Wenting Ren1, Peilin Lv2, Wei Deng3, Qizhu Wu1, Xiaoqi Huang1, Tao Li3, Qiyong Gong1, Su Lui1
1Huaxi MR Research Center (HMRRC), Department of Radiology,West China Hospital of Sichuan Universit, Chengdu, Sichuan, China; 2Department of Anesthesiology, West China Hospital of Sichuan University, Chengdu, Sichuan, China; 3Department of Psychiatry, West China Hospital of Sichuan University, Chengdu, Sichuan, China
A support vector machine classification approach combining gray matter volume and amplitude of low-frequency fluctuations information was utilized to distinguish schizophrenia patients from health controls. One hundred drug-naive first episode schizophrenia patients and 100 controls were scanned using a high-resolution 3D T1-weighted sequence and EPI sequence on a 3T MR imaging system. The classification yielded an accuracy of 83.5%. For the first time, we provides the evidence for evaluating the MR diagnostic value in a large sample of antipsychotic-naive first-episode schizophrenia, supporting the anatomical and functional deficits could be used as a biomarker for the diagnosis of schizophrenia.