This study aim to investigate the relationship between depression after onset of stroke and superior longitudinal fasciculus and cingulate gyrus with multi-parameter DTI comparisons. These two brain structures distal to infarct regions were obtained by automatic segmentation and four parameters based on intensity distribution (mean, standard deviation, skewness and kurtosis) were quantitatively measured on each structure. Significant difference in these two structures was found among major depression subjects, mild depression subjects and the normal control. Our results verified that PSD patients latently exhibit neuroanatomical changes in superior longitudinal fasciculus and cingulate gyrus.Post stroke depression (PSD) is one of the most prevalent emo­tional disorders afflicting stroke patients. Despite being one of the direct causes of depression, whether stroke-induced neuroanatomical deterioration plays an important role in the onset of PSD is still unknown. Over the recent years, voxel-based analysis is widely used to detect structural abnormalities in post stroke depression (PSD)¹. However, this method inevitably ignored the infarct lesion impact in calculating the positive clusters, indicating inaccurate results. Due to no clear relationship between the stroke lesion location and PSD occurrence in current research, we specifically hypothesized that some neuroanatomical abnormalities distal to infarct lesions would contribute to risk of PSD onset. Previous neuroimaging studies have reported that neuroanatomical changes in superior longitudinal fasciculus and cingulate gyrus were associated with late-life depression². In this study, we aim to investigate the relationship between depression after onset of stroke and superior longitudinal fasciculus and cingulate gyrus with multi-parameter DTI comparisons. Twenty-two PSD subjects (12 men and 10 women, mean age 65.5±7.3 years) and thirteen healthy subjects (5 men and 8 women, mean age 40.1±11.5 years) as normal control (NC) were included in the current study. PSD subjects were separated into the major depression group and the mild depression group according to the 24-item Hamilton Rating Scale for Depression (HAMD-24) within 1 month after stroke onset. All the subjects underwent 1.5T MR imaging with T1 and DTI scanning within two weeks after stroke onset. The typical MRI protocol consisted of T1-weighted sequence (TE=5 ms, TR=195 ms, Flip Angle=70°, acquisition matrix=768×624, FOV=230×187 mm2) and diffusion tensor imaging sequence (TE=88ms, TR=2700ms, Flip Angle=90°, acquisition matrix=256×256, FOV=250×250 mm2 , b0=1000 s/mm2). Diffusion weighting consists of 20 non-collinear directions, and a non-diffusion tensor image whose b0=0 s/mm2. Final infarcts with corresponding volume were manually defined on diffusion weighted images (DWI). After image preprocessing, all the subjects T1 images were parcellated through MRICloud platform, which is an atlas-based automatic T1 segmentation pipeline invented by JHU (www.mricloud.org). Then for each of our subject, the parcellation maps obtained from MRICloud were registered to corresponding DWI, apparent diffsuion coefficient (ADC) and FA images separately with manual correction. An experienced neurologist verified that the obtained regions of interests (ROI), i.e. superior longitudinal fasciculus and cingulate gyrus, were not included in the infarct regions. For each type of DTI coefficient images, four parameters based on intensity distribution (mean, standard deviation, skewness and kurtosis) were quantitatively measured on each ROI with ANOVA analysis. The demographic and focal information of all subject groups are shown in Table 1. Figure 1 shows diffusion parametric maps for one representative PSD subject with segmentation of superior longitudinal fasciculus and cingulate gyrus. Only those characteristics differing among three subject groups significantly are shown in Table 2. In the case of superior longitudinal fasciculus, SD of FA consistently decrease with severity of depress symptom on the left side (p=0.034). SD of DWI also declines with severity of depress symptom on the right side (p=0.009). As for cingulate gyrus, mean of ADC, SD of ADC, and SD of DWI in NC group are significantly lower than both PSD groups on both sides. It is also observed that skewness of FA exhibit high level significantly in NC group on the right side.In this study we verified that PSD patients latently exhibit neuroanatomical changes in superior longitudinal fasciculus and cingulate gyrus. This result may explained by the disconnection theory in vascular depression: focal damage to specific fiber tracts and neural circuits may contribute to depression by disrupting neural connections among regions regulating mood and cognition³. Our study coincides well with the fact that superior longitudinal fasciculus and cingulate gyrus have extensive connectivity within prefrontal-subcortical circuit, which plays an important role in cognition and mood control. This study may pro­vide new insights into the neuroanatomical mechanism of PSD.