Temporal And Spectral Patterns Of Resting State Brain Networks In Magnetoencephalography For Depression And Related Clinical Evaluations

Depression is a common mental disorder characterized by significant and persistent low mood,and in extreme cases,could lead to suicide.Over the past years,functional imaging studies have confirmed that depression is significantly associated with dysfunction in brain regions related to emotional cognition,and is closely associated with dysfunction of large-scale brain networks.Resting state brain networks have been proved to have rich spatial and temporal dynamics and multi-frequency oscillation information,but the abnormalities of time-frequency neuronal patterns in emotional processing and cognitive function in depression have not been fully explored.Suicide is considered to be a devastating consequence of depression,and its mechanisms and clinical evaluation still have great challenges to yet.The main purpose of this paper is to use the characteristics of magnetoencephalography(MEG)with high spatial and temporal resolution and multi-frequency signals,and combine with a variety of machine learning algorithms to study the time-frequency pattern changes of the brain functional network in patients with depression.The research aims are to explore the neural mechanism related to depressed disorder and clinical suicide from multiple perspectives,and to provide a variety of neuroimaging markers to assist the clinical diagnosis.The main work contents are as follows.1.The past findings from extensive researches on the depression of the default mode network(DMN)are inconsistency.We studied in the 25 depression patients and 25 healthy controls by the resting state MEG data,and used the dynamic functional connectivity within DMN followed by a clustering algorithm to divided the DMN activity in the alpha band(8-13Hz)to the sub-second functional connectivity microstates.In the power dominant state,depressed patients showed a transient decreased pattern that reflected inter/intra-subnetwork deregulation.A supplementary negatively correlated state simultaneously presented with increased connectivity between the ventromedial prefrontal cortex and the posterior cingulate cortex.The function compensation mechanism between the supplementary microstate and the major functional microstates is also proposed.The dynamic functional network transformation may be the reason for the different results of the static research.2.Since the neural mechanism of depression and suicide are still unknown in the spatiotemporal dynamics,we studied two groups of depressed patients who were divided into two groups according to their history of suicide attempt(34 suicide attempted and 44non-suicide patients).The hidden Markov Model(HMM)was used to decompose the information of the whole brain source-space MEG sequence both in temporal and frequency scale,and to obtain the transient activated HMM states at the sub-second level.Resting-state activity of suicide attempters switches more frequently into the fronto-temporal network,which is reflected by the time spent occupancy of fronto-temporal network is increased and interval time is decreased.Moreover,these changes are significantly correlated with the clinically assessed suicide risks.Suicide attempters also exhibit increased state-wise activations in the theta band(4-8Hz)in default mode network.The disorder of dynamic allocation of fronto-temporal network may increase the neuronal pain and disturb the cognitive control in depressed patients,and the abnormal dynamic activation of DMN also intensified the suicide attempt tendency.The perspective of time-frequency pattern provides a dynamic neural mechanisms and selective predictor for suicide attempts in depression.3.To explore the electrophysiological mechanisms underlying suicide attempt in depression,we conducted the research from the regional neuronal oscillation to the integrated neural network,used a novel fitting oscillation algorithm to remove neuronal background noise.The true power changes of brain regions in the oscillatary frequency band and the control centrality of the region to involved network were calculated in the same dataset of the above study.In suicide attempters,the beta-band(13-30Hz)activation power of dorsal medial prefrontal lobe is enhanced in a specific frequency band,and positively correlated with clinically assessed suicide risk scores,but its control ability of the central executive network is significantly decreased.Over-localization of the core node explains the disruption of neural circuit that support cognitive control,leads to increased suicidal tendency in patients with depression,and provides a potential electrophysiological indicator for clinical prediction of suicide.4.Based on the assumption that the development of suicidal thoughts and transformation into suicide attempts are based on different neural mechanisms,we used a semi-supervised learning framework by comparing the depressed patients,14 without the risk of suicide and 12 with suicide attempts and high risk,to extract suicide related characteristics and eliminate the disease effects.Then the greater amounts of 37 patients with suicide ideations were clustered to get the assessments of suicide risk level.The feature distance distribution in the suicide spectrum is positively correlated with the clinically assessed suicide risk scores.The following study on the alpha-gamma(30-48Hz)coupling of cortical-subcortical loop changing with the increased risk of suicide across the suicide spectrum reveals that,the development of suicide ideation for patients with depression to suicide attempt is the transformation from the abnormalities in the orbitofrontal area dominated ventral cognitive loop to the dorsolateral prefrontal region dominated dorsal cognitive loop.The suicide spectrum at different risk stages is quite meaningful to the understanding and assessment of suicide in clinical.In conclusion,based on the signal characteristics of resting state MEG,this paper explores the time-frequency patterns of brain networks from multiple perspectives,with findings to providing help for understanding the neural mechanism of depression and objective neural markers of suicide in clinical.