The Function Biomedical Informatics Research Network Data Repository

A spatially constrained independent component analysis jointly informed by structural and functional network connectivity

There are a growing number of neuroimaging studies motivating joint structural and functional brain connectivity. Brain connectivity of different modalities provides insight into brain functional organization by leveraging complementary information, especially for brain disorders such as schizophrenia. In this paper, we propose a multi-modal independent component analysis (ICA) model that utilizes information from both structural and functional brain connectivity guided by spatial maps to estimate intrinsic connectivity networks (ICNs). Structural connectivity is estimated through whole-brain tractography on diffusion-weighted MRI (dMRI), while functional connectivity is derived from resting-state functional MRI (rs-fMRI). The proposed structural-functional connectivity and spatially constrained ICA (sfCICA) model estimates ICNs at the subject level using a multi-objective optimization framework. We evaluated our model using synthetic and real datasets (including dMRI and rs-fMRI from 149 schizophrenia patients and 162 controls). Multi-modal ICNs revealed enhanced functional coupling between ICNs with higher structural connectivity, improved modularity, and network distinction, particularly in schizophrenia. Statistical analysis of group differences showed more significant differences in the proposed model compared to the unimodal model. In summary, the sfCICA model showed benefits from being jointly informed by structural and functional connectivity. These findings suggest advantages in simultaneously learning effectively and enhancing connectivity estimates using structural connectivity.

Assistive tools for classifying neurological disorders using fMRI and deep learning: A guide and example

BACKGROUND

Deep-learning (DL) methods are rapidly changing the way researchers classify neurological disorders. For example, combining functional magnetic resonance imaging (fMRI) and DL has helped researchers identify functional biomarkers of neurological disorders (e.g., brain activation and connectivity) and pilot innovative diagnostic models. However, the knowledge required to perform DL analyses is often domain-specific and is not widely taught in the brain sciences (e.g., psychology, neuroscience, and cognitive science). Conversely, neurological diagnoses and neuroimaging training (e.g., fMRI) are largely restricted to the brain and medical sciences. In turn, these disciplinary knowledge barriers and distinct specializations can act as hurdles that prevent the combination of fMRI and DL pipelines. The complexity of fMRI and DL methods also hinders their clinical adoption and generalization to real-world diagnoses. For example, most current models are not designed for clinical settings or use by nonspecialized populations such as students, clinicians, and healthcare workers. Accordingly, there is a growing area of assistive tools (e.g., software and programming packages) that aim to streamline and increase the accessibility of fMRI and DL pipelines for the diagnoses of neurological disorders.

OBJECTIVES AND METHODS

In this study, we present an introductory guide to some popular DL and fMRI assistive tools. We also create an example autism spectrum disorder (ASD) classification model using assistive tools (e.g., Optuna, GIFT, and the ABIDE preprocessed repository), fMRI, and a convolutional neural network.

RESULTS

In turn, we provide researchers with a guide to assistive tools and give an example of a streamlined fMRI and DL pipeline.

CONCLUSIONS

We are confident that this study can help more researchers enter the field and create accessible fMRI and deep-learning diagnostic models for neurological disorders.

A Brainwide Risk Score for Psychiatric Disorder Evaluated in a Large Adolescent Population Reveals Increased Divergence Among Higher-Risk Groups Relative to Control Participants

BACKGROUND

Accurate psychiatric risk assessment requires biomarkers that are both stable and adaptable to development. Functional network connectivity (FNC), which steadily reconfigures over time, potentially contains abundant information to assess psychiatric risks. However, the absence of suitable analytical methodologies has constrained this area of investigation.

METHODS

We investigated the brainwide risk score (BRS), a novel FNC-based metric that contrasts the relative distances of an individual's FNC to that of psychiatric disorders versus healthy control references. To generate group-level disorder and healthy control references, we utilized a large brain imaging dataset containing 5231 total individuals diagnosed with schizophrenia, autism spectrum disorder, major depressive disorder, and bipolar disorder and their corresponding healthy control individuals. The BRS metric was employed to assess the psychiatric risk in 2 new datasets: Adolescent Brain Cognitive Development (ABCD) Study (n = 8191) and Human Connectome Project Early Psychosis (n = 170).

RESULTS

The BRS revealed a clear, reproducible gradient of FNC patterns from low to high risk for each psychiatric disorder in unaffected adolescents. We found that low-risk ABCD Study adolescent FNC patterns for each disorder were strongly present in over 25% of the ABCD Study participants and homogeneous, whereas high-risk patterns of each psychiatric disorder were strongly present in about 1% of ABCD Study participants and heterogeneous. The BRS also showed its effectiveness in predicting psychosis scores and distinguishing individuals with early psychosis from healthy control individuals.

CONCLUSIONS

The BRS could be a new image-based tool for assessing psychiatric vulnerability over time and in unaffected individuals, and it could also serve as a potential biomarker, facilitating early screening and monitoring interventions.

More reliable biomarkers and more accurate prediction for mental disorders using a label-noise filtering-based dimensional prediction method

The integration of neuroimaging with artificial intelligence is crucial for advancing the diagnosis of mental disorders. However, challenges arise from incomplete matching between diagnostic labels and neuroimaging. Here, we propose a label-noise filtering-based dimensional prediction (LAMP) method to identify reliable biomarkers and achieve accurate prediction for mental disorders. Our method proposes to utilize a label-noise filtering model to automatically filter out unclear cases from a neuroimaging perspective, and then the typical subjects whose diagnostic labels align with neuroimaging measures are used to construct a dimensional prediction model to score independent subjects. Using fMRI data of schizophrenia patients and healthy controls (n = 1,245), our method yields consistent scores to independent subjects, leading to more distinguishable relabeled groups with an enhanced classification accuracy of 31.89%. Additionally, it enables the exploration of stable abnormalities in schizophrenia. In summary, our LAMP method facilitates the identification of reliable biomarkers and accurate diagnosis of mental disorders using neuroimages.

Structural variability of the cerebral cortex in schizophrenia and its association with clinical symptoms

BACKGROUND

Substantial evidence indicates structural abnormalities in the cerebral cortex of patients with schizophrenia (SCZ), although their clinical implications remain unclear. Previous case-control studies have investigated group-level differences in structural abnormalities, although the study design cannot account for interindividual differences. Recent research has focused on the association between the heterogeneity of the cerebral cortex morphometric features and clinical heterogeneity.

METHODS

We used neuroimaging data from 420 healthy controls and 695 patients with SCZ from seven studies. Four cerebral cortex measures were obtained: surface area, gray matter volume, thickness, and local gyrification index. We calculated the coefficient of variation (CV) and person-based similarity index (PBSI) scores and performed group comparisons. Associations between the PBSI scores and cognitive functions were evaluated using Spearman's rho test and normative modeling.

RESULTS

Patients with SCZ had a greater CV of surface area and cortical thickness than those of healthy controls. All PBSI scores across cortical measures were lower in patients with SCZ than in HCs. In the patient group, the PBSI scores for gray matter volume and all cortical measures taken together positively correlated with the full-scale IQ scores. Patients with deviant PBSI scores for gray matter volume and all cortical measures taken together had lower full-scale IQ scores than those of other patients.

CONCLUSIONS

The cerebral cortex in patients with SCZ showed greater regional and global structural variability than that in healthy controls. Patients with deviant similarity of cortical structural profiles exhibited a lower general intelligence than those exhibited by the other patients.

Identifying canonical and replicable multi-scale intrinsic connectivity networks in 100k+ resting-state fMRI datasets

Despite the known benefits of data-driven approaches, the lack of approaches for identifying functional neuroimaging patterns that capture both individual variations and inter-subject correspondence limits the clinical utility of rsfMRI and its application to single-subject analyses. Here, using rsfMRI data from over 100k individuals across private and public datasets, we identify replicable multi-spatial-scale canonical intrinsic connectivity network (ICN) templates via the use of multi-model-order independent component analysis (ICA). We also study the feasibility of estimating subject-specific ICNs via spatially constrained ICA. The results show that the subject-level ICN estimations vary as a function of the ICN itself, the data length, and the spatial resolution. In general, large-scale ICNs require less data to achieve specific levels of (within- and between-subject) spatial similarity with their templates. Importantly, increasing data length can reduce an ICN's subject-level specificity, suggesting longer scans may not always be desirable. We also find a positive linear relationship between data length and spatial smoothness (possibly due to averaging over intrinsic dynamics), suggesting studies examining optimized data length should consider spatial smoothness. Finally, consistency in spatial similarity between ICNs estimated using the full data and subsets across different data lengths suggests lower within-subject spatial similarity in shorter data is not wholly defined by lower reliability in ICN estimates, but may be an indication of meaningful brain dynamics which average out as data length increases.

