Functional Connectivity Networks with Latent Distributions for Mild Cognitive Impairment Identification

This work presents a novel approach to estimate brain functional connectivity networks via generative learning. Due to the complexity and variability of rs-fMRI signal, we consider it as a random variable, and utilize variational autoencoder networks to encode it as a confidence distribution in the latent space rather than as a fixed vector, so as to establish the relationship between them. First, the mean time series of each brain region of interest is mapped into a multivariate Gaussian distribution. The correlation between two brain regions is measured by the Jensen-Shannon divergence that describes the statistical similarity between two probability distributions, and then the adjacency matrix is created to indicate the functional connectivity strength of pairwise brain regions. Meanwhile, our findings show that the adjacency matrices obtained at VAE latent spaces of different dimensionalities have good complementarity for MCI identification in precision and recall, and the classification performance can be further boosted by an efficient cascade of classifiers. This proposal constructs brain functional networks from a statistical modeling standpoint, improving the statistical ability of population data and the generalization ability of observation data variability. We evaluate the proposed framework over the task of identifying subjects with MCI from normal controls, and the experimental results on the public dataset show that our method significantly outperforms both the baseline and current state-of-the-art methods.

Brain function changes reveal rapid antidepressant effects of nitrous oxide for treatment-resistant depression:Evidence from task-state EEG

Nitrous oxide has rapid antidepressant effects in patients with treatment-resistant depression (TRD), but its underlying mechanisms of therapeutic actions are not well understood. Moreover, most of the current studies lack objective biological indicators to evaluate the changes of nitrous oxide-induced brain function for TRD. Therefore, this study assessed the effect of nitrous oxide on brain function for TRD based on event-related potential (ERP) components and functional connectivity networks (FCNs) methods. In this randomized, longitudinal, placebo-controlled trial, all TRD participants were divided into two groups to receive either a 1-hour inhalation of nitrous oxide or a placebo treatment, and they took part in the same task-state electroencephalogram (EEG) experiment before and after treatment. The experimental results showed that nitrous oxide improved depressive symptoms better than placebo in terms of 17-Hamilton Depression Rating Scale score (HAMD-17). Statistical analysis based on ERP components showed that nitrous oxide-induced significant differences in amplitude and latency of N1, P1, N2, P2. In addition, increased brain functional connectivity was found after nitrous oxide treatment. And the change of network metrics has a significant correlation with decreased depressive symptoms. These findings may suggest that nitrous oxide improves depression symptoms for TRD by modifying brain function.

Construction of embedded fMRI resting-state functional connectivity networks using manifold learning

We construct embedded functional connectivity networks (FCN) from benchmark resting-state functional magnetic resonance imaging (rsfMRI) data acquired from patients with schizophrenia and healthy controls based on linear and nonlinear manifold learning algorithms, namely, Multidimensional Scaling, Isometric Feature Mapping, Diffusion Maps, Locally Linear Embedding and kernel PCA. Furthermore, based on key global graph-theoretic properties of the embedded FCN, we compare their classification potential using machine learning. We also assess the performance of two metrics that are widely used for the construction of FCN from fMRI, namely the Euclidean distance and the cross correlation metric. We show that diffusion maps with the cross correlation metric outperform the other combinations.

Diagnosis of early Alzheimer's disease based on dynamic high order networks

Machine learning methods have been widely used for early diagnosis of Alzheimer's disease (AD) via functional connectivity networks (FCNs) analysis from neuroimaging data. The conventional low-order FCNs are obtained by time-series correlation of the whole brain based on resting-state functional magnetic resonance imaging (R-fMRI). However, FCNs overlook inter-region interactions, which limits application to brain disease diagnosis. To overcome this drawback, we develop a novel framework to exploit the high-level dynamic interactions among brain regions for early AD diagnosis. Specifically, a sliding window approach is employed to generate some R-fMRI sub-series. The correlations among these sub-series are then used to construct a series of dynamic FCNs. High-order FCNs based on the topographical similarity between each pair of the dynamic FCNs are then constructed. Afterward, a local weight clustering method is used to extract effective features of the network, and the least absolute shrinkage and selection operation method is chosen for feature selection. A support vector machine is employed for classification, and the dynamic high-order network approach is evaluated on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset. Our experimental results demonstrate that the proposed approach not only achieves promising results for AD classification, but also successfully recognizes disease-related biomarkers.

Intrinsic functional neurocircuitry associated with treatment response to transdiagnostic CBT in bipolar disorder with anxiety

BACKGROUND

Anxiety in bipolar disorder (BD) exacerbates emotion dysregulation and reduces treatment response. We recently conducted a pilot trial of transdiagnostic CBT to target anxiety and emotion dysregulation in BD adjunctive to pharmacotherapy. Reductions in depression and anxiety symptoms were significantly predicted by baseline levels of neuroticism and perceived affective control, as well as changes over time in emotion regulation skills. The present study investigates mechanism of treatment response by examining the relationship between baseline emotion regulation-related neural circuitry and trial outcomes.

