈椎病（cervical spondylotic myelopathy, CSM）对大脑内在活动的影响，探讨CSM患者肢体功能障碍的脑重塑机制。方法　纳入12例CSM患者和18名健康志愿者，进行日本骨科协会评估治疗分数（Japanese Orthopaedic Association Scores, JOA）评定及静息态功能磁共振成像（functional magnetic resonance imaging, fMRI）扫描，计算低频振荡振幅（amplitude of low-frequency fluctuation， ALFF），将组间具有显著差异的ALFF作为特征值，通过支持向量机（support vector machine, SVM）方法对两组特征值进行二分类，并进行临床功能评估与分类决策值之间的相关性分析。结果　CSM患者的大脑内在自发活动性较健康对照升高，主要在左侧小脑、左侧中央前回、左侧辅助运动区、右侧前扣带回。将这些脑区作为判别分析的模板进行两组SVM分类，准确率为8667%（P=0001），灵敏度为7500%，特异度为9444%。CSM患者的SVM分类决策值与JOA运动功能评分显著负相关（r=-0761 5，P=0004），与JOA感觉功能评分无显著相关性（r=-0070 5，P=0828）。结论　SVM分类表明差异脑区的自发活动可以在个体水平较准确区分CSM患者和健康对照；CSM患者可能需要调动更多的运动中枢和运动调控脑区参与功能代偿，且以优势半球为主。

关键词: 脊髓型颈椎病支持向量机静息态功能磁共振成像

DOI：10.12289/j.issn.1008-0392.21082

投稿时间：2021-03-06

基金项目:国家自然科学基金青年项目(81902301)

Changes of intrinsic brain activity in patients with cervical spondylotic myelopathy based on fMRI classification

GE Zhi-chao,WU Jia-jia,SUN Hao,ZHAO Qing

(Dept. of Orthopedics and Traumatology, Yueyang Hospital of Integrated Traditional Chinese and Western Medicine, Shanghai University of Traditional Chinese Medicine, Shanghai 200437, China;Dept. of Rehabilitation, Yueyang Hospital of Integrated Traditional Chinese and Western Medicine, Shanghai University of Traditional Chinese Medicine, Shanghai 200437, China)

Abstract:

Objective　To analyze the intrinsic brain activity in patients with cervical spondylotic myelopathy（CSM) based on MRI classification. Methods　Twelve CSM patients and 18 healthy controls were enrolled, and the Japanese Orthopedic Association Scores（JOA) assessment and resting state functional magnetic resonance imaging（fMRI) scan were performed. The amplitude of low-frequency fluctuation(ALFF) were calculated, and CSM patients and healthy controls were classified with support vector machine（SVM) method based on ALFF feature values of brain areas which showed significant differences between two groups. The correlation analysis was performed between clinical assessments and classification decision values. Results　The intrinsic spontaneous activity of the brain in CSM patients was significantly higher than that in healthy controls, mainly in the left cerebellum, the left precentral gyrus, the left supplementary motor area, and the right anterior cingulate cortex. Using these brain regions as a mask for SVM classification, the accuracy rate was 8667%（P=0001), the sensitivity was 7500%, and the specificity was 9444%. Moreover, the SVM classification decision value of CSM patients was negatively correlated with JOA motor function score（r=-0761 5,P=0004), but not correlated with JOA sensory function score（r=-0070 5, P=0828). Conclusion　SVM classification indicates that the intrinsic spontaneous activities of brain regions which showed significant differences between groups may be used to classify CSM patients and healthy controls at the individual level. CSM patients might need to recruit more motor centers and motor-control brain regions to compensate the dysfunction, mainly involving the dominant hemisphere.

Key words: cervical spondylotic myelopathy support vector machine functional magnetic resonance imaging