Mindfulness-based stress reduction (MBSR) is modified from the Buddhist traditions and aims to develop an individual’s moment-to-moment and non-judgmental awareness through various practices_.¹ Previous study has suggested MBSR improves three components of self-regulation, including attention control, emotion regulation, and self-awareness.² However, brain mechanisms of these improvements remain unknown. In this study, we employed the resting-state functional MRI (rsfMRI) to investigate changes of functional connectivity (FC) before and after MBSR practice (self-awareness of the breath), and before and after 8-week MBSR training. We hypothesized changes in FC may reflect improvements of self-regulation after MBSR training.Study design: Fifteen subjects (9 females and 6 males, age = 41.67 ± 8.41 years), who were the beginners of MBSR practice with no history of neurological or psychological disorders, participated this study and 8-week MBSR training. All subjects received MRI scans twice before and after MBSR training (Figure 1). Each MRI scan consisted of T1W imaging and two runs of 6-min rsfMRI (before and after MBSR practice). Subjects were instructed to stop mindfulness practice, close eye, remain awake, and relax during each run of rsfMRI scan. In the period between two runs of rsfMRI, subjects were asked to focus on MBSR practice for 20 minutes. MRI data acquisition: MR scanning was performed on a 3T MRI system (TIM Trio, Siemens) with a 32 channel phased-array head coil. T1W imaging was performed using a 3D-MPRAGE sequence: TR / TE = 2000 ms / 3 ms, flip angle = 9°, FOV = 256 × 192 × 208 mm^3, matrix size = 256 × 192 × 208, and spatial resolution = 1 mm^3. rsfMRI was performed using a 2D gradient EPI sequence: TR / TE = 2000 ms / 24 ms, flip angle = 90°, FOV = 256 × 256 × 3 mm^3, matrix size = 64 × 64 × 34, and 180 volumes per run. Data processing: rsfMRI data was analyzed using data processing assistant for resting state fMRI³ (DPARSF) based on SPM8. The procedure of data preprocessing involved slice timing correction, motion correction using a linear affine transformation, co-registration, segmentation and spatial normalization to MNI space using DARTEL, smoothing with 4 mm FWHM, and band pass filtering (0.01-0.08 Hz). Finally, a nuisance regression was applied to correct for polynomial trend, six-parameter rigid-body parameters, head motion scrubbing, mean signals of white matter and cerebrospinal fluid. The seed-based analysis was performed to obtain whole brain FC maps by placing a spherical seed in the posterior cingulate cortex (PCC, the center coordinate: [-5, -49, 40], radius = 6 mm) in the Talairach space.⁴ FC maps were computed by voxel-wise Pearson correlation between a time course of each voxel over whole brain and averaged time course of predefined PCC seed. FC maps were then transformed to z-maps through Fisher Z-transform. We used SPM8 to perform two-way ANOVA (factor 1: pre-/post-mindfulness, factor 2: pre-/post-training) to access significant differences in z-maps between different contrasts.In the contrast of post-mindfulness > pre-mindfulness with post-training, we found significantly increased FC in the right lateral middle orbitofrontal cortex (LMOFC) and the subgenual anterior cingulate cortex (sgACC) (p < 0.01, uncorrected, size > 10 voxels) (Figure 2). In the contrast of post-mindfulness > pre-mindfulness with pre-training, we did not find any significant result. We further investigated the effect of MBSR training by comparing with the above two contrasts. Both the right sgACC and LMOFC showed significantly increased FC in the contrast of post-mindfulness > pre-mindfulness with post-training > pre-training (p < 0.05, uncorrected, cluster size > 20 voxels). There was no significant result in the contrast of post-mindfulness > pre-mindfulness with pre-training > post-training.In this study, we adopted a longitudinal experiment design to access FC changes in the resting state due to MBSR practice and training. Compared to pre-mindfulness, the results exhibited increased FC in the right sgACC and the LMOFC with the PCC in post-mindfulness after MBSR training (Figure 3). These two regions and the PCC have been considered to compose a brain network about emotional processing and regulation.⁵ Previous studies using functional imaging have suggested that the enhanced activations in these regions may be associated with reduced anxiety after mindfulness practice⁶ and improved emotion regulation.⁷MBSR strengthens FC couplings of the right sgACC and LMOFC with the PCC in the beginners after 8-week MBSR training. This finding reveals an underlying neural mechanism of the positive effects of MBSR practice on emotional processing and regulation.