METHODS

Nineteen patients completed baseline resting state fMRI scans prior to treatment randomization. Functional connectivity between the anterior insula (AI) and regions in the salience network (SN), default mode network (DMN), and executive control network (ECN) were examined as predictors of baseline and treatment-related changes in emotion regulation.

RESULTS

Greater improvements in emotion regulation were predicted by weaker right dorsal AI - right ventrolateral prefrontal cortex (VLPFC; SN) and stronger bilateral dorsal AI - bilateral amygdala functional connectivity. Baseline neuroticism was negatively correlated with right dorsal AI- inferior parietal lobule (ECN) functional connectivity, and baseline deficits in perceived affective control were positively associated with ventral AI - bilateral dACC (SN) connectivity.

LIMITATIONS

Small sample limits interpretability of treatment-specific effects.

CONCLUSION

Baseline functional connectivity of emotion-regulation related neural circuitry significantly predicted change in emotion regulation-related dimensions associated with anxiety and depression symptom reduction. Future studies are needed to determine if employing methods such as neuromodulation to rehabilitate relevant neural circuitry may improve ultimate treatment outcomes of transdiagnostic CBT for BD and anxiety.

Resting state brain network function in major depression - Depression symptomatology, antidepressant treatment effects, future research

The alterations of functional connectivity brain networks in major depressive disorder (MDD) have been subject of a large number of studies. Using different methodologies and focusing on diverse aspects of the disease, research shows heterogeneous results lacking integration. Disrupted network connectivity has been found in core MDD networks like the default mode network (DMN), the central executive network (CEN), and the salience network, but also in cerebellar and thalamic circuitries. Here we review literature published on resting state brain network function in MDD focusing on methodology, and clinical characteristics including symptomatology and antidepressant treatment related findings. There are relatively few investigations concerning the qualitative aspects of symptomatology of MDD, whereas most studies associate quantitative aspects with distinct resting state functional connectivity alterations. Such depression severity associated alterations are found in the DMN, frontal, cerebellar and thalamic brain regions as well as the insula and the subgenual anterior cingulate cortex. Similarly, different therapeutical options in MDD and their effects on brain function showed patchy results. Herein, pharmaceutical treatments reveal functional connectivity alterations throughout multiple brain regions notably the DMN, fronto-limbic, and parieto-temporal regions. Psychotherapeutical interventions show significant functional connectivity alterations in fronto-limbic networks, whereas electroconvulsive therapy and repetitive transcranial magnetic stimulation result in alterations of the subgenual anterior cingulate cortex, the DMN, the CEN and the dorsal lateral prefrontal cortex. While it appears clear that functional connectivity alterations are associated with the pathophysiology and treatment of MDD, future research should also generate a common strategy for data acquisition and analysis, as a least common denominator, to set the basis for comparability across studies and implementation of functional connectivity as a scientifically and clinically useful biomarker.

A variance components model for statistical inference on functional connectivity networks

We propose a variance components linear modeling framework to conduct statistical inference on functional connectivity networks that directly accounts for the temporal autocorrelation inherent in functional magnetic resonance imaging (fMRI) time series data and for the heterogeneity across subjects in the study. The novel method estimates the autocorrelation structure in a nonparametric and subject-specific manner, and estimates the variance due to the heterogeneity using iterative least squares. We apply the new model to a resting-state fMRI study to compare the functional connectivity networks in both typical and reading impaired young adults in order to characterize the resting state networks that are related to reading processes. We also compare the performance of our model to other methods of statistical inference on functional connectivity networks that do not account for the temporal autocorrelation or heterogeneity across the subjects using simulated data, and show that by accounting for these sources of variation and covariation results in more powerful tests for statistical inference.

Independent functional connectivity networks underpin food and monetary reward sensitivity in excess weight

Overvaluation of palatable food is a primary driver of obesity, and is associated with brain regions of the reward system. However, it remains unclear if this network is specialized in food reward, or generally involved in reward processing. We used functional magnetic resonance imaging (fMRI) to characterize functional connectivity during processing of food and monetary rewards. Thirty-nine adults with excess weight and 37 adults with normal weight performed the Willingness to Pay for Food task and the Monetary Incentive Delay task in the fMRI scanner. A data-driven graph approach was applied to compare whole-brain, task-related functional connectivity between groups. Excess weight was associated with decreased functional connectivity during the processing of food rewards in a network involving primarily frontal and striatal areas, and increased functional connectivity during the processing of monetary rewards in a network involving principally frontal and parietal areas. These two networks were topologically and anatomically distinct, and were independently associated with BMI. The processing of food and monetary rewards involve segregated neural networks, and both are altered in individuals with excess weight.