Chromatic fusion: Generative multimodal neuroimaging data fusion provides multi-informed insights into schizophrenia

This work proposes a novel generative multimodal approach to jointly analyze multimodal data while linking the multimodal information to colors. We apply our proposed framework, which disentangles multimodal data into private and shared sets of features from pairs of structural (sMRI), functional (sFNC and ICA), and diffusion MRI data (FA maps). With our approach, we find that heterogeneity in schizophrenia is potentially a function of modality pairs. Results show (1) schizophrenia is highly multimodal and includes changes in specific networks, (2) non-linear relationships with schizophrenia are observed when interpolating among shared latent dimensions, and (3) we observe a decrease in the modularity of functional connectivity and decreased visual-sensorimotor connectivity for schizophrenia patients for the FA-sFNC and sMRI-sFNC modality pairs, respectively. Additionally, our results generally indicate decreased fractional corpus callosum anisotropy, and decreased spatial ICA map and voxel-based morphometry strength in the superior frontal lobe as found in the FA-sFNC, sMRI-FA, and sMRI-ICA modality pair clusters. In sum, we introduce a powerful new multimodal neuroimaging framework designed to provide a rich and intuitive understanding of the data which we hope challenges the reader to think differently about how modalities interact.

Two neurostructural subtypes: results of machine learning on brain images from 4,291 individuals with schizophrenia

Machine learning can be used to define subtypes of psychiatric conditions based on shared clinical and biological foundations, presenting a crucial step toward establishing biologically based subtypes of mental disorders. With the goal of identifying subtypes of disease progression in schizophrenia, here we analyzed cross-sectional brain structural magnetic resonance imaging (MRI) data from 4,291 individuals with schizophrenia (1,709 females, age=32.5 years±11.9) and 7,078 healthy controls (3,461 females, age=33.0 years±12.7) pooled across 41 international cohorts from the ENIGMA Schizophrenia Working Group, non-ENIGMA cohorts and public datasets. Using a machine learning approach known as Subtype and Stage Inference (SuStaIn), we implemented a brain imaging-driven classification that identifies two distinct neurostructural subgroups by mapping the spatial and temporal trajectory of gray matter (GM) loss in schizophrenia. Subgroup 1 (n=2,622) was characterized by an early cortical-predominant loss (ECL) with enlarged striatum, whereas subgroup 2 (n=1,600) displayed an early subcortical-predominant loss (ESL) in the hippocampus, amygdala, thalamus, brain stem and striatum. These reconstructed trajectories suggest that the GM volume reduction originates in the Broca's area/adjacent fronto-insular cortex for ECL and in the hippocampus/adjacent medial temporal structures for ESL. With longer disease duration, the ECL subtype exhibited a gradual worsening of negative symptoms and depression/anxiety, and less of a decline in positive symptoms. We confirmed the reproducibility of these imaging-based subtypes across various sample sites, independent of macroeconomic and ethnic factors that differed across these geographic locations, which include Europe, North America and East Asia. These findings underscore the presence of distinct pathobiological foundations underlying schizophrenia. This new imaging-based taxonomy holds the potential to identify a more homogeneous sub-population of individuals with shared neurobiological attributes, thereby suggesting the viability of redefining existing disorder constructs based on biological factors.

Multimodal subspace independent vector analysis captures latent subspace structures in large multimodal neuroimaging studies

We present Multimodal Subspace Independent Vector Analysis (MSIVA), a methodology to capture both joint and unique vector sources across multiple data modalities by defining linked and modality-specific subspaces. In particular, MSIVA enables the estimation of independent subspaces of various sizes within modalities and their one-to-one linkage to corresponding subspaces across modalities. We compare MSIVA to a fully unimodal initialization baseline and a fully multimodal initialization baseline, and evaluate all three approaches with five distinct subspace structures on synthetic and neuroimaging datasets. We first demonstrate that MSIVA and the unimodal baseline can identify the correct ground-truth subspace structures from the incorrect ones in multiple synthetic datasets, while the multimodal baseline fails at detecting high-dimensional subspace structures. We then show that MSIVA can better capture the latent subspace structure with the minimum loss value in two large multimodal neuroimaging datasets compared to the unimodal baseline. Our results from subsequent per-subspace canonical correlation analysis (CCA) and brain-phenotype modeling demonstrate that the sources from the optimal subspace structure are strongly associated with phenotype measures, including age, sex and schizophrenia-related effects. Our proposed methodology MSIVA can be applied to capture linked and unique biomarkers from multimodal neuroimaging data.

Alterations in grey matter structure linked to frequency-specific cortico-subcortical connectivity in schizophrenia via multimodal data fusion

Schizophrenia (SZ) is a complex psychiatric disorder that is currently defined by symptomatic and behavioral, rather than biological, criteria. Neuroimaging is an appealing avenue for SZ biomarker development, as several neuroimaging-based studies comparing individuals with SZ to healthy controls (HC) have shown measurable group differences in brain structure, as well as functional brain alterations in both static and dynamic functional network connectivity (sFNC and dFNC, respectively). The recently proposed filter-banked connectivity (FBC) method extends the standard dFNC sliding-window approach to estimate FNC within an arbitrary number of distinct frequency bands. The initial implementation used a set of filters spanning the full connectivity spectral range, providing a unified approach to examine both sFNC and dFNC in a single analysis. Initial FBC results found that individuals with SZ spend more time in a less structured, more disconnected low-frequency (i.e., static) FNC state than HC, as well as preferential SZ occupancy in high-frequency connectivity states, suggesting a frequency-specific component underpinning the functional dysconnectivity observed in SZ. Building on these findings, we sought to link such frequency-specific patterns of FNC to covarying data-driven structural brain networks in the context of SZ. Specifically, we employ a multi-set canonical correlation analysis + joint independent components analysis (mCCA + jICA) data fusion framework to study the connection between grey matter volume (GMV) maps and FBC states across the full connectivity frequency spectrum. Our multimodal analysis identified two joint sources that captured co-varying patterns of frequency-specific functional connectivity and alterations in GMV with significant group differences in loading parameters between the SZ group and HC. The first joint source linked frequency-modulated connections between the subcortical and sensorimotor networks and GMV alterations in the frontal and temporal lobes, while the second joint source identified a relationship between low-frequency cerebellar-sensorimotor connectivity and structural changes in both the cerebellum and motor cortex. Together, these results show a strong connection between cortico-subcortical functional connectivity at both high and low frequencies and alterations in cortical GMV that may be relevant to the pathogenesis and pathophysiology of SZ.

Network Differential in Gaussian Graphical Models from Multimodal Neuroimaging Data. ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. ANNUAL INTERNATIONAL CONFERENCE 2023;2023:1-6. [PMID: 38083176 DOI: 10.1109/embc40787.2023.10340856] [Citation(s) in RCA: 0] [Impact Index Per Article: 0]

Multimodal brain network analysis has the potential to provide insights into the mechanisms of brain disorders. Most previous studies have analyzed either unimodal brain graphs or focused on local/global graphic metrics with little consideration of details of disrupted paths in the patient group. As we show, the combination of multimodal brain graphs and disrupted path-based analysis can be highly illuminating to recognize path-based disease biomarkers. In this study, we first propose a way to estimate multimodal brain graphs using static functional network connectivity (sFNC) and gray matter features using a Gaussian graphical model of schizophrenia versus controls. Next, applying the graph theory approach we identify disconnectors or connectors in the patient group graph that create additional paths or cause absent paths compared to the control graph. Results showed several edges in the schizophrenia group graph that trigger missing or additional paths. Identified edges associated with these disrupted paths were identified both within and between dFNC and gray matter which highlights the importance of considering multimodal studies and moving beyond pairwise edges to provide a more comprehensive understanding of brain disorders.Clinical Relevance- We identified a path-based biomarker in schizophrenia, by imitating the structure of paths in a multimodal (sMIR+fMRI) brain graph of the control group. Identified cross-modal edges associated with disrupted paths were related to the middle temporal gyrus and cerebellar regions.

