Search

Background Depressive episodes (DEPs), characterized by abnormalities in cognitive functions and mood, are a leading cause of disability. Electroconvulsive therapy (ECT), which involves a brief electrical stimulation of the anesthetized brain, is one of the most effective treatments used in patients with DEP due to its rapid efficacy. Methods In this work, we investigated how dynamic brain functional connectivity responds to ECT and whether the dynamic responses are associated with treatment outcomes and side effects in patients. We applied a fully automated independent component analysis–based pipeline to 110 patients with DEP (including diagnosis of unipolar depression or bipolar depression) and 60 healthy control subjects. The dynamic functional connectivity was analyzed by a combination of the sliding window approach and clustering analysis. Results Five recurring connectivity states were identified, and patients with DEPs had fewer occurrences in one brain state (state 1) with strong positive and negative connectivity. Patients with DEP changed the occupancy of two states (states 3 and 4) after ECT, resulting in significantly different occurrences of one additional state (state 3) compared with healthy control subjects. We further found that patients with DEP had diminished global metastate dynamism, two of which recovered to normal after ECT. The changes in dynamic connectivity characteristics were associated with the changes in memory recall and Hamilton Depression Rating Scale of DEP after ECT. Conclusions These converging results extend current findings on subcortical-cortical dysfunction and dysrhythmia in DEP and demonstrate that ECT might cause remodeling of brain functional dynamics that enhance the neuroplasticity of the diseased brain.

Inferring dynamic functional connectivity (dFC) from functional magnetic resonance imaging (fMRI) is crucial to understand the time-variant functional inter-relationships among brain regions. Because of the sparse property of functional connectivity networks, sparsity-promoting dFC estimation methods, which are mainly based on l 1 -norm regularization, are gaining popularity. However, l 1 -norm regularization cannot provide the maximum sparsity solution as the most natural sparsity promoting norm, the l 0 -norm. But l 0 -norm is seldom used to infer sparse dFC because an efficient algorithm to address the non-convexity problem of l 0 -norm is lacking. In this work, we develop a new l 0 -norm regularization-based inverse covariance estimation method for estimating dFC from fMRI. This novel method employs l 0 -norm regularizations on both spatial and temporal scales to enhance the spatial sparsity and temporal smoothness of dFC estimates. To overcome the non-convexity of l 0 -norm, we further propose an effective optimization algorithm based on the coordinate descent (CD). The performance of the proposed l 0 -norm-based sparse-smooth regularization (L0-SSR) method is examined using a series of synthetic datasets concerning various types of network topology. We further apply the proposed L0-SSR method to real fMRI data recorded in block-design motor tasks from 45 participants for the exploration of task induced dFC. Results on synthetic and real-world fMRI data show that, the L0-SSR method can achieve more accurate and interpretable dFC estimates than conventional l 1 -norm-based dFC estimation methods. Hence, the proposed L0-SSR method could serve as a powerful analytical tool to infer highly complex, variable, and sparse dFC patterns.

Background: Alzheimer's disease (AD) is a chronic neurodegenerative disorder frequently accompanied by cerebral small vessel disease (CSVD). However, the influence of CSVD on the brain functional connectivity in subjects along the AD continuum is still largely unknown. The current study combined the static and dynamic functional network connectivity (FNC) to explore the underlying mechanism. Materials and Methods: In this study, we included 182 healthy controls, 27 individuals with subjective cognitive decline (SCD), 27 with SCD+CSVD, 104 with mild cognitive impairment (MCI), 123 with MCI+CSVD, 16 with AD, and 62 with AD+CSVD. We examined the static and dynamic FNC within the default mode, salience, and cognitive control domains. We also assessed the association between atypical FNC patterns and cognitive impairments, as well as the pathologies. Results: Static FNC results showed progressively increased within-domain connectivity and decreased between-domain connectivity along the AD continuum, especially in CSVD subjects. Dynamic FNC in CSVD subjects showed more occurrences in a highly modularized state and fewer occurrences in the diffusely connected state. Further analysis showed that neuropathology and CSVD burden divergently affect the FNC changes. Conclusions: The overall results demonstrate divergent abnormalities of FNC in CSVD and non-CSVD individuals along the AD continuum, which were divergently affected by neuropathology and CSVD burden. Specifically, those with CSVD show more static and dynamic FNC impairments, associated with cognitive decline. These findings may advance our understanding of the effect of CSVD on AD onset and progression, and provide potential hints for clinical treatment. IMPACT STATEMENT: By combining the static and dynamic brain functional network connectivity (FNC), our study explored the effect of vascular pathology on FNC in subjects along the Alzheimer's disease (AD) continuum. Results demonstrate divergent FNC abnormalities in individuals with/without cerebral small vessel disease (CSVD) along the AD continuum, which could be divergently affected by neuropathology (amyloid and tau) and CSVD burden. Notably, AD-continuum subjects with CSVD show more static and dynamic FNC impairments, associated with cognitive decline. Conclusively, these findings may advance our understanding of the effect of vascular pathology on AD onset and progression, and provide potential hints for clinical treatment.

