CAJ | 학술논문

目的利用支持向量机(SVM)分类器,评价全脑扩散张量成像(DTI)对社交焦虑障碍(SAD)的诊断价值.方法采集19例SAD患者和19名年龄、性别及受教育程度与之相匹配的健康志愿者的全脑DTI数据.通过Diffusion Toolkit软件获取全脑的各向异性分数(FA).根据全脑FA数据,以基于SVM算法的脑影像数据模式识别(PROBID)软件对SAD患者和健康对照进行模式识别分析,评价基于SVM,以全脑FA数据诊断SAD的诊断价值.以阈值P≤0.001行置换检验评价诊断试验的可靠性;计算各体素的FA差异占全脑总差异的比率,得到其对分类所贡献的权重向量值.用MRIcro软件显示最高30％权重向量值且体素个数≥10的脑区.结果基于SVM,以全脑FA数据诊断SAD,准确度为92.11％(35/38),特异度为94.44％(17/18),敏感度为90.00％(18/20),阳性似然比为17.01,阴性似然比为0.11,诊断指数为184.22％.置换检验显示诊断结果可靠.对识别SAD贡献更大的白质脑区位于胼胝体膝部和压部、左侧钩束、左侧下纵束、左侧下额枕束、双侧额上回及左侧枕叶(体素数目21 ～ 136);对诊断健康对照贡献更大的白质区域包括双侧前扣带、左侧小脑中脚及左侧顶下小叶(体素数目14～112).结论基于SVM,以全脑FA数据诊断SAD准确性较高,全脑DTI特征有望成为诊断SAD的特异性指标;在神经精神疾病的白质研究中,SVM或许可用于验证研究结果和提供感兴趣脑区.
목적 이용지지향량궤(SVM)분류기,평개전뇌확산장량성상(DTI)대사교초필장애(SAD)적진단개치.방법 채집19례SAD환자화19명년령、성별급수교육정도여지상필배적건강지원자적전뇌DTI수거.통과Diffusion Toolkit연건획취전뇌적각향이성분수(FA).근거전뇌FA수거,이기우SVM산법적뇌영상수거모식식별(PROBID)연건대SAD환자화건강대조진행모식식별분석,평개기우SVM,이전뇌FA수거진단SAD적진단개치.이역치P≤0.001행치환검험평개진단시험적가고성;계산각체소적FA차이점전뇌총차이적비솔,득도기대분류소공헌적권중향량치.용MRIcro연건현시최고30％권중향량치차체소개수≥10적뇌구.결과 기우SVM,이전뇌FA수거진단SAD,준학도위92.11％(35/38),특이도위94.44％(17/18),민감도위90.00％(18/20),양성사연비위17.01,음성사연비위0.11,진단지수위184.22％.치환검험현시진단결과가고.대식별SAD공헌경대적백질뇌구위우변지체슬부화압부、좌측구속、좌측하종속、좌측하액침속、쌍측액상회급좌측침협(체소수목21 ～ 136);대진단건강대조공헌경대적백질구역포괄쌍측전구대、좌측소뇌중각급좌측정하소협(체소수목14～112).결론 기우SVM,이전뇌FA수거진단SAD준학성교고,전뇌DTI특정유망성위진단SAD적특이성지표;재신경정신질병적백질연구중,SVM혹허가용우험증연구결과화제공감흥취뇌구.
Objective To evaluate the clinical value of whole-brain diffusion tensor imaging(DTI) in diagnosing patients with social anxiety disorder(SAD) using an automated method based on support vector machine(SVM) classification.Methods Whole brain DTI data were collected from 19 patients with SAD and 19 age-,gender-and education-matched healthy control(HC) subjects.Fractional anisotropy(FA) of whole brain was obtained by input all tensor images into Diffusion Toolkit software.Based upon the characteristics of brain FA,the pattern recognition of brain image data(PROBID) toolbox on the grounds of SVM algorithm was employed to classify the subjects,evaluate the diagnostic value of whole-brain FA data based SVM in diagnosing SAD patients and verify the robustness of the diagnostic results using permutation test with the threshold at P≤0.001.The weight vector score of each voxel was calculated according to the ratio between this voxel and whole brain in FA differences of the two groups.The white matter regions identified by setting the threshold to the top 30％ of the weight vector scores with at least 10 contiguous voxels were demonstrated by MRIcro software.Results Diagnostic accuracy of whole-brain FA based SVM in diagnosing SAD was 92.11％ (35/38) in which the specificity was 94.44％ (17/18),the sensitivity was 90.00％(18/20),the positive likelihood ratio was 17.01,the negative likelihood ratio was 0.11 and the diagnostic index was 184.22％.Permutation test suggested that the diagnostic results were significantly reliable.White matter regions showing major contributions favoring SAD over HC were located in the genu and splenium of the corpus callosum,the left uncinate fasciculus,the left inferior longitudinal fasciculus,the left inferior fronto-occipital fasciculus,bilateral frontal gyri and the left occipital lobe.Whereas,white matter in bilateral anterior cingula,the left middle cerebellar peduncle and the left inferior parietal lobule showed more contributions to diagnose HC than to diagnose SAD.Conclusions As whole brain FA data based on SVM showing a high accuracy in diagnosing SAD,brain DTI characteristics have the potential to be the specific indicators in the diagnosis of SAD.SVM might be used as a tool to verify the reliability of white matter abnormalities and provide regions of interest in DTI study of neurological and psychiatric diseases.