A Deep Learning Approach for Psychosis Spectrum Label Noise Detection from Multimodal Neuroimaging Data

Understanding the structural and functional mechanisms of the brain is challenging for mood and mental disorders. Many neuroimaging techniques are widely used to reveal hidden patterns from different brain imaging modalities. However, these findings are bounded by the limitation of each modality. In addition, the lack of validity of current psychosis nosology created more complications in understanding biomarkers. In this study, we introduced a deep convolutional framework to classify and identify label noises using structural and functional magnetic resonance imaging data. We applied our method to functional and structural MRI data from a schizophrenia dataset and evaluated the model's performance in a cross-validated form. In addition, we introduced a noise criterion to distinguish a potentially noisy subject for each modality. Our results show the learned model using resting-state functional MRI data is more informative and has higher performance in comparison with structural MRI data. Lastly, based on the noise level, we investigated potential borderline subjects as possible subtypes and made a statistical analysis to distinguish differences between resting-state static functional connectivity features.Clinical Relevance- Results show schizophrenia patients are separable from the healthy control group based on their neuroimaging data and resting-state functional MRI data is more informative than structural MRI data and hence contains less label noise.

GLACIER: GLASS-BOX TRANSFORMER FOR INTERPRETABLE DYNAMIC NEUROIMAGING

Deep learning models can perform as well or better than humans in many tasks, especially vision related. Almost exclusively, these models are used to perform classification or prediction. However, deep learning models are usually of black-box nature, and it is often difficult to interpret the model or the features. The lack of interpretability causes a restrain from applying deep learning to fields such as neuroimaging, where the results must be transparent, and interpretable. Therefore, we present a 'glass-box' deep learning model and apply it to the field of neuroimaging. Our model mixes spatial and temporal dimensions in succession to estimate dynamic connectivity between the brain's intrinsic networks. The interpretable connectivity matrices produced by our model result in beating state-of-the-art models on many tasks using multiple functional MRI datasets. More importantly, our model estimates task-based flexible connectivity matrices, unlike static methods such as Pearson's correlation coefficients.

Chromatic fusion: generative multimodal neuroimaging data fusion provides multi-informed insights into schizophrenia

This work proposes a novel generative multimodal approach to jointly analyze multimodal data while linking the multimodal information to colors. By linking colors to private and shared information from modalities, we introduce chromatic fusion, a framework that allows for intuitively interpreting multimodal data. We test our framework on structural, functional, and diffusion modality pairs. In this framework, we use a multimodal variational autoencoder to learn separate latent subspaces; a private space for each modality, and a shared space between both modalities. These subspaces are then used to cluster subjects, and colored based on their distance from the variational prior, to obtain meta-chromatic patterns (MCPs). Each subspace corresponds to a different color, red is the private space of the first modality, green is the shared space, and blue is the private space of the second modality. We further analyze the most schizophrenia-enriched MCPs for each modality pair and find that distinct schizophrenia subgroups are captured by schizophrenia-enriched MCPs for different modality pairs, emphasizing schizophrenia's heterogeneity. For the FA-sFNC, sMRI-ICA, and sMRI-ICA MCPs, we generally find decreased fractional corpus callosum anisotropy and decreased spatial ICA map and voxel-based morphometry strength in the superior frontal lobe for schizophrenia patients. To additionally highlight the importance of the shared space between modalities, we perform a robustness analysis of the latent dimensions in the shared space across folds. These robust latent dimensions are subsequently correlated with schizophrenia to reveal that for each modality pair, multiple shared latent dimensions strongly correlate with schizophrenia. In particular, for FA-sFNC and sMRI-sFNC shared latent dimensions, we respectively observe a reduction in the modularity of the functional connectivity and a decrease in visual-sensorimotor connectivity for schizophrenia patients. The reduction in modularity couples with increased fractional anisotropy in the left part of the cerebellum dorsally. The reduction in the visual-sensorimotor connectivity couples with a reduction in the voxel-based morphometry generally but increased dorsal cerebellum voxel-based morphometry. Since the modalities are trained jointly, we can also use the shared space to try and reconstruct one modality from the other. We show that cross-reconstruction is possible with our network and is generally much better than depending on the variational prior. In sum, we introduce a powerful new multimodal neuroimaging framework designed to provide a rich and intuitive understanding of the data that we hope challenges the reader to think differently about how modalities interact.

Reliability and clinical utility of spatially constrained estimates of intrinsic functional networks from very short fMRI scans

Resting-state functional network connectivity (rsFNC) has shown utility for identifying characteristic functional brain patterns in individuals with psychiatric and mood disorders, providing a promising avenue for biomarker development. However, several factors have precluded widespread clinical adoption of rsFNC diagnostics, namely a lack of standardized approaches for capturing comparable and reproducible imaging markers across individuals, as well as the disagreement on the amount of data required to robustly detect intrinsic connectivity networks (ICNs) and diagnostically relevant patterns of rsFNC at the individual subject level. Recently, spatially constrained independent component analysis (scICA) has been proposed as an automated method for extracting ICNs standardized to a chosen network template while still preserving individual variation. Leveraging the scICA methodology, which solves the former challenge of standardized neuroimaging markers, we investigate the latter challenge of identifying a minimally sufficient data length for clinical applications of resting-state fMRI (rsfMRI). Using a dataset containing rsfMRI scans of individuals with schizophrenia and controls (M = 310) as well as simulated rsfMRI, we evaluated the robustness of ICN and rsFNC estimates at both the subject- and group-level, as well as the performance of diagnostic classification, with respect to the length of the rsfMRI time course. We found individual estimates of ICNs and rsFNC from the full-length (5 min) reference time course were sufficiently approximated with just 3-3.5 min of data (r = 0.85, 0.88, respectively), and significant differences in group-average rsFNC could be sufficiently approximated with even less data, just 2 min (r = 0.86). These results from the shorter clinical data were largely consistent with the results from validation experiments using longer time series from both simulated (30 min) and real-world (14 min) datasets, in which estimates of subject-level FNC were reliably estimated with 3-5 min of data. Moreover, in the real-world data we found rsFNC and ICN estimates generated across the full range of data lengths (0.5-14 min) more reliably matched those generated from the first 5 min of scan time than those generated from the last 5 min, suggesting increased influence of "late scan" noise factors such as fatigue or drowsiness may limit the reliability of FNC from data collected after 10+ min of scan time, further supporting the notion of shorter scans. Lastly, a diagnostic classification model trained on just 2 min of data retained 97%-98% classification accuracy relative to that of the full-length reference model. Our results suggest that, when decomposed with scICA, rsfMRI scans of just 2-5 min show good clinical utility without significant loss of individual FNC information of longer scan lengths.

Estimating uncertainty in read-out patterns: Application to controls-based denoising and voxel-based morphometry patterns in neurodegenerative and neuropsychiatric diseases

Quantifying pathology-related patterns in patient data with pattern expression score (PES) is a standard approach in medical image analysis. In order to estimate the PES error, we here propose to express the uncertainty contained in read-out patterns in terms of the expected squared Euclidean distance between the read-out pattern and the unknown "true" pattern (squared standard error of the read-out pattern, SE² ). Using SE² , we predicted and optimized the net benefit (NBe) of the recently suggested method controls-based denoising (CODE) by weighting patterns of nonpathological variance (NPV). Multi-center MRI (1192 patients with various neurodegenerative and neuropsychiatric diseases, 1832 healthy controls) were analysed with voxel-based morphometry. For each pathology, accounting for SE² , NBe correctly predicted classification improvement and allowed to optimize NPV pattern weights. Using these weights, CODE improved classification performances in all but one analyses, for example, for prediction of conversion to Alzheimer's disease (AUC 0.81 vs. 0.75, p = .01), diagnosis of autism (AUC 0.66 vs. 0.60, p < .001), and of major depressive disorder (AUC 0.62 vs. 0.50, p = .03). We conclude that the degree of uncertainty in a read-out pattern should generally be reported in PES-based analyses and suggest using weighted CODE as a complement to PES-based analyses.