BackgroundThe heterogeneity inherent in autism spectrum disorder (ASD) presents a substantial challenge to diagnosis and precision treatment. Heterogeneity across biological etiologies, genetics, neural systems, neurocognitive attributes and clinical subtypes or phenotypes has been observed across individuals with ASD.MethodsIn this study, we aim to investigate the heterogeneity in ASD from a multimodal brain imaging perspective. The Autism Diagnostic Observation Schedule (ADOS) was used as a reference to guide functional and structural MRI fusion. DSM-IV-TR diagnosed Asperger’s disorder (n = 79), pervasive developmental disorder-not otherwise specified [PDD-NOS] (n = 58) and Autistic disorder (n = 92) from ABIDE II were used as discovery cohort, and ABIDE I (n = 400) was used for replication.ResultsDorsolateral prefrontal cortex and superior/middle temporal cortex are the primary common functional–structural covarying cortical brain areas shared among Asperger’s, PDD-NOS and Autistic subgroups. Key differences among the three subtypes are negative functional features within subcortical brain areas, including negative putamen–parahippocampus fractional amplitude of low-frequency fluctuations (fALFF) unique to the Asperger’s subtype; negative fALFF in anterior cingulate cortex unique to PDD-NOS subtype; and negative thalamus–amygdala–caudate fALFF unique to the Autistic subtype. Furthermore, each subtype-specific brain pattern is correlated with different ADOS subdomains, with social interaction as the common subdomain. The identified subtype-specific patterns are only predictive for ASD symptoms manifested in the corresponding subtypes, but not the other subtypes.ConclusionsAlthough ASD has a common neural basis with core deficits linked to social interaction, each ASD subtype is strongly linked to unique brain systems and subdomain symptoms, which may help to better understand the underlying mechanisms of ASD heterogeneity from a multimodal neuroimaging perspective.LimitationsThis study is male based, which cannot be generalized to the female or the general ASD population.

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.

Schizophrenia (SZ) is frequently concurrent with substance use, depressive symptoms, social communication and attention deficits. However, the relationship between common brain networks (e.g., SZ vs. substance use, SZ vs. depression, SZ vs. developmental disorders) with SZ on specific symptoms and cognition is unclear. Symptom scores were used as a reference to guide fMRI-sMRI fusion for SZ (n = 94), substance use with drinking (n = 313), smoking (n = 104), major depressive disorder (MDD, n = 260), developmental disorders with autism spectrum disorder (ASD, n = 421) and attention-deficit/hyperactivity disorder (ADHD, n = 244) respectively. Common brain regions were determined by overlapping the symptom-related components between SZ and these other groups. Correlation between the identified common brain regions and cognition/symptoms in an independent SZ dataset (n = 144) was also performed. Results show that (1): substance use was related with cognitive deficits in schizophrenia through gray matter volume (GMV) in anterior cingulate cortex and thalamus; (2) depression was linked to PANSS negative dimensions and reasoning in SZ through a network involving caudate-thalamus-middle/inferior temporal gyrus in GMV; (3) developmental disorders pattern was correlated with poor attention, speed of processing and reasoning in SZ through inferior temporal gyrus in GMV. This study reveals symptom driven transdiagnostic shared networks between SZ and other mental disorders via multi-group data mining, indicating that some potential common underlying brain networks associated with schizophrenia differently with respect to symptoms and cognition. These results have heuristic value and advocate specific approaches to refine available treatment strategies for comorbid conditions in schizophrenia.

