Although morphological changes of subcortical gray matter structures have been discovered in subjects with methamphetamine(METH) psychosis[1, 2] , the volume changes of amygdala and hippocampus in METH psychosis is still in debate. Fewer neuroimaging studies have focused on the microstructure of these structures in METH users. The purpose of this study is to evaluate the fractional anisotrophy (FA) index on diffusion weighted imaging (DTI), as well as volume, of subcortical gray matter structures in METH users with psychosis. Human Subjects: Eighteen subjects with METH psychosis (age, 34.2±7.1 yrs; duration of METH use, 3.98±2.58 yrs) and 18 normal volunteers (age, 30.2±6.8 yrs; sex, parental socio-economic background, and IQ matched) were enrolled. MRI Acquisition: MR imaging was employed on 3T MR scanner (GE Discovery MR750). 3D-FSPGRIR acquisition was performed to obtain structural T1WI (FOV = 256 × 256 mm2, matrix = 256 × 256, resolution = 1 × 1 × 1 mm3, slices = 192, TR/TE/TI = 7.4/3.2/450 ms, and bandwidth = 27.78 Hz/pixel). SSSE-EPI acquisition was used to obtain DTI (FOV = 256 × 256 mm2, matrix = 128 × 128, resolution = 2 × 2 × 2 mm3, slices = 68, TR/TE = 8175/80.8 ms, FA = 90°, bandwidth = 250 Hz/pixel). Image Analyses: FSL-First carried out with FSL tools was used to assess all subcortical gray matter structures, i.e. hippocampus, putamen, caudate, amygdala, accumbens, pallidum (Figure 1). A comparison of the volume and mean FA values of these structures was carried out between METH and control groups using ANCOVA. Correlation analysis were performed between volume, FA of each structure and METH use duration, dose, and clinical psychotic measures, such as Hamilton Anxiety Scale (HAMA) and Brief Psychiatric Rating Scale (BPRS). Stepwise Binary logistic regression analysis and receiver-operating characteristic (ROC) curves were applied to evaluate the diagnostic significance of subcortical gray matter structure.No volume differences in subcortical gray matter structures were detected between METH and normal control groups. The METH group was witnessed a significant reduction in FA values of bilateral amygdalae and hippocampi (p<0.05, FDR corrected), while the degree of FA reduction was the most severe in the left amygdala. No evidence showed a correlation between subcortical gray matter structures and METH use duration, dose, as well as clinical psychotic measures. In logistical regression analysis, only left amygdala FA value entered the final model with an accuracy of 80.56% (Figure 2).In this study, we demonstrate alteration of microstructure in bilateral amygdalae and hippocampi in METH users with psychosis, who presents no evident volume changes in the two structures. The finding agrees with existing study results that hippocampus and amygdala do play a role in the pathophysiology of psychostimulant addiction, affective psychosis and schizophrenia [1]. Based on [3]), we further deduce a hypothesis: Reduced FA in amygdala may indicate enhanced information transmission involved in high relative activity in amygdala, which is believed to produce psychotic illness[4]. Damage to hippocampal neurogenesis in METH users has been revealed [5], which may cause the FA reduction in hippocampus. This study indicates that DTI-FA can be useful in detecting amygdala and hippocampal microstructural changes prior to presentation of morphologcial changes in METH users with psychosis. The prominent FA reduction in amygdala rather than that in hippocampus could be relatively specific characteristics of METH psychosis. Subregions of amygdala and hippocampus is the focus of further neuroimaging studies.