A method for estimating and characterizing explicitly nonlinear dynamic functional network connectivity in resting-state fMRI data

The past 10 years have seen an explosion of approaches that focus on the study of time-resolved change in functional connectivity (FC). FC characterization among networks at a whole-brain level is frequently termed functional network connectivity (FNC). Time-resolved or dynamic functional network connectivity (dFNC) focuses on the estimation of transient, recurring, whole-brain patterns of FNC. While most approaches in this area have attempted to capture dynamic linear correlation, we are particularly interested in whether explicitly nonlinear relationships, above and beyond linear, are present and contain unique information. This study thus proposes an approach to assess explicitly nonlinear dynamic functional network connectivity (EN dFNC) derived from the relationship among independent component analysis time courses. Linear relationships were removed at each time point to evaluate, typically ignored, explicitly nonlinear dFNC using normalized mutual information (NMI). Simulations showed the proposed method estimated explicitly nonlinearity over time, even within relatively short windows of data. We then, applied our approach on 151 schizophrenia patients, and 163 healthy controls fMRI data and found three unique, highly structured, mostly long-range, functional states that also showed significant group differences. In particular, explicitly nonlinear relationships tend to be more widespread than linear ones. Results also highlighted a state with long range connections to the visual domain, which were significantly reduced in schizophrenia. Overall, this work suggests that quantifying EN dFNC may provide a complementary and potentially valuable tool for studying brain function by exposing relevant variation that is typically ignored.

Through the looking glass: Deep interpretable dynamic directed connectivity in resting fMRI

Brain network interactions are commonly assessed via functional (network) connectivity, captured as an undirected matrix of Pearson correlation coefficients. Functional connectivity can represent static and dynamic relations, but often these are modeled using a fixed choice for the data window Alternatively, deep learning models may flexibly learn various representations from the same data based on the model architecture and the training task. However, the representations produced by deep learning models are often difficult to interpret and require additional posthoc methods, e.g., saliency maps. In this work, we integrate the strengths of deep learning and functional connectivity methods while also mitigating their weaknesses. With interpretability in mind, we present a deep learning architecture that exposes a directed graph layer that represents what the model has learned about relevant brain connectivity. A surprising benefit of this architectural interpretability is significantly improved accuracy in discriminating controls and patients with schizophrenia, autism, and dementia, as well as age and gender prediction from functional MRI data. We also resolve the window size selection problem for dynamic directed connectivity estimation as we estimate windowing functions from the data, capturing what is needed to estimate the graph at each time-point. We demonstrate efficacy of our method in comparison with multiple existing models that focus on classification accuracy, unlike our interpretability-focused architecture. Using the same data but training different models on their own discriminative tasks we are able to estimate task-specific directed connectivity matrices for each subject. Results show that the proposed approach is also more robust to confounding factors compared to standard dynamic functional connectivity models. The dynamic patterns captured by our model are naturally interpretable since they highlight the intervals in the signal that are most important for the prediction. The proposed approach reveals that differences in connectivity among sensorimotor networks relative to default-mode networks are an important indicator of dementia and gender. Dysconnectivity between networks, specially sensorimotor and visual, is linked with schizophrenic patients, however schizophrenic patients show increased intra-network default-mode connectivity compared to healthy controls. Sensorimotor connectivity was important for both dementia and schizophrenia prediction, but schizophrenia is more related to dysconnectivity between networks whereas, dementia bio-markers were mostly intra-network connectivity.

Nonlinear functional network connectivity in resting functional magnetic resonance imaging data

In this work, we focus on explicitly nonlinear relationships in functional networks. We introduce a technique using normalized mutual information (NMI) that calculates the nonlinear relationship between different brain regions. We demonstrate our proposed approach using simulated data and then apply it to a dataset previously studied by Damaraju et al. This resting-state fMRI data included 151 schizophrenia patients and 163 age- and gender-matched healthy controls. We first decomposed these data using group independent component analysis (ICA) and yielded 47 functionally relevant intrinsic connectivity networks. Our analysis showed a modularized nonlinear relationship among brain functional networks that was particularly noticeable in the sensory and visual cortex. Interestingly, the modularity appears both meaningful and distinct from that revealed by the linear approach. Group analysis identified significant differences in explicitly nonlinear functional network connectivity (FNC) between schizophrenia patients and healthy controls, particularly in the visual cortex, with controls showing more nonlinearity (i.e., higher normalized mutual information between time courses with linear relationships removed) in most cases. Certain domains, including subcortical and auditory, showed relatively less nonlinear FNC (i.e., lower normalized mutual information), whereas links between the visual and other domains showed evidence of substantial nonlinear and modular properties. Overall, these results suggest that quantifying nonlinear dependencies of functional connectivity may provide a complementary and potentially important tool for studying brain function by exposing relevant variation that is typically ignored. Beyond this, we propose a method that captures both linear and nonlinear effects in a "boosted" approach. This method increases the sensitivity to group differences compared to the standard linear approach, at the cost of being unable to separate linear and nonlinear effects.

Deep multimodal predictome for studying mental disorders

Characterizing neuropsychiatric disorders is challenging due to heterogeneity in the population. We propose combining structural and functional neuroimaging and genomic data in a multimodal classification framework to leverage their complementary information. Our objectives are two-fold (i) to improve the classification of disorders and (ii) to introspect the concepts learned to explore underlying neural and biological mechanisms linked to mental disorders. Previous multimodal studies have focused on naïve neural networks, mostly perceptron, to learn modality-wise features and often assume equal contribution from each modality. Our focus is on the development of neural networks for feature learning and implementing an adaptive control unit for the fusion phase. Our mid fusion with attention model includes a multilayer feed-forward network, an autoencoder, a bi-directional long short-term memory unit with attention as the features extractor, and a linear attention module for controlling modality-specific influence. The proposed model acquired 92% (p < .0001) accuracy in schizophrenia prediction, outperforming several other state-of-the-art models applied to unimodal or multimodal data. Post hoc feature analyses uncovered critical neural features and genes/biological pathways associated with schizophrenia. The proposed model effectively combines multimodal neuroimaging and genomics data for predicting mental disorders. Interpreting salient features identified by the model may advance our understanding of their underlying etiological mechanisms.

Interpreting models interpreting brain dynamics

Brain dynamics are highly complex and yet hold the key to understanding brain function and dysfunction. The dynamics captured by resting-state functional magnetic resonance imaging data are noisy, high-dimensional, and not readily interpretable. The typical approach of reducing this data to low-dimensional features and focusing on the most predictive features comes with strong assumptions and can miss essential aspects of the underlying dynamics. In contrast, introspection of discriminatively trained deep learning models may uncover disorder-relevant elements of the signal at the level of individual time points and spatial locations. Yet, the difficulty of reliable training on high-dimensional low sample size datasets and the unclear relevance of the resulting predictive markers prevent the widespread use of deep learning in functional neuroimaging. In this work, we introduce a deep learning framework to learn from high-dimensional dynamical data while maintaining stable, ecologically valid interpretations. Results successfully demonstrate that the proposed framework enables learning the dynamics of resting-state fMRI directly from small data and capturing compact, stable interpretations of features predictive of function and dysfunction.

Mind the gap: functional network connectivity interpolation between schizophrenia patients and controls using a variational autoencoder

Mental disorders such as schizophrenia have been challenging to characterize due in part to their heterogeneous presentation in individuals. Most studies have focused on identifying groups differences and have typically ignored the heterogeneous patterns within groups. Here we propose a novel approach based on a variational autoencoder (VAE) to interpolate static functional network connectivity (sFNC) across individuals, with group-specific patterns between schizophrenia patients and controls captured simultaneously. We then visualize the original sFNC in a 2D grid according to the samples in the VAE latent space. We observe a high correspondence between the generated and the original sFNC. The proposed framework facilitates data visualization and can potentially be applied to predict the stage that a subject falls within a disorder continuum as well as characterize individual heterogeneity within and between groups.

Learning Active Multimodal Subspaces in the Brain

Here we introduce a multimodal framework to identify subspaces in the human brain that are defined by collective changes in structural and functional measures and are actively linked to demographic, biological and cognitive indicators in a population. We determine the multimodal subspaces using principles of active subspace learning (ASL) and demonstrate its application on a sample learning task (biological ageing) on a schizophrenia dataset. The proposed multimodal ASL method successfully identifies latent brain representations as subsets of brain regions and connections forming covarying subspaces in association with biological age. We show that schizophrenia is characterized by different subspace patterns compared to those in a cognitively normal brain. The multimodal features generated by projecting structural and functional MRI components onto these active subspaces perform better than several PCA-based transformations and equally well when compared to non-transformed features on the studied learning task. In essence, the proposed method successfully learns active brain subspaces associated with a specific brain condition but inferred from the brain imaging data along with the biological/cognitive traits of interest. Clinical relevance- The work introduces a novel way to create multimodal brain biomarkers based on subspaces computed in association with cognitive or biological traits of interest. These subspaces collectively covary maximally in association with a given trait and successfully retain predictive information.

