Altered Dynamic Interactions Of Resting State Brain Network In Depression And Associated Clinical Effect By Antidepressants

Abnormal interactions within and between resting state networks(RSNs),is one of the characteristics of major depressive disorder(MDD).The development of functional neuroimaging technology provides an opportunity to reveal the neural basis of brain dysfunction in patients with MDD.Most importantly,RSNs is a promising predictor of treatment response in MDD.In recent years,more researchers have explored the RSNs’ interactions from the perspective of temporal dynamics.The neuropathological mechanisms related to diagnosis and treatment of MDD are investigated in a more comprehensive way than traditional static analysis and the abnormal RSNs’ dynamic interactions in MDD patients are confirmed.However,there are still some challenges.(1)Though the abnormal RSNs’ dynamic interactions in MDD patients are compared with healthy controls,few studies have examined the relationship between the dynamic interactions and antidepressant treatment outcome.The individualized prediction of the response to antidepressant treatment by analyzing the abnormal RSNs’ dynamic interactions in MDD at baseline should be clarified.(2)With the remission of MDD patients,it remains to be seen whether the abnormal dynamic interactions are recovered.The correlation between the recovery of different brain networks’ abnormalities and the reduced severity of multidimensional symptoms has not been explored.In this paper,functional magnetic resonance imaging(f MRI)and magnetoencephalography(MEG)are used to investigate brain networks’ communications in MDD patients from a dynamic perspective,so as to explore the potential neuropathological mechanism of MDD and the neural circuit basis of antidepressant efficacy.The main contents are as follows:1.Traditional static analysis on RSNs and prediction of early antidepressant response in MDD patients based on MEG data.Brain networks involved in depression are selected as the regions of interest,including the default mode network(DMN),central executive network(CEN),salience network(SN),subcortical,motor areas and visual network.We analyze the neural power spectrum density(PSD)after removing the non-periodic background signals,and then explore the abnormal changes of traditional static functional connections among RSNs at specific frequency bands in MDD patients.Compared with controls,MDD patients show significantly decreased beta power(t= 3.22,p< 0.001),which is related to baseline depression severity(r=-0.46,p< 0.05).A significant increase(p< 0.001)in beta functional connectivity is associated with 2-week reductions in core six items of the Hamilton rating scale for depression(HRSD)(r= 0.47,p< 0.05).The results of receiver operating characteristic curve(ROC)indicate that abnormalities of RSNs at baseline in MDD patients can predict early antidepressant responses(area under curve= 0.773,p< 0.05).2.The RSNs’ dynamic interactions in MDD patients based on MEG data.Based on the traditional static analysis of RSNs,we continue to analyze the abnormal transient dynamic interactions of DMN,CEN and SN in MDD patients.Both MDD patients and healthy controls show dynamic modular characteristics of RSNs,but MDD patients have fewer modules(p< 0.05).The internal bindings within the CEN and DMN in MDD patients are enhanced(p< 0.05),and SN is involved in the stronger external integrations between networks(p< 0.01).3.Prediction of antidepressant response to escitalopram in MDD patients based on f MRI data.Based on the discovery of dynamic interaction abnormalities in the brain network in MDD patients,we explore the dynamic interaction abnormalities in RSNs to predict antidepressant response.We enroll first-episode drug-na(?)ve patients from three centers.Responders are defined as with a HRSD-17 reduction greater than 50% after treatment with escitalopram.The RSNs’ flexibility is an accurate predictor of antidepressant efficacy at baseline(accuracy=79.41%),which is independently validated at three centers(accuracy:79.41%,94.44%,82.86%).The anterior cingulum cortex(ACC)cored default subnetwork constitutes the main contributing brain regions of the predictive model.Modular allegiance(MA)of ACC is higher in the antidepressant responders than non-responders at baseline(t=3.045,p< 0.01).Although the drug treatment target is not ACC,higher activity in this region can benefit MDD patients and make them more likely to respond to antidepressant treatment.4.Normalization of RSNs in remitted MDD patients based on f MRI data.We aim to answer,“are large-scale brain network dynamic interaction abnormalities in remitted patients normalized after treatment?” We analyzed f MRI data at baseline and after treatment.A two-sample paired test compares the dynamic interaction indexes of brain networks before and after treatment.Following antidepressant treatment,the reduced network flexibility is significantly normalized with the remission of depressive symptoms(p< 0.001).The recovery of whole brain flexibility is related to symptom factors,including somatic anxiety(r=-0.75,p< 0.001),cognition(r=-0.62,p< 0.001),hysteresis(r=-0.74,p< 0.001)and dyssomnia(r=-0.59,p< 0.001).Interestingly,the flexibility of cognitive control network(CCN)is selectively related to somatic anxiety factor(r=-0.37,p< 0.05),while the flexibility of DMN is related to hysteresis factor(r=-0.36,p< 0.05).These findings indicate that the antidepressants-induced normalizations of different hypoflexible brain networks are related to specific depressive symptoms.