Prediction Of Long-term After-effects Of Theta-burst Stimulation On Supplementary Motor Network Through One-session Response

Background:To understand the neural mechanism of repetitive transcranial magnetic stimulation(rTMS),the after-effects following one session or multiple days of stimulation have been widely investigated.However,the relation between the short-term effect(STE)and long-term effect(LTE)of rTMS is largely unknown.Objective:This study aims to explore whether the after-effects of five-days rTMS on supplementary motor area(SMA)network could be predicted by one-session response.Method:A primary cohort of 38 healthy participants underwent five daily sessions of real or sham continuous theta-burst stimulation(c TBS)on the left SMA.Resting-state functional magnetic resonance imaging(rs-f MRI)data were acquired at the first(before and after the first stimulation)and sixth experimental day.The SMA connectivity changes after the first c TBS and after five days of stimulation were defined as STE and LTE,respectively.The predicting value of STE for LTE was estimated by leave-one-out cross validation.Briefly,we sequentially selected one subject as a test,and the others as training data.In the training data,resting-state functional connectivity changes of the target(i.e.,STE)was estimated by comparing the post-and pre-rTMS conditions.Voxels with significant STE(Pcorr < 0.05)was defined as Region-of-interest(ROI).RSFC alterations in these ROIs were correlated between short-term and long-term conditions.Based on the information of the voxel with the highest correlation coefficient,a linear function between STE and LTE could be established.Then,predicted LTE of the test subject could be computed through the function and STE.Finally,Pearson’s correlation was performed between the real and predicted LTE across subjects.Results:Compared to the baseline,significant STE and LTE were found in the paracentral gyrus after real stimulation,suggesting shared neural correlates of short-and long-term stimulations.Region-of-interest(ROI)analysis indicated that the resting-state functional connectivity between SMA and paracentral gyrus increased after real stimulation,while no significant changes was found after sham stimulation.Leave-one-out cross validation(LOOCV)indicated that the LTE in could be predicted by the STE after real but not sham stimulations.In an independent cohort,the after-effects of rTMS on paracentral gyrus and short to long-term prediction was reproduced at the ROI level.The second experiment was designed to reproduce the findings of the primary experiment.All participants(n=16)received real c TBS for consecutive five days.Imaging data were acquired as results for the primary experiment.Conclusion:These imaging evidences indicate that one-session rTMS can aid to predict the regions responsive to long-term stimulation and the individualized response degree.