Spatially Constrained ICA Enables Robust Detection of Schizophrenia from Very Short Resting-state fMRI

Resting-state functional network connectivity (rsFNC) has shown utility for identifying characteristic functional brain patterns in individuals with psychiatric and mood disorders, providing a promising avenue for biomarker development. However, several factors have precluded widespread clinical adoption of rsFNC diagnostics, namely the lack of standardized approaches for capturing comparable and reproducible imaging markers across individuals, as well as the disagreement on the amount of data required to robustly detect intrinsic connectivity networks (ICNs) and diagnostically relevant patterns of rsFNC. Here, we investigate the robustness of (1) subject-specific ICNs standardized to an a priori network template via spatially constrained ICA (scICA), and (2) rsFNC differences between schizophrenia and control groups with respect to the length of the fMRI. Our results suggest clinical rsFMRI scans, when decomposed with scICA, could potentially be shortened to just 2-4 minutes without significant loss of individual rsFNC information or classification performance of longer scan lengths. Clinical Relevance - This work shows diagnostically relevant rsFNC patterns for schizophrenia can be identified from just 2-4 minutes of rsfMRI using an scICA approach. These results can influence future work in neuroimaging biomarker development.

Building Models of Functional Interactions Among Brain Domains that Encode Varying Information Complexity: A Schizophrenia Case Study

Revealing associations among various structural and functional patterns of the brain can yield highly informative results about the healthy and disordered brain. Studies using neuroimaging data have more recently begun to utilize the information within as well as across various functional and anatomical domains (i.e., groups of brain networks). However, most whole-brain approaches assume similar complexity of interactions throughout the brain. Here we investigate the hypothesis that interactions between brain networks capture varying amounts of complexity, and that we can better capture this information by varying the complexity of the model subspace structure based on available training data. To do this, we employ a Bayesian optimization-based framework known as the Tree Parzen Estimator (TPE) to identify, exploit and analyze patterns of variation in the information encoded by temporal information extracted from functional magnetic resonance imaging (fMRI) subdomains of the brain. Using a repeated cross-validation procedure on a schizophrenia classification task, we demonstrate evidence that interactions between specific functional subdomains are better characterized by more sophisticated model architectures compared to less complicated ones required by the others for optimally contributing towards classification and understanding the brain's functional interactions. We show that functional subdomains known to be involved in schizophrenia require more complex architectures to optimally unravel discriminatory information about the disorder. Our study points to the need for adaptive, hierarchical learning frameworks that cater differently to the features from different subdomains, not only for a better prediction but also for enabling the identification of features predicting the outcome of interest.

Path Analysis: A Method to Estimate Altered Pathways in Time-varying Graphs of Neuroimaging Data

Graph-theoretical methods have been widely used to study human brain networks in psychiatric disorders. However, the focus has primarily been on global graphic metrics with little attention to the information contained in paths connecting brain regions. Details of disruption of these paths may be highly informative for understanding disease mechanisms. To detect the absence or addition of multistep paths in the patient group, we provide an algorithm estimating edges that contribute to these paths with reference to the control group. We next examine where pairs of nodes were connected through paths in both groups by using a covariance decomposition method. We apply our method to study resting-state fMRI data in schizophrenia versus controls. Results show several disconnectors in schizophrenia within and between functional domains, particularly within the default mode and cognitive control networks. Additionally, we identify new edges generating additional paths. Moreover, although paths exist in both groups, these paths take unique trajectories and have a significant contribution to the decomposition. The proposed path analysis provides a way to characterize individuals by evaluating changes in paths, rather than just focusing on the pairwise relationships. Our results show promise for identifying path-based metrics in neuroimaging data.

Statelets: Capturing recurrent transient variations in dynamic functional network connectivity

Dynamic functional network connectivity (dFNC) analysis is a widely used approach for capturing brain activation patterns, connectivity states, and network organization. However, a typical sliding window plus clustering (SWC) approach for analyzing dFNC models the system through a fixed sequence of connectivity states. SWC assumes connectivity patterns span throughout the brain, but they are relatively spatially constrained and temporally short‐lived in practice. Thus, SWC is neither designed to capture transient dynamic changes nor heterogeneity across subjects/time. We propose a state‐space time series summarization framework called “statelets” to address these shortcomings. It models functional connectivity dynamics at fine‐grained timescales, adapting time series motifs to changes in connectivity strength, and constructs a concise yet informative representation of the original data that conveys easily comprehensible information about the phenotypes. We leverage the earth mover distance in a nonstandard way to handle scale differences and utilize kernel density estimation to build a probability density profile for local motifs. We apply the framework to study dFNC of patients with schizophrenia (SZ) and healthy control (HC). Results demonstrate SZ subjects exhibit reduced modularity in their brain network organization relative to HC. Statelets in the HC group show an increased recurrence across the dFNC time‐course compared to the SZ. Analyzing the consistency of the connections across time reveals significant differences within visual, sensorimotor, and default mode regions where HC subjects show higher consistency than SZ. The introduced approach also enables handling dynamic information in cross‐modal and multimodal applications to study healthy and disordered brains.

An Approach to Automatically Label and Order Brain Activity/Component Maps

Background: Functional magnetic resonance imaging (fMRI) is a brain imaging technique that provides detailed insights into brain function and its disruption in various brain disorders. The data-driven analysis of fMRI brain activity maps involves several postprocessing steps, the first of which is identifying whether the estimated brain network maps capture signals of interest, for example, intrinsic connectivity networks (ICNs), or artifacts. This is followed by linking the ICNs to standardized anatomical and functional parcellations. Optionally, as in the study of functional network connectivity (FNC), rearranging the connectivity graph is also necessary to facilitate interpretation. Methods: Here we develop a novel and efficient method (Autolabeler) for implementing and integrating all of these processes in a fully automated manner. The Autolabeler method is pretrained on a cross-validated elastic-net regularized general linear model from the noisecloud toolbox to separate neuroscientifically meaningful ICNs from artifacts. It is capable of automatically labeling activity maps with labels from several well-known anatomical and functional parcellations. Subsequently, this method also maximizes the modularity within functional domains to generate a more systematically structured FNC matrix for post hoc network analyses. Results: Results show that our pretrained model achieves 86% accuracy at classifying ICNs from artifacts in an independent validation data set. The automatic anatomical and functional labels also have a high degree of similarity with manual labels selected by human raters. Discussion: At a time of ever-increasing rates of generating brain imaging data and analyzing brain activity, the proposed Autolabeler method is intended to automate such analyses for faster and more reproducible research. Impact statement Our proposed method is capable of implementing and integrating some of the crucial tasks in functional magnetic resonance imaging (fMRI) studies. It is the first to incorporate such tasks without the need for expert intervention. We develop an open-source toolbox for the proposed method that can function as stand-alone software and additionally provides seamless integration with the widely used group independent component analysis for fMRI toolbox (GIFT). This integration can aid investigators to conduct fMRI studies in an end-to-end automated manner.

Tri-Clustering Dynamic Functional Network Connectivity Identifies Significant Schizophrenia Effects Across Multiple States in Distinct Subgroups of Individuals

Background: Brain imaging data collected from individuals are highly complex with unique variation; however, such variation is typically ignored in approaches that focus on group averages or even supervised prediction. State-of-the-art methods for analyzing dynamic functional network connectivity (dFNC) subdivide the entire time course into several (possibly overlapping) connectivity states (i.e., sliding window clusters). However, such an approach does not factor in the homogeneity of underlying data and may result in a less meaningful subgrouping of the data set. Methods: Dynamic-N-way tri-clustering (dNTiC) incorporates a homogeneity benchmark to approximate clusters that provide a more "apples-to-apples" comparison between groups within analogous subsets of time-space and subjects. dNTiC sorts the dFNC states by maximizing similarity across individuals and minimizing variance among the pairs of components within a state. Results: Resulting tri-clusters show significant differences between schizophrenia (SZ) and healthy control (HC) in distinct brain regions. Compared with HC subjects, SZ show hypoconnectivity (low positive) among subcortical, default mode, cognitive control, but hyperconnectivity (high positive) between sensory networks in most tri-clusters. In tri-cluster 3, HC subjects show significantly stronger connectivity among sensory networks and anticorrelation between subcortical and sensory networks than SZ. Results also provide a statistically significant difference in SZ and HC subject's reoccurrence time for two distinct dFNC states. Conclusions: Outcomes emphasize the utility of the proposed method for characterizing and leveraging variance within high-dimensional data to enhance the interpretability and sensitivity of measurements in studying a heterogeneous disorder such as SZ and unconstrained experimental conditions as resting functional magnetic resonance imaging. Impact statement The current methods for analyzing dynamic functional network connectivity (dFNC) run k-means on a collection of dFNC windows, and each window includes all the pairs of independent component analysis networks. As such, it depicts a short-time connectivity pattern of the entire brain, and the k-means clusters fixed-length signatures that have an extent throughout the neural system. Consequently, there is a chance of missing connectivity signatures that span across a smaller subset of pairs. Dynamic-N-way tri-clustering further sorts the dFNC states by maximizing similarity across individuals, minimizing variance among the pairs of components within a state, and reporting more complex and transient patterns.