Functional connectivity (FC) studies have demonstrated the overarching value of studying the brain and its disorders through the undirected weighted graph of fMRI correlation matrix. Most of the work with the FC, however, depends on the way the connectivity is computed, and 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 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 learned by the model exhibit strong class discrimination and the sparse subset of relevant regions are consistent with the schizophrenia literature., Accepted at Algorithms 2021, 14(3), 75

BackgroundLate-onset Alzheimer’s disease (AD) is a polygenic neurodegenerative disease. Identifying the neuroimaging phenotypes behind the genetic predisposition of AD is critical to the understanding of AD pathogenesis. Two major questions which previous studies have led to are: (1) should the general “polygenic hazard score” (PHS) be a good choice to identify the individual genetic risk for AD; and (2) should researchers also include inter-modality relationships in the analyses considering these may provide complementary information about the AD etiology.MethodsWe collected 88 healthy controls, 77 patients with mild cognitive impairment (MCI), and 22 AD patients to simulate the AD continuum included from the ADNI database. PHS-guided multimodal fusion was used to investigate the impact of PHS on multimodal brain networks in AD-continuum by maximizing both inter-modality association and reference-modality correlation. Fractional amplitude of low frequency fluctuations, gray matter (GM) volume, and amyloid standard uptake value ratios were included as neuroimaging features. Eventually, the changes in neuroimaging features along AD continuum were investigated, and relationships between cognitive performance and identified PHS associated multimodal components were established.ResultsWe found that PHS was associated with multimodal brain networks, which showed different functional and structural impairments under increased amyloid deposits. Notably, along with AD progression, functional impairment occurred before GM atrophy, amyloid deposition started from the MCI stage and progressively increased throughout the disease continuum.ConclusionPHS is associated with multi-facets of brain impairments along the AD continuum, including cognitive dysfunction, pathological deposition, which might underpin the AD pathogenesis.

Psychosis disorders share overlapping symptoms and are characterized by a wide-spread breakdown in functional brain integration. Although neuroimaging studies have identified numerous connectivity abnormalities in affective and non-affective psychoses, whether they have specific or unique connectivity abnormalities, especially within the early stage is still poorly understood. The early phase of psychosis is a critical period with fewer chronic confounds and when treatment intervention may be most effective. In this work, we examined whole-brain functional network connectivity (FNC) from both static and dynamic perspectives in patients with affective psychosis (PAP) or with non-affective psychosis (PnAP) and healthy controls (HCs). A fully automated independent component analysis (ICA) pipeline called “Neuromark” was applied to high-quality functional magnetic resonance imaging (fMRI) data with 113 early-phase psychosis patients (32 PAP and 81 PnAP) and 52 HCs. Relative to the HCs, both psychosis groups showed common abnormalities in static FNC (sFNC) between the thalamus and sensorimotor domain, and between subcortical regions and the cerebellum. PAP had specifically decreased sFNC between the superior temporal gyrus and the paracentral lobule, and between the cerebellum and the middle temporal gyrus/inferior parietal lobule. On the other hand, PnAP showed increased sFNC between the fusiform gyrus and the superior medial frontal gyrus. Dynamic FNC (dFNC) was investigated using a combination of a sliding window approach, clustering analysis, and graph analysis. Three reoccurring brain states were identified, among which both psychosis groups had fewer occurrences in one antagonism state (state 2) and showed decreased network efficiency within an intermediate state (state 1). Compared with HCs and PnAP, PAP also showed a significantly increased number of state transitions, indicating more unstable brain connections in affective psychosis. We further found that the identified connectivity features were associated with the overall positive and negative syndrome scale, an assessment instrument for general psychopathology and positive symptoms. Our findings support the view that subcortical-cortical information processing is disrupted within five years of the initial onset of psychosis and provide new evidence that abnormalities in both static and dynamic connectivity consist of shared and unique features for the early affective and non-affective psychoses.

