%0 Journal Article %T 颞叶癫痫病患者的脑皮层多特征分类<br>Classification using multiply cortical features of temporal lobe epilepsy patients %A 郭圣文 %A 赖春任 %A 汪文胜 %A 成丽娜 %A 吴凯 %A 岑桂英 %A 池敏越 %J 东南大学学报(自然科学版) %D 2015 %R 10.3969/j.issn.1001-0505.2015.06.006 %X 为了自动识别左、右侧颞叶癫痫病患者,提出了一种基于脑皮层结构特征的分类方法.采集了21例左侧TLE患者(LTLE)、18例右侧TLE患者(RTLE)与28例健康正常人(NC)的脑结构磁共振图像,提取与分析其脑皮层形态结构特征.为了降低特征维数,利用递归特征消除法获取对分类有效的主要特征.然后,采用支持向量机对3组人群进行分类. 结果表明,大脑皮层表面积为最佳分类特征,灰质体积也具有突出的分类能力,皮层厚度识别不同侧TLE的能力稍差,平均曲率对右侧TLE与健康人群的区分能力相对较弱.对分类有重要影响的脑区主要分布于颞叶、额叶与顶叶,尤其是大脑左侧脑区.由此证明利用脑皮层特征可以有效地区分不同侧TLE患者和健康正常人,TLE患者的左脑更易受损.研究结果有助于理解TLE患者的病理机制与进展规律、确定病灶位置及实现自动诊断.<br>A classification method based on cortical features is proposed to automatically discriminate left temporal lobe epilepsy(LTLE)and right temporal lobe epilepsy(RTLE). Brain structural magnetic resonance images from 21 LTLE patients, 18 RTLE patients and 28 NCs(normal controls)are collected, and the morphological cortical features are extracted and analyzed. In order to reduce features dimensions, the recursive feature elimination(RFE)strategy is applied to determine salient features for classification. Then, support vector machine(SVM)is used to classify the three cohorts. The results show that the surface area is the best feature for classification, and the gray matter volume has also prominent capability in discrimination. The cortical thickness exhibits limited ability to differentiate LTLE patients from RTLE patients, and the mean curvature behaves relatively weakly to distinguish RTLE patients from NCs. The dominant regions for classification are located in the temporal, the frontal and parietal lobe, especially in the left hemisphere. The effectiveness of cortical features for discrimination of three cohorts is proved, and the brain regions in left hemisphere of TLE patients are more vulnerable to injury. This work provides potential information to understand the pathogenic mechanism and progress, determine the focus and facilitate computer-aided diagnosis of TLE patients %K 颞叶癫痫 %K 脑皮层特征 %K 支持向量机 %K 分类< %K br> %K temporal lobe epilepsy %K cortical features %K support vector machine %K classification %U http://journal.seu.edu.cn/oa/darticle.aspx?type=view&id=201506006