The association between language-based task-functional magnetic resonance imaging hemodynamics and baseline GABA+ and glutamate-glutamine measured in pre-supplementary motor area: A pilot study in an aging model

Aging is a natural phenomenon that elicits slow and progressive cerebrovascular and neurophysiological changes that eventually lead to cognitive decline. The objective of this pilot study is to examine the association of GABA+ and glutamate-glutamine (Glx) complex with language-based blood oxygen level dependent (BOLD) hemodynamics in an aging model. More specifically, using standard BOLD we will first attempt to validate whether previously reported findings for BOLD amplitude and resting neurochemical relationships hold in an aging model. Secondly, we will investigate how our recently established neurosensitized task-BOLD energetics relate to resting GABA+ and Glx, especially accounting for titration of task difficulty. To support the above endeavors, we optimize the baseline fitting for edited magnetic resonance spectroscopy (MRS) difference spectra to sensitize GABA+ and Glx concentrations to aging-related differences. We identify a spline-knot spacing of 0.6ppm to yield the optimal aging-related differences in GABA+ and Glx. The optimized MRS values were then graduated to relate to task-BOLD hemodynamics. Our results did not replicate previous findings that relate task-BOLD amplitude and resting GABA+ and Glx. However, we did identify neurochemistry relationships with the vascularly-driven dispersion component of the hemodynamic response function, specifically in older participants. In terms of neuro-sensitized BOLD energetics and the underlying role of GABA+ and Glx, our data suggests that the task demands are supported by both neurometabolites depending on the difficulty of the task stimuli. Another novelty is that we developed task-based functional parcellation of pre-SMA using both groups. In sum, we are the first to demonstrate that multimodal task-fMRI and MRS studies are beneficial to improve our understanding of the aging brain physiology, and to set the platform to better inform approaches for clinical care in aging-related neurovascular diseases. We also urge future studies to replicate our findings in a larger population incorporating a lifespan framework.

ENIGMA + COINSTAC: Improving Findability, Accessibility, Interoperability, and Re-usability

The FAIR principles, as applied to clinical and neuroimaging data, reflect the goal of making research products Findable, Accessible, Interoperable, and Reusable. The use of the Collaborative Informatics and Neuroimaging Suite Toolkit for Anonymized Computation (COINSTAC) platform in the Enhancing Neuroimaging Genetics through Meta-Analysis (ENIGMA) consortium combines the technological approach of decentralized analyses with the sociological approach of sharing data. In addition, ENIGMA + COINSTAC provides a platform to facilitate the use of machine-actionable data objects. We first present how ENIGMA and COINSTAC support the FAIR principles, and then showcase their integration with a decentralized meta-analysis of sex differences in negative symptom severity in schizophrenia, and finally present ongoing activities and plans to advance FAIR principles in ENIGMA + COINSTAC. ENIGMA and COINSTAC currently represent efforts toward improved Access, Interoperability, and Reusability. We highlight additional improvements needed in these areas, as well as future connections to other resources for expanded Findability.

Ruminative reflection is associated with anticorrelations between the orbitofrontal cortex and the default mode network in depression: implications for repetitive transcranial magnetic stimulation

Patients with depression who ruminate repeatedly focus on depressive thoughts; however, there are two cognitive subtypes of rumination, reflection and brooding, each associated with different prognoses. Reflection involves problem-solving and is associated with positive outcomes, whereas brooding involves passive, negative, comparison with other people and is associated with poor outcomes. Rumination has also been related to atypical functional hyperconnectivity between the default mode network and subgenual prefrontal cortex. Repetitive pulse transcranial magnetic stimulation of the prefrontal cortex has been shown to alter functional connectivity, suggesting that the abnormal connectivity associated with rumination could potentially be altered. This study examined potential repetitive pulse transcranial magnetic stimulation prefrontal cortical targets that could modulate one or both of these rumination subtypes. Forty-three patients who took part in a trial of repetitive pulse transcranial magnetic stimulation completed the Rumination Response Scale questionnaire and resting-state functional magnetic resonance imaging. Seed to voxel functional connectivity analyses identified an anticorrelation between the left lateral orbitofrontal cortex (-44, 26, -8; k = 172) with the default mode network-subgenual region in relation to higher levels of reflection. Parallel analyses were not significant for brooding or the RRS total score. These findings extend previous studies of rumination and identify a potential mechanistic model for symptom-based neuromodulation of rumination.

Learning Active Multimodal Subspaces in the Brain.. [DOI: 10.1101/2021.12.06.471396] [Citation(s) in RCA: 0] [Impact Index Per Article: 0]

Here we introduce a multimodal framework to identify subspaces in the human brain that are defined by collective changes in structural and functional measures and are actively linked to demographic, biological and cognitive indicators in a population. We determine the multimodal subspaces using principles of active subspace learning (ASL) and demonstrate its application on a sample learning task (biological ageing) on a Schizophrenia dataset. The proposed multimodal ASL method successfully identifies latent brain representations as subsets of brain regions and connections forming co-varying subspaces in association with biological age. We show that Schizophrenia is characterized by different subspace patterns compared to those in a cognitively normal brain. The multimodal features generated by projecting structural and functional MRI components onto these active subspaces perform better than several PCA-based transformations and equally well when compared to non-transformed features on the studied learning task. In essence, the proposed method successfully learns active brain subspaces associated with a specific brain condition but inferred from the brain imaging data along with the biological/cognitive traits of interest.

Three-way parallel group independent component analysis: Fusion of spatial and spatiotemporal magnetic resonance imaging data

Advances in imaging acquisition techniques allow multiple imaging modalities to be collected from the same subject. Each individual modality offers limited yet unique views of the functional, structural, or dynamic temporal features of the brain. Multimodal fusion provides effective ways to leverage these complementary perspectives from multiple modalities. However, the majority of current multimodal fusion approaches involving functional magnetic resonance imaging (fMRI) are limited to 3D feature summaries that do not incorporate its rich temporal information. Thus, we propose a novel three‐way parallel group independent component analysis (pGICA) fusion method that incorporates the first‐level 4D fMRI data (temporal information included) by parallelizing group ICA into parallel ICA via a unified optimization framework. A new variability matrix was defined to capture subject‐wise functional variability and then link it to the mixing matrices of the other two modalities. Simulation results show that the three‐way pGICA provides highly accurate cross‐modality linkage estimation under both weakly and strongly correlated conditions, as well as comparable source estimation under different noise levels. Results using real brain imaging data identified one linked functional–structural–diffusion component associated to differences between schizophrenia and controls. This was replicated in an independent cohort, and the identified components were also correlated with major cognitive domains. Functional network connectivity revealed visual–subcortical and default mode‐cerebellum pairs that discriminate between schizophrenia and controls. Overall, both simulation and real data results support the use of three‐way pGICA to identify multimodal spatiotemporal links and to pursue the study of brain disorders under a single unifying multimodal framework.

