DS-GCNs: Connectome Classification using Dynamic Spectral Graph Convolution Networks with Assistant Task Training.

Functional connectivity (FC) matrices measure the regional interactions in the brain and have been widely used in neurological brain disease classification. A brain network, also named as connectome, could form a graph structure naturally, the nodes of which are brain regions and the edges are interregional connectivity. Thus, in this study, we proposed novel graph convolutional networks (GCNs) to extract efficient disease-related features from FC matrices. Considering the time-dependent nature of brain activity, we computed dynamic FC matrices with sliding windows and implemented a graph convolution-based LSTM (long short-term memory) layer to process dynamic graphs. Moreover, the demographics of patients were also used as additional outputs to guide the classification. In this paper, we proposed to utilize the demographic information as extra outputs and to share parameters among three networks predicting subject status, gender, and age, which serve as assistant tasks. We tested the performance of the proposed architecture in ADNI II dataset to classify Alzheimer's disease patients from normal controls. The classification accuracy, sensitivity, and specificity reach 90.0%, 91.7%, and 88.6%, respectively, on ADNI II dataset.

Classification of schizophrenia by intersubject correlation in functional connectome.

Functional connectomes have been suggested as fingerprinting for individual identification. Accordingly, we hypothesized that subjects in the same phenotypic group have similar functional connectome features, which could help to discriminate schizophrenia (SCH) patients from healthy controls (HCs) and from depression patients. To this end, we included resting-state functional magnetic resonance imaging data of SCH, depression patients, and HCs from three centers. We first investigated the characteristics of connectome similarity between individuals, and found higher similarity between subjects belonging to the same group (i.e., SCH-SCH) than different groups (i.e., HC-SCH). These findings suggest that the average connectome within group (termed as group-specific functional connectome [GFC]) may help in individual classification. Consistently, significant accuracy (75-77%) and area under curve (81-86%) were found in discriminating SCH from HC or depression patients by GFC-based leave-one-out cross-validation. Cross-center classification further suggests a good generalizability of the GFC classification. We additionally included normal aging data (255 young and 242 old subjects with different scanning sequences) to show factors could be improved for better classification performance, and the findings emphasized the importance of increasing sample size but not temporal resolution during scanning. In conclusion, our findings suggest that the average functional connectome across subjects contained group-specific biological features and may be helpful in clinical diagnosis for schizophrenia.

Statistical testing and power analysis for brain-wide association study.

The identification of connexel-wise associations, which involves examining functional connectivities between pairwise voxels across the whole brain, is both statistically and computationally challenging. Although such a connexel-wise methodology has recently been adopted by brain-wide association studies (BWAS) to identify connectivity changes in several mental disorders, such as schizophrenia, autism and depression, the multiple correction and power analysis methods designed specifically for connexel-wise analysis are still lacking. Therefore, we herein report the development of a rigorous statistical framework for connexel-wise significance testing based on the Gaussian random field theory. It includes controlling the family-wise error rate (FWER) of multiple hypothesis testings using topological inference methods, and calculating power and sample size for a connexel-wise study. Our theoretical framework can control the false-positive rate accurately, as validated empirically using two resting-state fMRI datasets. Compared with Bonferroni correction and false discovery rate (FDR), it can reduce false-positive rate and increase statistical power by appropriately utilizing the spatial information of fMRI data. Importantly, our method bypasses the need of non-parametric permutation to correct for multiple comparison, thus, it can efficiently tackle large datasets with high resolution fMRI images. The utility of our method is shown in a case-control study. Our approach can identify altered functional connectivities in a major depression disorder dataset, whereas existing methods fail. A software package is available at https://github.com/weikanggong/BWAS.

A comprehensive analysis of resting state fMRI measures to classify individual patients with Alzheimer's disease.

