Objective:To investigate the associations of brain white matter integrity deficits, and to explore the association of fractional anisotropy (FA) abnormality with clinical symptoms and cognitive impairments, as well as the prediction effect of the FA alterations on symptoms and cognitive function outcomes in the acute stage of schizophrenia from the whole brain level based on magnetic resonance diffusion tensor imaging (DTI).Methods:From November 2019 to December 2020, thirty-eight patients with first-episode schizophrenia and 38 healthy controls were recruited in this study. Wisconsin card classification test (WCST), digit span test (DST forward/backward), verbal fluency test, Stroop (A/B/C), trail making test (TMT-A/B), as well as positive and negative syndrome scale (PANSS) were utilized to evaluate patients' cognitive function and clinical symptoms both before and after 8 weeks of risperidone monotherapy. T1-weighted images and DTI data of all the subjects were collected . FSL and SPM8 were used to preprocess MRI data and compare the between-group differences of FA. Support vector machine (SVM) analysis was used to evaluate the accuracy of abnormal FA values in differentiating schizophrenic patients from healthy controls. Finally, stepwise multiple regression analysis or generalized linear models were used to explore the associations between white matter integrity and symptoms or cognition.Results:Before treatment, patients' FA values of right medial temporal lobe (mTL), cuneus, anterior cingulate gyrus (ACG) and inferior parietal lobe (IPL) were significantly lower than those in healthy controls ( P<0.01, GRF corrected), and patients' FA values of bilateral middle cingulate gyrus (mCG) were significantly higher than those in the control group ( P<0.01, GRF corrected). SVM analysis showed that four combinations could distinguish the patients from the control with the most accurate rate of 89.47%. Patients' baseline decreased FA values in the right IPL were positively associated with their increased total response time in WCST ( β=0.489, P=0.003, FDR corrected), correct response time in WCST ( β=0.450, P=0.008, FDR corrected), error response time in WCST ( β=0.435, P=0.008, FDR corrected), TMT-B ( β=0.296, P=0.042, FDR corrected), Stroop-C ( β=0.345, P=0.035, FDR corrected), and PANSS-P ( β=0.321, P=0.042, FDR corrected). Reduced FA values in the right mTL in patient group were significantly negatively related to the total time spent on the TMT-A ( β=-0.425, P=0.009, FDR corrected) and TMT-B ( β=-0.325, P=0.026 with FDR correction). Increased FA values in right mCG in patient group demonstrated positive associations with total response time in the WCST ( β=0.585, P=0.002, FDR corrected), correct response time in WCST ( β=0.524, P=0.003, FDR corrected), error response time in WCST ( β=0.536, P=0.003, FDR corrected) and total time consuming in TMT-B ( β=0.484, P=0.004, FDR corrected), as well as negative associations with DST-forward ( β=-0.319, P=0.042, FDR corrected). After treatment, patients' percentage changes in total response time of WCST ( β=0.715, P<0.001, FDR corrected), correct response time of WCST ( β=0.752, P<0.001, FDR corrected), as well as total time-consuming of TMT-A ( β=1.333, P=0.001, FDR corrected) showed positive correlations with baseline increased FA values in the left mCG. Percentage alteration of Stroop-B was negatively correlated with baseline FA values in the right cuneus ( β=-0.745, P=0.015, FDR corrected). Conclusions:The combination of abnormal FA values in multiple brain regions may be potential biomarkers to distinguish schizophrenic patients from healthy volunteers. There was regional dependence in the associations of the impairment of white matter integrity with the cognitive impairment, the severity of psychopathological symptoms as well as the prognosis of patients in the acute stage.