Temporal-Adaptive Graph Convolutional Network for Automated Identification of Major Depressive Disorder Using Resting-State fMRI

Altered brain functional connectivity and correlation with psychological status in patients with unilateral pulsatile tinnitus

Developmental malformations (dysgenesis) of the corpus callosum lead to neurological conditions with a broad range of clinical presentations. Investigating the altered brain connectivity patterns is crucial to understanding both adaptive and maladaptive neuroplasticity in corpus callosum dysgenesis patients. Here, we acquired structural diffusion-weighted and resting-state functional MRI data from a cohort of 11 corpus callosum dysgenesis patients (five with agenesis and six with hypoplasia) and compared their structural and functional connectivity patterns to healthy subjects selected from the Human Connectome Project. We found that these patients have fewer structural inter- and intra-hemispheric brain connections relative to healthy controls. Interestingly, the patients with callosal agenesis have a scant number of inter-hemispheric connections but manage to maintain the full integrity of functional connectivity between the same cortical regions as the healthy subjects. On the other hand, the hypoplasic group presented abnormal structural and functional connectivity patterns relative to healthy controls while maintaining the same total amount of functional connections. These results demonstrate that acallosal patients can compensate for having fewer structural brain connections and present functional adaptation. However, hypoplasics present atypical structural connections to different brain regions, leading to entirely new and abnormal functional brain connectivity patterns.

The promise of machine learning has fueled the hope for developing diagnostic tools for psychiatry. Initial studies showed high accuracy for the identification of major depressive disorder (MDD) with resting-state connectivity, but progress has been hampered by the absence of large datasets. Here we used regular machine learning and advanced deep learning algorithms to differentiate patients with MDD from healthy controls and identify neurophysiological signatures of depression in two of the largest resting-state datasets for MDD. We obtained resting-state functional magnetic resonance imaging data from the REST-meta-MDD (N = 2338) and PsyMRI (N = 1039) consortia. Classification of functional connectivity matrices was done using support vector machines (SVM) and graph convolutional neural networks (GCN), and performance was evaluated using 5-fold cross-validation. Features were visualized using GCN-Explainer, an ablation study and univariate t-testing. The results showed a mean classification accuracy of 61% for MDD versus controls. Mean accuracy for classifying (non-)medicated subgroups was 62%. Sex classification accuracy was substantially better across datasets (73–81%). Visualization of the results showed that classifications were driven by stronger thalamic connections in both datasets, while nearly all other connections were weaker with small univariate effect sizes. These results suggest that whole brain resting-state connectivity is a reliable though poor biomarker for MDD, presumably due to disease heterogeneity as further supported by the higher accuracy for sex classification using the same methods. Deep learning revealed thalamic hyperconnectivity as a prominent neurophysiological signature of depression in both multicenter studies, which may guide the development of biomarkers in future studies.

The whole-brain functional connectivity (FC) pattern obtained from resting-state functional magnetic resonance imaging data are commonly applied to study neuropsychiatric conditions such as autism spectrum disorder (ASD) by using different machine learning models. Recent studies indicate that both hyper- and hypo- aberrant ASD-associated FCs were widely distributed throughout the entire brain rather than only in some specific brain regions. Deep neural networks (DNN) with multiple hidden layers have shown the ability to systematically extract lower-to-higher level information from high dimensional data across a series of neural hidden layers, significantly improving classification accuracy for such data. In this study, a DNN with a novel feature selection method (DNN-FS) is developed for the high dimensional whole-brain resting-state FC pattern classification of ASD patients vs. typical development (TD) controls. The feature selection method is able to help the DNN generate low dimensional high-quality representations of the whole-brain FC patterns by selecting features with high discriminating power from multiple trained sparse auto-encoders. For the comparison, a DNN without the feature selection method (DNN-woFS) is developed, and both of them are tested with different architectures (i.e., with different numbers of hidden layers/nodes). Results show that the best classification accuracy of 86.36% is generated by the DNN-FS approach with 3 hidden layers and 150 hidden nodes (3/150). Remarkably, DNN-FS outperforms DNN-woFS for all architectures studied. The most significant accuracy improvement was 9.09% with the 3/150 architecture. The method also outperforms other feature selection methods, e.g., two sample t-test and elastic net. In addition to improving the classification accuracy, a Fisher's score-based biomarker identification method based on the DNN is also developed, and used to identify 32 FCs related to ASD. These FCs come from or cross different pre-defined brain networks including the default-mode, cingulo-opercular, frontal-parietal, and cerebellum. Thirteen of them are statically significant between ASD and TD groups (two sample t-test p < 0.05) while 19 of them are not. The relationship between the statically significant FCs and the corresponding ASD behavior symptoms is discussed based on the literature and clinician's expert knowledge. Meanwhile, the potential reason of obtaining 19 FCs which are not statistically significant is also provided.