Alzheimer's disease (AD) patients show altered patterns of functional connectivity (FC) on resting state functional magnetic resonance imaging (RSfMRI) scans. It is yet unclear which RSfMRI measures are most informative for the individual classification of AD patients. We investigated this using RSfMRI scans from 77 AD patients (MMSE = 20.4 ± 4.5) and 173 controls (MMSE = 27.5 ± 1.8). We calculated i) FC matrices between resting state components as obtained with independent component analysis (ICA), ii) the dynamics of these FC matrices using a sliding window approach, iii) the graph properties (e.g., connection degree, and clustering coefficient) of the FC matrices, and iv) we distinguished five FC states and administered how long each subject resided in each of these five states. Furthermore, for each voxel we calculated v) FC with 10 resting state networks using dual regression, vi) FC with the hippocampus, vii) eigenvector centrality, and viii) the amplitude of low frequency fluctuations (ALFF). These eight measures were used separately as predictors in an elastic net logistic regression, and combined in a group lasso logistic regression model. We calculated the area under the receiver operating characteristic curve plots (AUC) to determine classification performance. The AUC values ranged between 0.51 and 0.84 and the highest were found for the FC matrices (0.82), FC dynamics (0.84) and ALFF (0.82). The combination of all measures resulted in an AUC of 0.85. We show that it is possible to obtain moderate to good AD classification using RSfMRI scans. FC matrices, FC dynamics and ALFF are most discriminative and the combination of all the resting state measures improves classification accuracy slightly.

Spontaneous eyelid closures link vigilance fluctuation with fMRI dynamic connectivity states.

Fluctuations in resting-state functional connectivity occur but their behavioral significance remains unclear, largely because correlating behavioral state with dynamic functional connectivity states (DCS) engages probes that disrupt the very behavioral state we seek to observe. Observing spontaneous eyelid closures following sleep deprivation permits nonintrusive arousal monitoring. During periods of low arousal dominated by eyelid closures, sliding-window correlation analysis uncovered a DCS associated with reduced within-network functional connectivity of default mode and dorsal/ventral attention networks, as well as reduced anticorrelation between these networks. Conversely, during periods when participants' eyelids were wide open, a second DCS was associated with less decoupling between the visual network and higher-order cognitive networks that included dorsal/ventral attention and default mode networks. In subcortical structures, eyelid closures were associated with increased connectivity between the striatum and thalamus with the ventral attention network, and greater anticorrelation with the dorsal attention network. When applied to task-based fMRI data, these two DCS predicted interindividual differences in frequency of behavioral lapsing and intraindividual temporal fluctuations in response speed. These findings with participants who underwent a night of total sleep deprivation were replicated in an independent dataset involving partially sleep-deprived participants. Fluctuations in functional connectivity thus appear to be clearly associated with changes in arousal.

Ensemble Hierarchical High-Order Functional Connectivity Networks for MCI Classification.

Conventional functional connectivity (FC) and corresponding networks focus on characterizing the pairwise correlation between two brain regions, while the high-order FC (HOFC) and networks can model more complex relationship between two brain region "pairs" (i.e., four regions). It is eye-catching and promising for clinical applications by its irreplaceable function of providing unique and novel information for brain disease classification. Since the number of brain region pairs is very large, clustering is often used to reduce the scale of HOFC network. However, a single HOFC network, generated by a specific clustering parameter setting, may lose multifaceted, highly complementary information contained in other HOFC networks. To accurately and comprehensively characterize such complex HOFC towards better discriminability of brain diseases, in this paper, we propose a novel HOFC based disease diagnosis framework, which can hierarchically generate multiple HOFC networks and further ensemble them with a selective feature fusion method. Specifically, we create a multi-layer HOFC network construction strategy, where the networks in upper layers are formed by hierarchically clustering the nodes of the networks in lower layers. In such a way, information is passed from lower layers to upper layers by effectively removing the most redundant part of information and, at the same time, retaining the most unique part. Then, the retained information/features from all HOFC networks are fed into a selective feature fusion method, which combines sequential forward selection and sparse regression, to further select the most discriminative feature subset for classification. Experimental results confirm that our novel method outperforms all the single HOFC networks corresponding to any single parameter setting in diagnosis of mild cognitive impairment (MCI) subjects.

Three dysconnectivity patterns in treatment-resistant schizophrenia patients and their unaffected siblings.