Multi-modal deep learning of functional and structural neuroimaging and genomic data to predict mental illness

Neuropsychiatric disorders such as schizophrenia are very heterogeneous in nature and typically diagnosed using self-reported symptoms. This makes it difficult to pose a confident prediction on the cases and does not provide insight into the underlying neural and biological mechanisms of these disorders. Combining neuroimaging and genomic data with a multi-modal 'predictome' paves the way for biologically informed markers and may improve prediction reliability. With that, we develop a multi-modal deep learning framework by fusing data from different modalities to capture the interaction between the latent features and evaluate their complementary information in characterizing schizophrenia. Our deep model uses structural MRI, functional MRI, and genome-wide polymorphism data to perform the classification task. It includes a multi-layer feed-forward network, an encoder, and a long short-term memory (LSTM) unit with attention to learn the latent features and adopt a joint training scheme capturing synergies between the modalities. The hybrid network also uses different regularizers for addressing the inherent overfitting and modality-specific bias in the multi-modal setup. Next, we run the network through a saliency model to analyze the learned features. Integrating modalities enhances the performance of the classifier, and our framework acquired 88% (P < 0.0001) accuracy on a dataset of 437 subjects. The trimodal accuracy is comparable to the state-of-the-art performance on a data collection of this size and outperforms the unimodal and bimodal baselines we compared. Model introspection was used to expose the salient neural features and genes/biological pathways associated with schizophrenia. To our best knowledge, this is the first approach that fuses genomic information with structural and functional MRI biomarkers for predicting schizophrenia. We believe this type of modality blending can better explain the disorder's dynamics by adding cross-modal prospects.Clinical Relevance- This study combinedly learns imaging and genomic features for the classification of schizophrenia. The data fusion scheme extracts modality interactions, and the saliency experiments report multiple functional and structural networks closely connected to the disorder.

Converting scores between the PANSS and SAPS/SANS beyond the positive/negative dichotomy

Previous work provided conversion equations for overall indices of positive and negative symptomatology between the Positive and Negative Syndrome Scale (PANSS) and the Scales for the Assessment of Positive/Negative Symptoms (SAPS/SANS). Our objective was to provide such conversion equations for subdomains of positive and negative symptomatology in order to better account for the diversity of symptom profiles in schizophrenia. Symptoms severity was assessed using both the PANSS and SAPS/SANS in 205 patients with schizophrenia. Two exploratory factor analyses combining items from both scales were first performed separately in the positive and negative symptom domains. Positive factors were termed 'Hallucinations', 'Delusions' and 'Disorganization', while negative factors were associated with 'Expressivity', 'Amotivation' and 'Cognition', consistent with current descriptions of symptom dimensions in schizophrenia. For each factor, linear regression analyses were conducted on 80% of the data to obtain conversion equations from the PANSS to the SAPS/SANS and vice versa. Reliability was then evaluated on the 20% remaining data, with good to excellent intra-class correlation coefficients between the original and predicted scores for all but the cognition factor. These findings show that symptom severity scores can be converted with good accuracy between clinical scales beyond the positive/negative symptom dichotomy.

Quantity and quality: Normative open-access neuroimaging databases

The focus of this article is to compare twenty normative and open-access neuroimaging databases based on quantitative measures of image quality, namely, signal-to-noise (SNR) and contrast-to-noise ratios (CNR). We further the analysis through discussing to what extent these databases can be used for the visualization of deeper regions of the brain, such as the subcortex, as well as provide an overview of the types of inferences that can be drawn. A quantitative comparison of contrasts including T1-weighted (T1w) and T2-weighted (T2w) images are summarized, providing evidence for the benefit of ultra-high field MRI. Our analysis suggests a decline in SNR in the caudate nuclei with increasing age, in T1w, T2w, qT1 and qT2* contrasts, potentially indicative of complex structural age-dependent changes. A similar decline was found in the corpus callosum of the T1w, qT1 and qT2* contrasts, though this relationship is not as extensive as within the caudate nuclei. These declines were accompanied by a declining CNR over age in all image contrasts. A positive correlation was found between scan time and the estimated SNR as well as a negative correlation between scan time and spatial resolution. Image quality as well as the number and types of contrasts acquired by these databases are important factors to take into account when selecting structural data for reuse. This article highlights the opportunities and pitfalls associated with sampling existing databases, and provides a quantitative backing for their usage.

A Deep Learning Model for Data-Driven Discovery of Functional Connectivity

Functional connectivity (FC) studies have demonstrated the overarching value of studying the brain and its disorders through the undirected weighted graph of functional magnetic resonance imaging (fMRI) correlation matrix. However, most of the work with the FC depends on the way the connectivity is computed, and it further depends on the manual post-hoc analysis of the FC matrices. In this work, we propose a deep learning architecture BrainGNN that learns the connectivity structure as part of learning to classify subjects. It simultaneously applies a graphical neural network to this learned graph and learns to select a sparse subset of brain regions important to the prediction task. We demonstrate that the model’s state-of-the-art classification performance on a schizophrenia fMRI dataset and demonstrate how introspection leads to disorder relevant findings. The graphs that are learned by the model exhibit strong class discrimination and the sparse subset of relevant regions are consistent with the schizophrenia literature.

Multiple overlapping dynamic patterns of the visual sensory network in schizophrenia

Although visual processing impairments have been explored in schizophrenia (SZ), their underlying neurobiology of the visual processing impairments has not been widely studied. Also, while some research has hinted at differences in information transfer and flow in SZ, there are few investigations of the dynamics of functional connectivity within visual networks. In this study, we analyzed resting-state fMRI data of the visual sensory network (VSN) in 160 healthy control (HC) subjects and 151 SZ subjects. We estimated 9 independent components within the VSN. Then, we calculated the dynamic functional network connectivity (dFNC) using the Pearson correlation. Next, using k-means clustering, we partitioned the dFNCs into five distinct states, and then we calculated the portion of time each subject spent in each state, which we termed the occupancy rate (OCR). Using OCR, we compared HC with SZ subjects and investigated the link between OCR and visual learning in SZ subjects. Besides, we compared the VSN functional connectivity of SZ and HC subjects in each state. We found that this network is indeed highly dynamic. Each state represents a unique connectivity pattern of fluctuations in VSN FNC, and all states showed significant disruption in SZ. Overall, HC showed stronger connectivity within the VSN in states. SZ subjects spent more time in a state in which the connectivity between the middle temporal gyrus and other regions of VNS is highly negative. Besides, OCR in a state with strong positive connectivity between the middle temporal gyrus and other regions correlated significantly with visual learning scores in SZ.

Application of deep learning in detecting neurological disorders from magnetic resonance images: a survey on the detection of Alzheimer's disease, Parkinson's disease and schizophrenia

Neuroimaging, in particular magnetic resonance imaging (MRI), has been playing an important role in understanding brain functionalities and its disorders during the last couple of decades. These cutting-edge MRI scans, supported by high-performance computational tools and novel ML techniques, have opened up possibilities to unprecedentedly identify neurological disorders. However, similarities in disease phenotypes make it very difficult to detect such disorders accurately from the acquired neuroimaging data. This article critically examines and compares performances of the existing deep learning (DL)-based methods to detect neurological disorders-focusing on Alzheimer's disease, Parkinson's disease and schizophrenia-from MRI data acquired using different modalities including functional and structural MRI. The comparative performance analysis of various DL architectures across different disorders and imaging modalities suggests that the Convolutional Neural Network outperforms other methods in detecting neurological disorders. Towards the end, a number of current research challenges are indicated and some possible future research directions are provided.

Meta-Modal Information Flow: A Method for Capturing Multimodal Modular Disconnectivity in Schizophrenia

OBJECTIVE

Multimodal measurements of the same phenomena provide complementary information and highlight different perspectives, albeit each with their own limitations. A focus on a single modality may lead to incorrect inferences, which is especially important when a studied phenomenon is a disease. In this paper, we introduce a method that takes advantage of multimodal data in addressing the hypotheses of disconnectivity and dysfunction within schizophrenia (SZ).

METHODS

We start with estimating and visualizing links within and among extracted multimodal data features using a Gaussian graphical model (GGM). We then propose a modularity-based method that can be applied to the GGM to identify links that are associated with mental illness across a multimodal data set. Through simulation and real data, we show our approach reveals important information about disease-related network disruptions that are missed with a focus on a single modality. We use functional MRI (fMRI), diffusion MRI (dMRI), and structural MRI (sMRI) to compute the fractional amplitude of low frequency fluctuations (fALFF), fractional anisotropy (FA), and gray matter (GM) concentration maps. These three modalities are analyzed using our modularity method.

RESULTS

Our results show missing links that are only captured by the cross-modal information that may play an important role in disconnectivity between the components.

