Abstract: Background: Animal models and cytological studies suggested that Rett syndrome(RTT) had neuronal development and connection deficits,but it still remains unclear how the brain white matter develops and connects in patients. Objective: To explore the abnormalities of multiple white matter parameters using diffusion tensor imaging (DTI) in RTT children and its correlation with disease severity and developmental level. Design: Case-control study. Methods: Children aged 2 to 6 years old, diagnosed with typical RTT and confirmed with MECP2 mutation or deletion, were recruited as the RTT group. Routine MRI scan, Griffith Developmental Scale-Chinese Version (GDS-C) and RTT Severity Score (RSSS) were taken as evaluations. Children aged 2 to 6 years with normal Developmental Screening Test (DST) scores, undergoing MR scan because of paroxysmal dizziness or headache but without recognizable lesions or abnormalities on scans, were recruited as typically developing (TD) controls. GDS-C, RSSS and DST were performed in the Health Care Department of Children Hospital of Fudan University. MR was scanned in the same GE 750 3T scanner in the Radiology Department of our hospital. All of the GDS-C, RSSS, DST and MR scans were completed within one month. Four DTI parameters were collected and tract-based spatial statistics (TBSS) were applied to obtain DTI parameter differences and locations between two groups. Main outcome measures: Abnormal white matter brain areas in RTT and the correlations with RSSS and developmental domains of GDS-C. Results: From June 2014 to June 2021, we recruited 22 children as RTT group and 23 children as TD controls. The age difference was not statistically significant. The RSSS score in RTT was 5.5±1.33. The total DQ of GDS-C was 15.8±6.8 and as in domains ADQ was 30.2±10.6, BDQ was 14.6±9.1, CDQ was 15±7.7, DDQ was 9.9±6.2, EDQ was 9.2±3.9. DST score in TD group ranged from 94 to 118 （103±7）. The general FA value in RTT group was lower than that in TD group, MD and DR values in RTT group was higher than that in TD group, and no group difference was found as for DA. DTI-TBSS analysis results (FDR corrected, P<0.05)showed the fibers with lower FA values in RTT group were mainly located in the forceps minor, splenium of corpus callosum, bilateral superior longitudinal fasciculus, left inferior longitudinal fasciculus, retrolenticular part of internal capsule and uncinate fasciculus;the fibers with increased MD values in the RTT group were mainly located in the body of corpus callosum;the fibers with increased DR values in the RTT group were mainly located in the genu of corpus callosum;the fibers with increased DA values in the RTT group were mainly located in the body and splenium of corpus callosum. In RTT group, RSSS was negatively correlated with FA values in splenium of corpus callosum and left superior longitudinal fasciculus, and was positively correlated with MD value and DR value in the body of corpus callosum. FA value of retrolenticular part of internal capsule was positively correlated with the total DQ and DQs of each domain from the GDS-C. DR value of right sagittal stratum was negatively correlated with the total DQ and DQs of each domain from the GDS-C. Conclusion: This study applied DTI combined with TBSS method in children with RTT and typical development. Results revealed that corpus callosum dysplasia, pervasive brain connectivity deficit and subcortical-cortical circuit structure and connection abnormalities could be found in RTT children.

Key words: Rett syndrome, Diffusion tensor imaging, Tract-based spatial statistics, White matter, Children