There is growing evidence that rather than using a single brain imaging modality to study its association with physiological or symptomatic features, the field is paying more attention to fusion of multimodal information. However, most current multimodal fusion approaches that incorporate functional magnetic resonance imaging (fMRI) are restricted to second-level 3D features, rather than the original 4D fMRI data. This trade-off is that the valuable temporal information is not utilized during the fusion step. Here we are motivated to propose a novel approach called "parallel group ICA+ICA" that incorporates temporal fMRI information from group independent component analysis (GICA) into a parallel independent component analysis (ICA) framework, aiming to enable direct fusion of first-level fMRI features with other modalities (e.g., structural MRI), which thus can detect linked functional network variability and structural covariations. Simulation results show that the proposed method yields accurate intermodality linkage detection regardless of whether it is strong or weak. When applied to real data, we identified one pair of significantly associated fMRI-sMRI components that show group difference between schizophrenia and controls in both modalities, and this linkage can be replicated in an independent cohort. Finally, multiple cognitive domain scores can be predicted by the features identified in the linked component pair by our proposed method. We also show these multimodal brain features can predict multiple cognitive scores in an independent cohort. Overall, results demonstrate the ability of parallel GICA+ICA to estimate joint information from 4D and 3D data without discarding much of the available information up front, and the potential for using this approach to identify imaging biomarkers to study brain disorders.

ObjectiveTo investigate the dynamic functional connectivity of thalamocortical networks in interictal migraine patients and whether clinical features are associated with abnormal connectivity.MethodsWe investigated dynamic functional network connectivity (dFNC) of the migraine brain in 89 interictal migraine patients and 70 healthy controls. We focused on the temporal properties of thalamocortical connectivity using sliding window cross-correlation, clustering state analysis, and graph-theory methods. Relationships between clinical symptoms and abnormal dFNC were evaluated using a multivariate linear regression model.ResultsFive dFNC brain states were identified to characterize and compare dynamic functional connectivity patterns. We demonstrated that migraineurs spent more time in a strongly interconnected between-network state, but they spent less time in a sparsely connected state. Interestingly, we found that abnormal posterior thalamus (pulvinar nucleus) dFNC with the visual cortex and the precuneus were significantly correlated with headache frequency of migraine. Further topologic measures revealed that migraineurs had significantly lower efficiency of information transfer in both global and local dFNC.ConclusionOur results demonstrated a transient pathologic state with atypical thalamocortical connectivity in migraineurs and extended current findings regarding abnormal thalamocortical networks and dysrhythmia in migraine.

Approaches involving the use of learning architectures on multimodal neuroimaging data tend to assume uniformity in the way information is stored in various sub-domains of the brain, thus not catering to the differences across functional and structural sub-domains. We introduce a learning framework to effectively incorporate multimodal features using structural and functional MRI data from a dataset of schizophrenia patients and controls, accounting for and exploiting the heterogeneity in the sub-domains of the brain. We analyze these sub-domains in terms of their functional interactions (i.e. within and between network connectivity) and structural properties (gray matter volume). By using Bayesian optimization on a search space of flexible multimodal architectures with multiple branches, we demonstrate that the discriminatory information from structural and functional sub-domains can be better recovered if the complexity of subspace structure in the model can be tuned to reflect the extent of non-linearity with which each sub-domain encodes the information. Our repeated cross-validated results from a schizophrenia classification problem show that for better classification and interpretation, sub-domains known for their role or disruption in Schizophrenia require more sophisticated subspace structure in the model compared to others. Our work emphasizes on the requirement to create multimodal frameworks that can adapt based on differences in the way various sub-domains of the brain encode discriminatory information. This is important to not only have better-performing prediction models but also to reveal sub-domains associated with the outcome at hand.

