Multidimensional Brain Connectome Signatures In Majore Depressive Disorder

Major depressive disorder(MDD)is a common mental disorder that severely affects psychosocial functioning and reduces the quality of life.MDD is characterized with high incidence and d treatment-refractory,causing heavy burden to the family and society.Despite decades of research in basic science,clinical neuroscience,and psychiatry,the pathophysiology of MDD is still not well understood.Magnetic resonance imaging(MRI)is a safe and non-invasive technique for observing the brain structures and function of neuroanatomical brain regions.Blood-oxygenation-level-dependent functional MRI(BOLD-fMRI),as the most commonly used fMRI technique,can be used to study the flow of blood through the brain and oxygen uptake in resting state/task-induced conditions,thus broadening our understanding of how the brain works in healthy and diseased brain.Although BOLD-fMRI technique has continued to gain popularity in scientific research and clinical trials in recent years,the technique's full potential remains untapped.Because most of the previous studies did not consider the detection of neural activities in the brain white-matter(WM)tissue,the studies on the brain function and the expression of abnormal brain function in diseases only focused on the gray matter(GM)organization.In previous studies,BOLD-fMRI signals in WM were often considered as physiological noises to be eliminated,or using GM mask to remove the functional activationof WM,or selectively reporting the functional activiation in GM.The nature,interpretation,and correlation of BOLD-fMRI signals in WM as potential indicators of brain function are still being explored and even controversial.This dissertation mainly used brain connectome method based on MRI data to explore abnormal patterns of MDD from multiple dimensions.The main contents include two parts as follows:The first part is to investigate the functional information in WM,which is the methodological part of this dissertation.This part constructed functional connectome in WM,and built the predictive model based on WM functional connectivity(WMFC)and linked WMFC to gene expression,providing a more comprehensive perspective for the exploration of functional abnormalities in MDD.Firstly,this dissertation constructed WM functional connectome using graph theory analysis on healthy longitudinal BOLD-fMRI data,to explore topological properties in WM functional connectome.WM functional connectome exhibited stable and reliable small-worldness and non-randome modular parttern.These topological properties of WM functional connectome were not affected by confounding factors,such as head motion,spatial distance,global signal,cerebral spinal signal(CSF),node parcellation,hemodynamic response function(HRF),and threshold selection.In addition,WM functional connectome exhibited decreased small-worldness and global efficiency,and increased network strength and local efficiency compared to GM functional connectome,suggesting that WM functional connectome showed a tendency to random network.This study indicates that functional information in WM is not noise,and WM functional connectome also shows small-worldness with high efficiency and low cost in information transports.Secondly,to explore the relationship between WM functional connectome and intelligence,predictive model for general fluid intelligence(Gf)was contructed based on WMFC from healthy longitudinal BOLD-fMRI data and a completely independent data.For internal validation,WM functional connectivity on time 1 data was used as train data,and time 1,time 2,and time 3 data were used as test data,respectively.The results exhibited that no matter what the test data was used,the constructed predictive model could predict individual's Gf.In addition,the predictive performance was not affected by confounding factors,including head motion,feature selection and WM mask.For external validation,the predictive model based on WMFC from time 1 could predict individual's Gf on a completely independent data.The consensue features were mainly located on superior longitudinal fascicus,deep frontal WM and ventral frontal-parietal system.This dissertation suggests that WM functional connectome is a novel neuromarker for predictiong individual's Gf,and complement functional information to explore the relationship of brain-behavior.Finally,this dissertation further developed intersubject variability of WMFC mothod based on healthy longitudinal BOLD-fMRI data,miscroscopic gene expression data,and mesoscale MRI data,to explore the spatial distribution of intersubject variability and its transcriptomic mechanism.The spatial distribution of human brain WMFC exhibited nonrandom pattern,with high variability in higher-order cognitive networks and low variability in lower-order perceptural networks.Using Allen trascriptomic data,neuronal cells-related genes were mainly overexpressed in regions with high intersubject variability,enriched in synapse-related terms and glutamic pathways,and related to most psychiatric diseases,such as bipolar depression,major depressive disorder(MDD),autism and schizophrenia.Howver,glial cells-related genes were primarily overexpressed in regions with low intersubject variability,enriched in glial-related terms,and related to most neurodegenerative disorders.More importantly,intersubject variability of WMFC and GM functional connectivity(GMFC)exhibited specific and common enchirment pathways,suggesting the the specificity and complemental and reverberating functional information between WM and GM.In addition,intersubject variability of WMFC was related to mesoscale functional variability(i.e.,cerebral blood fluid)and structural variability(i.e.,myelin,WM volum,and fractional anistropy).This dissertation highlights the potential significanc of combing intersubject variaibity of WMFC and GMFC for understanding brain evolution and development and guiding disease interventions.The second part is to investigate abnormal patterns of MDD from multiple dimensions.This part explored abnormal topological properties of WMFC in MDD,and linked abnormal brain network of MDD to gene expression data,revealing the associations between brain organizations of different levels in MDD.Firstly,employing WM functional connectome,this dissertation explored the abnormal topological properties in MDD,and built predictive and classified models.Unmedicated MDD patients exhibited decreased normalized clustering coefficient and small-worldness compred to healthy controls,suggesting that WM functional connectome in patients with MDD showed tendenty to random networks.Based on topological propertis(i.e.,normalized clustering coefficient,normalized shortest path length,smallworldness)of WM functional connectome,the constructed predictive model could be used to predict depressive degree,and the constructed classification model could be used to distinguish patients with MDD from healthy controls.These results indicated the potential clinical application of topological propertis in WM functional connectome,and provided a novol neuromarker for exploring biological mechanisms of MDD.Secondly,this dissertation linked abnormal pattern of morphometric similarity network pattern in MDD to gene expression data.Compared with healthy controls,patients with MDD showed replicable morphometric changes across two independent samples.Using Allen trascriptomic data,the genes that were correlated with morphometric differences of MDD were encriched with synapse-related pathways.In addition,microglial and neuronal cells account for most of the observed correlation with MDD-related morphometric differences.These results suggested that the combination of brain network information from multiple dimensions may provide new clues for exploring the biological causes of depression.In conclusion,this dissertation develops WMFC techinique and constructs classification and prediction models based on WM functional connectome using BOLDfMRI data.In addition,this dissertation explores functional information and clinical applications in WM and explores abnormal brain networks in MDD from micromacroscales,enriching the perspective of understanding brain function and providing a comprehensive insight(i.e.,WM functional neuromarker)for exploration of MDD.