Neural disruptions during emotion regulation are common of generalized anxiety disorder (GAD). Identifying distinct functional and effective connectivity patterns in GAD may provide biomarkers for their diagnoses. This study aims to investigate the differences of features of brain network connectivity between GAD patients and healthy controls (HC), and to assess whether those differences can serve as biomarkers to distinguish GAD from controls. Independent component analysis (ICA) with hierarchical partner matching (HPM-ICA) was conducted on resting-state functional magnetic resonance imaging data collected from 20 GAD patients with medicine-free and 20 matched HC, identifying nine highly reproducible and significantly different functional brain connectivity patterns across diagnostic groups. We then utilized Granger causality (GC) to study the effective connectivity between the regions that identified by HPM-ICA. The linear discriminant analysis was finally used to distinguish GAD from controls with these measures of neural connectivity. The GAD patients showed stronger functional connectivity in amygdala, insula, putamen, thalamus, and posterior cingulate cortex, but weaker in frontal and temporal cortex compared with controls. Besides, the effective connectivity in GAD was decreased from the cortex to amygdala and basal ganglia. Applying the ICA and GC features to the classifier led to a classification accuracy of 87.5%, with a sensitivity of 90.0% and a specificity of 85.0%. These findings suggest that the presence of emotion dysregulation circuits may contribute to the pathophysiology of GAD, and these aberrant brain features may serve as robust brain biomarkers for GAD.

Machine learning and pattern recognition have been widely investigated in order to look for the biomarkers of Alzheimer’s disease (AD). However, most existing methods extract features by seed-based correlation, which not only requires prior information but also ignores the relationship between resting state functional magnetic resonance imaging (rs-fMRI) voxels. In this study, we proposed a deep learning classification framework with multivariate data-driven based feature extraction for automatic diagnosis of AD. Specifically, a three-level hierarchical partner matching independent components analysis (3LHPM-ICA) approach was proposed first in order to address the issues in spatial individual ICA, including the uncertainty of the numbers of components, the randomness of initial values, and the correspondence of ICs of multiple subjects, resulting in stable and reliable ICs which were applied as the intrinsic brain functional connectivity (FC) features. Second, Granger causality (GC) was utilized to infer directional interaction between the ICs that were identified by the 3LHPM-ICA method and extract the effective connectivity features. Finally, a deep learning classification framework was developed to distinguish AD from controls by fusing the functional and effective connectivities. A resting state fMRI dataset containing 34 AD patients and 34 normal controls (NCs) was applied to the multivariate deep learning platform, leading to a classification accuracy of 95.59%, with a sensitivity of 97.06% and a specificity of 94.12% with leave-one-out cross validation (LOOCV). The experimental results demonstrated that the measures of neural connectivities of ICA and GC followed by deep learning classification represented the most powerful methods of distinguishing AD clinical data from NCs, and these aberrant brain connectivities might serve as robust brain biomarkers for AD. This approach also allows for expansion of the methodology to classify other psychiatric disorders.

The purpose of this study was to examine the neural substrates and mechanisms that generate memory deficits, seizures and neuropsychiatric abnormalities in encephalitis with LGI1 antibodies using a data-driven, multimodal magnetic resonance imaging (MRI) approach. Methods: Functional MRI data were acquired from 14 anti-LGI1 encephalitis patients and 14 age and gender matched normal controls. Independent component analysis with hierarchical partner matching (HPM-ICA) was used to assess the whole-brain intrinsic functional connectivity. Granger causality (GC) was applied to investigate the effective connectivity among the brain regions that identified by HPM-ICA. Diffusion tensor imaging (DTI) was utilized to investigate white matter microstructural changes of the patients. Results: Participants with LGI1 antibodies encephalitis presented reduced functional connectivity in the brain areas associated with memory, cognition and motion circuits, while increased functional connectivity in putamen and caudate in comparison to the normal controls. Moreover, the effective connectivity in patients was decreased from the frontal cortex to supplementary motor area. Finally, patients had significant reductions in fractional anisotropy (FA) for the corpus callosum, internal capsule, corona radiata and superior longitudinal fasciculus, accompanied by increases in mean diffusivity (MD) for these regions as compared to controls. Conclusion: Our findings suggest that the neural disorder and behavioral deficits of anti-LGI1 encephalitis may be associated with extensive changes in brain connectivity and microstructure. These pathological alterations affect the basal ganglia and limbic system besides the temporal and frontal lobe.