CONCLUSION

We identified multimodal (fALFF, FA and GM) disconnectivity in the default mode network area in patients with SZ, which would not have been detectable in a single modality.

SIGNIFICANCE

The proposed approach provides an important new tool for capturing information that is distributed among multiple imaging modalities.

Covarying structural alterations in laterality of the temporal lobe in schizophrenia: A case for source-based laterality

The human brain is asymmetrically lateralized for certain functions (such as language processing) to regions in one hemisphere relative to the other. Asymmetries are measured with a laterality index (LI). However, traditional LI measures are limited by a lack of consensus on metrics used for its calculation. To address this limitation, source-based laterality (SBL) leverages an independent component analysis for the identification of laterality-specific alterations, identifying covarying components between hemispheres across subjects. SBL is successfully implemented with simulated data with inherent differences in laterality. SBL is then compared with a voxel-wise analysis utilizing structural data from a sample of patients with schizophrenia and controls without schizophrenia. SBL group comparisons identified three distinct temporal regions and one cerebellar region with significantly altered laterality in patients with schizophrenia relative to controls. Previous work highlights reductions in laterality (ie, reduced left gray matter volume) in patients with schizophrenia compared with controls without schizophrenia. Results from this pilot SBL project are the first, to our knowledge, to identify covarying laterality differences within discrete temporal brain regions. The authors argue SBL provides a unique focus to detect covarying laterality differences in patients with schizophrenia, facilitating the discovery of laterality aspects undetected in previous work.

Correcting Task fMRI Signals for Variability in Baseline CBF Improves BOLD-Behavior Relationships: A Feasibility Study in an Aging Model

Blood Oxygen Level Dependent (BOLD) functional MRI is a complex neurovascular signal whose magnitude depends on baseline physiological factors such as cerebral blood flow (CBF). Because baseline CBF varies across the brain and is altered with aging, the interpretation of stand-alone aging-related BOLD changes can be misleading. The primary objective of this study was to develop a methodology that combines task fMRI and arterial spin labeling (ASL) techniques to sensitize task-induced BOLD activity by covarying out the baseline physiology (i.e., CBF) in an aging model. We recruited 11 younger and 13 older healthy participants who underwent ASL and an overt language fMRI task (semantic category member generation). We measured in-scanner language performance to investigate the effect of BOLD sensitization on BOLD-behavior relationships. The results demonstrate that our correction approach is effective at enhancing the specificity and sensitivity of the BOLD signal in both groups. In addition, the correction strengthens the statistical association between task BOLD activity and behavioral performance. Although CBF has inherent age dependence, our results show that retaining the age factor within CBF aides in greater sensitization of task fMRI signals. From a cognitive standpoint, compared to young adults, the older participants showed a delayed domain-general language-related task activity possibly due to compromised vessel compliance. Further, assessment of functional evolution of corrected BOLD activity revealed biphasic BOLD dynamics in both groups where BOLD deactivation may reflect greater semantic demand or increased premium on domain general executive functioning in response to task difficulty. Although it was promising to note that the predictability of behavior using the proposed methodology outperforms other methodologies (i.e., no correction and normalization by division), and provides moderate stability and adequate power, further work with a larger cohort and other task designs is necessary to improve the stability of predicting associated behavior. In summary, we recommend correction of task fMRI signals by covarying out baseline CBF especially when comparing groups with different neurovascular properties. Given that ASL and BOLD fMRI are well established and widely employed techniques, our proposed multi-modal methodology can be readily implemented into data processing pipelines to obtain more accurate BOLD activation maps.

Efficient Regularized Field Map Estimation in 3D MRI

Magnetic field inhomogeneity estimation is important in some types of magnetic resonance imaging (MRI), including field-corrected reconstruction for fast MRI with long readout times, and chemical shift based water-fat imaging. Regularized field map estimation methods that account for phase wrapping and noise involve nonconvex cost functions that require iterative algorithms. Most existing minimization techniques were computationally or memory intensive for 3D datasets, and are designed for single-coil MRI. This paper considers 3D MRI with optional consideration of coil sensitivity, and addresses the multi-echo field map estimation and water-fat imaging problem. Our efficient algorithm uses a preconditioned nonlinear conjugate gradient method based on an incomplete Cholesky factorization of the Hessian of the cost function, along with a monotonic line search. Numerical experiments show the computational advantage of the proposed algorithm over state-of-the-art methods with similar memory requirements.

A decade of test-retest reliability of functional connectivity: A systematic review and meta-analysis

BACKGROUND

Once considered mere noise, fMRI-based functional connectivity has become a major neuroscience tool in part due to early studies demonstrating its reliability. These fundamental studies revealed only the tip of the iceberg; over the past decade, many test-retest reliability studies have continued to add nuance to our understanding of this complex topic. A summary of these diverse and at times contradictory perspectives is needed.

OBJECTIVES

We aimed to summarize the existing knowledge regarding test-retest reliability of functional connectivity at the most basic unit of analysis: the individual edge level. This entailed (1) a meta-analytic estimate of reliability and (2) a review of factors influencing reliability.

METHODS

A search of Scopus was conducted to identify studies that estimated edge-level test-retest reliability. To facilitate comparisons across studies, eligibility was restricted to studies measuring reliability via the intraclass correlation coefficient (ICC). The meta-analysis included a random effects pooled estimate of mean edge-level ICC, with studies nested within datasets. The review included a narrative summary of factors influencing edge-level ICC.

RESULTS

From an initial pool of 212 studies, 44 studies were identified for the qualitative review and 25 studies for quantitative meta-analysis. On average, individual edges exhibited a "poor" ICC of 0.29 (95% CI = 0.23 to 0.36). The most reliable measurements tended to involve: (1) stronger, within-network, cortical edges, (2) eyes open, awake, and active recordings, (3) more within-subject data, (4) shorter test-retest intervals, (5) no artifact correction (likely due in part to reliable artifact), and (6) full correlation-based connectivity with shrinkage.

CONCLUSION

This study represents the first meta-analysis and systematic review investigating test-retest reliability of edge-level functional connectivity. Key findings suggest there is room for improvement, but care should be taken to avoid promoting reliability at the expense of validity. By pooling existing knowledge regarding this key facet of accuracy, this study supports broader efforts to improve inferences in the field.

Identifying schizophrenia subgroups using clustering and supervised learning

Schizophrenia has a 1% incidence rate world-wide and those diagnosed present with positive (e.g. hallucinations, delusions), negative (e.g. apathy, asociality), and cognitive symptoms. However, both symptom burden and associated brain alterations are highly heterogeneous and intimately linked to prognosis. In this study, we present a method to predict individual symptom profiles by first deriving clinical subgroups and then using machine learning methods to perform subject-level classification based on magnetic resonance imaging (MRI) derived neuroanatomical measures. Symptomatic and MRI data of 167 subjects were used. Subgroups were defined using hierarchical clustering of clinical data resulting in 3 stable clusters: 1) high symptom burden, 2) predominantly positive symptom burden, and 3) mild symptom burden. Cortical thickness estimates were obtained in 78 regions of interest and were input, along with demographic data, into three machine learning models (logistic regression, support vector machine, and random forest) to predict subgroups. Random forest performance metrics for predicting the group membership of the high and mild symptom burden groups exceeded those of the baseline comparison of the entire schizophrenia population versus normal controls (AUC: 0.81 and 0.78 vs. 0.75). Additionally, an analysis of the most important features in the random forest classification demonstrated consistencies with previous findings of regional impairments and symptoms of schizophrenia.

Knowing What You Know in Brain Segmentation Using Bayesian Deep Neural Networks

In this paper, we describe a Bayesian deep neural network (DNN) for predicting FreeSurfer segmentations of structural MRI volumes, in minutes rather than hours. The network was trained and evaluated on a large dataset (n = 11,480), obtained by combining data from more than a hundred different sites, and also evaluated on another completely held-out dataset (n = 418). The network was trained using a novel spike-and-slab dropout-based variational inference approach. We show that, on these datasets, the proposed Bayesian DNN outperforms previously proposed methods, in terms of the similarity between the segmentation predictions and the FreeSurfer labels, and the usefulness of the estimate uncertainty of these predictions. In particular, we demonstrated that the prediction uncertainty of this network at each voxel is a good indicator of whether the network has made an error and that the uncertainty across the whole brain can predict the manual quality control ratings of a scan. The proposed Bayesian DNN method should be applicable to any new network architecture for addressing the segmentation problem.