Objective To investigate the relationship between MRI signal,pathological changes and neurological function after sciatic nerve injury in rabbits.Methods Twenty New Zealand white rabbits were randomly and evenly divided into 5 groups,and the right sciatic nerve crush models were established.T2 fat suppression fast recovery spin echo (T2 fs FRFSE) sequence scanning was performed 3 days,7 days,2 weeks,3 weeks and 4 weeks after injury,and TE was set as 30,60 and 90 ms,respectively.Signal intensity ratio (SIR) and relative signal intensity (△S) of proximal and distal part of injured nerve and control side nerve were measured.The relationship between SIR,△S,pathology and rabbit lower limb nerve function were analyzed.Results In the distal part of injured nerve,SIR and △S increased 3-7 days after injury,pathological results showed vacuolar degeneration,and basic toe function lost was found.SIR and △S reached the peak 2 weeks after injury,with most serious disintegration of myelin and toe function disable.SIR,△S and toe function disable gradually recovered,and the nerve regenerated at 3-4 weeks after injury.The injure display rate of T2 fs FRFSE images with TE=90 and 60 ms,SIR of both distal and proximal part of injured nerve were higher than those on images with TE=30 ms (all P＜0.05).Conclusion SIR and △S changes on T2 fs FRFSE imaging can be used to predict rabbit nerve injury.

Objective To develop a deep learning network based on attention mechanism to identify the number of the vertical man-dibular third molar(MTM)roots(single or double)on panoramic radiographs in an automatic way.Methods The sample consisted of 1 045 patients with 1 642 MTMs on paired panoramic radiographs and Cone-beam computed tomography(CBCT)and were randomly grouped into the training(80%),the validation(10%),and the test(10%).The evaluation of CBCT was defined as the ground truth.A deep learning network based on attention mechanism,which was named as RN-MTMnet,was trained to judge if the MTM on pano-ramic radiographs had one or two roots.Diagnostic performance was evaluated by accuracy,sensitivity,specificity,and positive predict value(PPV),and the receiver operating characteristic(ROC)curve with the area under the ROC curve(AUC).Its diagnostic perform-ance was compared with dentists'diagnosis,Faster-RCNN,CenterNet,and SSD using evaluation metrics.Results On CBCT images,single-rooted MTM was observed on 336(20.46%)sides,while two-rooted MTM was 1 306(79.54%).The RN-MTMnet achieved an accuracy of 0.888,a sensitivity of 0.885,a specificity of 0.903,a PPV of 0.976,and the AUC value of 0.90.Conclusion RN-MTM-net is developed as a novel,robust and accurate method for detecting the numberof MTM roots on panoramic radiographs.