BackgroundSustained attention deficits were reported more significant in patients with major depressive disorder (MDD) than in healthy controls (HCs), and are pivotal in both the development and aggravation of depression. Childhood trauma is also common in MDD and the exposure to childhood trauma may impede sustained attention and increase the treatment resistance in MDD. However, the underlying neuro-mechanisms link the childhood trauma to sustained attention deficits in MDD remain unclear.MethodsWe collected resting-state functional magnetic resonance imaging data, and measured childhood trauma severity using the Childhood Trauma Questionnaire and sustained attention using the Continuous Performance Test, Identical Pairs version. After excluding subjects with significant head movement, 45 MDDs and 54 HCs were included in the analysis. We compared fractional amplitude of low-frequency fluctuation (fALFF) between the groups, conducted whole-brain correlation analysis between the fALFF and sustained attention in the MDD group, and defined significant regions as the regions of interest for the seed-to-whole brain functional connectivity (FC) analysis. We further performed mediation analyses to investigate the relationships among the childhood trauma, fALFF and FC values, and the level of sustained attention in the MDD group.ResultsCompared with HCs, MDDs exhibited higher fALFF in the right middle frontal gyrus and left inferior frontal gyrus, and lower fALFF in the bilateral insular cortex, left medial orbital superior frontal gyrus and left angular gyrus (ANG.L). Whole-brain correlation analysis showed that impaired sustained attention was associated with increased fALFF in the left postcentral gyrus (PoCG.L), and FC of PoCG.L-left precentral gyrus (PreCG.L) and ANG.L-right superior temporal gyrus (STG.R) in the MDD group. Furthermore, mediation analyses showed that the fALFF in PoCG.L, and FC of PoCG.L-PreCG.L and ANG.L-STG.R mediated the relationship between the childhood trauma and sustained attention in the MDD group.ConclusionThe fALFF in PoCG.L, and FC of PoCG.L-PreCG.L and ANG.L-STG.R might be potential neural substrate in the association between the childhood trauma and poor sustained attention in the MDDs, and might serve as potential intervention targets for the treatment of sustained attention deficits in MDDs with childhood trauma history.

The human entorhinal-hippocampal circuitry is characterized by an information-specific functional organization where two routes, that are preferentially connected to the parahippocampal cortex or the perirhinal and retrosplenial cortices, divide the entorhinal cortex as well as hippocampal subiculum and CA1 subregions.

Altered brain entropy and functional connectivity patterns in generalized anxiety disorder patients

While there’s been clinical success and growing research interest in hypnosis, neurobiological underpinnings induced by hypnosis remain unclear. In this fMRI study (which is part of a larger hypnosis project) with 50 hypnosis-experienced participants, we analyzed neural and physiological responses during two hypnosis states, comparing them to non-hypnotic control conditions and to each other. An unbiased whole-brain analysis (multi-voxel- pattern analysis, MVPA), pinpointed key neural hubs in parieto-occipital-temporal areas, cuneal/precuneal and occipital cortices, lingual gyri, and the occipital pole. Comparing directly both hypnotic states revealed depth-dependent connectivity changes, notably in left superior temporal/supramarginal gyri, cuneus, planum temporale, and lingual gyri. Multi-voxel- pattern analysis (MVPA) based seeds were implemented in a seed-to-voxel analysis unveiling region-specific increases and decreases in functional connectivity patterns. Physiologically, the respiration rate significantly slowed during hypnosis. Summarized, these findings foster fresh insights into hypnosis-induced functional connectivity changes and illuminate further knowledge related with the neurobiology of altered consciousness.

Objective: To study resting-state functional connectivity (FC) of default mode network (DMN) in adolescent patients with first-episode drug-naive major depressive disorder (MDD). Methods: We enrolled thirty first-episode and drug-naive adolescent MDD patients and twenty-nine adolescent healthy control (HC) participants in the First Affiliated Hospital of Zhengzhou University. There were no differences in age, sex, and education between the MDD and HC group. Resting-state functional magnetic resonance images (fMRI) was performed. We selected posterior cingulate cortex (PCC) and medial prefrontal cortex (MPFC) of DMN as regions of interests (ROI). The differences of these regions from the whole brain functional connectivity were analyzed. The relations between abnormalities in FCs of DMN and clinical variables were further investigated. Results: Compared to the HCs, the MDD patients had congruently reduced FCs between the PCC and cerebellum, temporal cortices, occipital cortices, fusiform, dorsolateral prefrontal cortex. MPFC not only had reduced FCs with fusiform, temporal cortices, anterior cingulate cortex, but also had enhanced FCs with occipital cortices, parietal cortices, and precentral gyrus. In addition, the increased FC between the right MPFC and right precentral gyrus was positive correlated with Hamilton Rating Scale for Depression (HAMD) scores (r=0.38, P=0.04). The reduced FC between the left middle temporal gyrus and left PCC as well as the enhanced FC between the right middle cingulum and right MPFC were positive correlated with the duration of depression since onset (r=0.39, P=0.03; r=0.38, P=0.04). Conclusions: These findings show dysfunctional DMN connectivity of adolescent MDD patients. Neurodevelopmental abnormalities in DMN may present in adolescent MDD.