UNLABELLED: Among individuals diagnosed with schizophrenia, approximately 20%-33% are recognized as treatment-resistant schizophrenia (TRS) patients. These TRS patients suffer more severely from the disease but struggle to benefit from existing antipsychotic treatments. A few recent studies suggested that schizophrenia may be caused by impaired synaptic plasticity that manifests as functional dysconnectivity in the brain, however, few of those studies focused on the functional connectivity changes in the brains of TRS groups. In this study, we compared the whole brain connectivity variations in TRS patients, their unaffected siblings, and healthy controls. Connectivity network features between and within the 116 automated anatomical labeling (AAL) brain regions were calculated and compared using maps created with three contrasts: patient vs. control, patient vs. sibling, and sibling vs. CONTROL: To evaluate the predictive power of the selected features, we performed a multivariate classification approach. We also evaluated the influence of six important clinical measures (e.g. age, education level) on the connectivity features. This study identified abnormal significant connectivity changes of three patterns in TRS patients and their unaffected siblings: 1) 69 patient-specific connectivity (PCN); 2) 102 shared connectivity (SCN); and 3) 457 unshared connectivity (UCN). While the first two patterns were widely reported by previous non-TRS specific studies, we were among the first to report widespread significant connectivity differences between TRS patient groups and their healthy sibling groups. Observations of this study may provide new insights for the understanding of the neurophysiological mechanisms of TRS.

Diagnostic classification of intrinsic functional connectivity highlights somatosensory, default mode, and visual regions in autism.

Despite consensus on the neurological nature of autism spectrum disorders (ASD), brain biomarkers remain unknown and diagnosis continues to be based on behavioral criteria. Growing evidence suggests that brain abnormalities in ASD occur at the level of interconnected networks; however, previous attempts using functional connectivity data for diagnostic classification have reached only moderate accuracy. We selected 252 low-motion resting-state functional MRI (rs-fMRI) scans from the Autism Brain Imaging Data Exchange (ABIDE) including typically developing (TD) and ASD participants (n = 126 each), matched for age, non-verbal IQ, and head motion. A matrix of functional connectivities between 220 functionally defined regions of interest was used for diagnostic classification, implementing several machine learning tools. While support vector machines in combination with particle swarm optimization and recursive feature elimination performed modestly (with accuracies for validation datasets <70%), diagnostic classification reached a high accuracy of 91% with random forest (RF), a nonparametric ensemble learning method. Among the 100 most informative features (connectivities), for which this peak accuracy was achieved, participation of somatosensory, default mode, visual, and subcortical regions stood out. Whereas some of these findings were expected, given previous findings of default mode abnormalities and atypical visual functioning in ASD, the prominent role of somatosensory regions was remarkable. The finding of peak accuracy for 100 interregional functional connectivities further suggests that brain biomarkers of ASD may be regionally complex and distributed, rather than localized.

Salience network-based classification and prediction of symptom severity in children with autism.

IMPORTANCE: Autism spectrum disorder (ASD) affects 1 in 88 children and is characterized by a complex phenotype, including social, communicative, and sensorimotor deficits. Autism spectrum disorder has been linked with atypical connectivity across multiple brain systems, yet the nature of these differences in young children with the disorder is not well understood. OBJECTIVES: To examine connectivity of large-scale brain networks and determine whether specific networks can distinguish children with ASD from typically developing (TD) children and predict symptom severity in children with ASD. DESIGN, SETTING, AND PARTICIPANTS: Case-control study performed at Stanford University School of Medicine of 20 children 7 to 12 years old with ASD and 20 age-, sex-, and IQ-matched TD children. MAIN OUTCOMES AND MEASURES: Between-group differences in intrinsic functional connectivity of large-scale brain networks, performance of a classifier built to discriminate children with ASD from TD children based on specific brain networks, and correlations between brain networks and core symptoms of ASD. RESULTS: We observed stronger functional connectivity within several large-scale brain networks in children with ASD compared with TD children. This hyperconnectivity in ASD encompassed salience, default mode, frontotemporal, motor, and visual networks. This hyperconnectivity result was replicated in an independent cohort obtained from publicly available databases. Using maps of each individual's salience network, children with ASD could be discriminated from TD children with a classification accuracy of 78%, with 75% sensitivity and 80% specificity. The salience network showed the highest classification accuracy among all networks examined, and the blood oxygen-level dependent signal in this network predicted restricted and repetitive behavior scores. The classifier discriminated ASD from TD in the independent sample with 83% accuracy, 67% sensitivity, and 100% specificity. CONCLUSIONS AND RELEVANCE: Salience network hyperconnectivity may be a distinguishing feature in children with ASD. Quantification of brain network connectivity is a step toward developing biomarkers for objectively identifying children with ASD.