Impaired functional default mode network in patients with mild neurological Wilson’s disease

The resting-state cerebro-cerebellar function connectivity and associations with verbal working memory performance

Linked functional network abnormalities during intrinsic and extrinsic activity in schizophrenia as revealed by a data-fusion approach

As the main input hub of the basal ganglia, the striatum receives projections from the cerebral cortex. Many studies have provided evidence for multiple parallel corticostriatal loops based on the structural and functional connectivity profiles of the human striatum. A recent resting-state fMRI study revealed the topography of striatum by assigning each voxel in the striatum to its most strongly correlated cortical network among the cognitive, affective, and motor networks. However, it remains unclear what patterns of striatal parcellation would result from performing the clustering without subsequent assignment to cortical networks. Thus, we applied unsupervised clustering algorithms to parcellate the human striatum based on its functional connectivity patterns to other brain regions without any anatomically or functionally defined cortical targets. Functional connectivity maps of striatal subdivisions, identified through clustering analyses, were also computed. Our findings were consistent with recent accounts of the functional distinctions of the striatum as well as with recent studies about its functional and anatomical connectivity. For example, we found functional connections between dorsal and ventral striatal clusters and the areas involved in cognitive and affective processes, respectively, and between rostral and caudal putamen clusters and the areas involved in cognitive and motor processes, respectively. This study confirms prior findings, showing similar striatal parcellation patterns between the present and prior studies. Given such striking similarity, it is suggested that striatal subregions are functionally linked to cortical networks involving specific functions rather than discrete portions of cortical regions. Our findings also demonstrate that the clustering of functional connectivity patterns is a reliable feature in parcellating the striatum into anatomically and functionally meaningful subdivisions. The striatal subdivisions identified here may have important implications for understanding the relationship between corticostriatal dysfunction and various neurodegenerative and psychiatric disorders.

Several functional magnetic resonance imaging (fMRI) investigations of patients with vestibular migraine (VM) have revealed abnormal functionality in different networks, indicating that VM is related to alterations in brain function. We sought to investigate the resting-state functional connectivity (FC) patterns during the interictal period in VM by combining data-driven voxel-wise degree centrality (DC) calculations and seed-based FC analyses, and thereby determine the associations between cerebral function and clinical symptoms. Thirty-eight patients with VM and 33 matched normal controls were recruited. DC was calculated and compared between the groups, and the FC of locations showing DC alterations was further tested using a seed-based technique. The participants' clinical indicators were correlated with the DC and FC values of the brain areas. In contrast to the control group, the VM group showed considerably lower DC values in the bilateral medial prefrontal cortex (mPFC) and significantly higher DC values in the right occipital lobe. In the seed-based FC analyses, patients with VM demonstrated fewer connections of the bilateral mPFC with the bilateral posterior cingulate cortex, right parahippocampus, right cerebellar posterior lobe, bilateral cuneus, and left precuneus. In addition, clinical data from patients, such as pain intensity, episode frequency, and the Dizziness Handicap Inventory score, were negatively related to these FC and DC impairments. Our findings showed changes in the default mode network and visual cortex in patients with VM, providing further insights into the complexity of the mechanisms underlying VM.

Brain functional connectivity (FC) changes and microstructural abnormalities are reported in infants born moderate and late preterm (MLPT). We evaluated the effect of low-grade (grades I, II) intraventricular hemorrhage (IVH) in MLPT babies on brain structural connectivity (SC) and FC. Babies born MLPT between January 2014 and May 2017 underwent brain ultrasound (US) at 72h and 7days after birth, and MRI at around term equivalent. The MRI protocol comprised T1- and T2-weighted sequences, diffusion tensor imaging (DTI), and resting-state functional MRI (fMRI). SC and FC were assessed using graph analysis. Of 350 MLPT neonates, 15 showed low-grade IVH on US at 72h, for which brain MRI was available in 10. These 10 infants, with mean gestational age (GA) 34.0 ± 0.8weeks, comprised the study group, and 10 MLPT infants of mean GA 33.9 ± 1.1weeks, with no abnormalities on brain US and MRI, were control subjects. All study subjects presented modularity, small world topology, and rich club organization for both SC and FC. The patients with low-grade IVH had lower FC rich club coefficient and lower SC betweenness centrality in the left frontoparietal operculum, and lower SC rich club coefficient in the right superior orbitofrontal cortex than the control subjects. Topological and functional properties of mature brain connectivity are present in MLPT infants. IVH in these infants was associated with structural and functional abnormalities in the left frontoparietal operculum and right orbitofrontal cortex, regions related to language and cognition.