BackgroundElectroconvulsive Therapy (ECT) is one of the most effective treatments for major depressive disorder (DEP). There is recently increasing attention to evaluate ECT’s effect on resting-state functional magnetic resonance imaging (rs-fMRI). This study aims to compare rs-fMRI of DEP patients with healthy participants, investigate whether dynamic functional network connectivity network (dFNC) estimated from rs-fMRI predicts the ECT outcome, and explore the effect of ECT on brain network states.MethodResting-state fMRI data were collected from 119 patients with depression or DEP (76 females), and 61 Healthy (HC) participants (34 females) with an age mean of 52.25 (N=180) years old. The pre-ECT and post-ECT Hamilton Depression Rating Scale (HDRS) were 25.59±6.14 and 11.48±9.07, respectively. Twenty-four independent components from default mode (DMN) and cognitive control network (CCN) were extracted using group-independent component analysis from pre-ECT and post-ECT rs-fMRI. Then, the sliding window approach was used to estimate the pre-and post-ECT dFNC of each participant. Next, k-means clustering was separately applied to pre-ECT dFNC and post-ECT dFNC to assess three distinct states from each participant. We calculated the amount of time each individual spends in each state, called occupancy rate or OCR. Next, we compared OCR values between HC and DEP participants. We also calculated the partial correlation between pre-ECT OCRs and HDRS change while controlling for age, gender, number of treatment, and site. Finally, we evaluated the effectiveness of ECT by comparing pre-and post-ECT OCR of DEP and HC participants.ResultsThe main findings include: 1) DEP patients had significantly lower OCR values than the HC group in a state, where connectivity between CCN and DMN was relatively higher than other states (corrected p= 0.015), 2) Pre-ECT OCR of state, with more negative connectivity between CCN and DMN components, predicted the HDRS changes (R=0.23 corrected p=0.03). This means that those DEP patients who spend less time in this state showed more HDRS change, and 3) The post-ECT OCR analysis suggested that ECT increased the amount of time DEP patients spend in state 2 (corrected p=0.03). Finally, we found ECT increases the total traveled distance in DEP.ConclusionOur finding suggests that dFNC features, estimated from CCN and DMN, show promise as a predictive biomarker of the ECT outcome of DEP patients. Also, this study identified a possible underlying mechanism associated with the ECT effect in DEP patients.

Individuals can be characterized in a population according to their brain measurements and activity, given the inter-subject variability in brain anatomy, structure-function relationships, or life experience. Many neuroimaging studies have demonstrated the potential of functional network connectivity patterns estimated from resting functional magnetic resonance imaging (fMRI) to discriminate groups and predict information about individual subjects. However, the predictive signal present in the spatial heterogeneity of brain connectivity networks is yet to be extensively studied. In this study, we investigate, for the first time, the use of pairwise-relationships between resting-state independentspatial mapsto characterize individuals. To do this, we develop a deep Siamese framework comprising three-dimensional convolution neural networks for contrastive learning based on individual-level spatial maps estimated via a fully automated fMRI independent component analysis approach. The proposed framework evaluates whether pairs of spatial networks (e.g., visual network and auditory network) are capable of subject identification and assesses the spatial variability in different network pairs’ predictive power in an extensive whole-brain analysis. Our analysis on nearly 12,000 unaffected individuals from the UK Biobank study demonstrates that the proposed approach can discriminate subjects with an accuracy of up to 88% for a single network pair on the test set (best model, after several runs), and 82% average accuracy at the subcortical domain level, notably the highest average domain level accuracy attained. Further investigation of our network’s learned features revealed a higher spatial variability in predictive accuracy among younger brains and significantly higher discriminative power among males. In sum, the relationship among spatial networks appears to be both informative and discriminative of individuals and should be studied further as putative brain-based biomarkers.

We introduce an extension of independent component analysis (ICA), called multiscale ICA (msICA), and design an approach to capture dynamic functional source interactions within and between multiple spatial scales. msICA estimates functional sources at multiple spatial scales without imposing direct constraints on the size of functional sources, overcomes the limitation of using fixed anatomical locations, and eliminates the need for model-order selection in ICA analysis. We leveraged this approach to study sex-specific and -common connectivity patterns in schizophrenia.Results show dynamic reconfiguration and interaction within and between multi-spatial scales. Sex-specific differences occur (1) within the subcortical domain, (2) between the somatomotor and cerebellum domains, and (3) between the temporal domain and several others, including the subcortical, visual, and default mode domains. Most of the sex-specific differences belong to between-spatial scale functional interactions and are associated with a dynamic state with strong functional interactions between the visual, somatomotor, and temporal domains and their anticorrelation patterns with the rest of the brain. We observed significant correlations between multi-spatial scale functional interactions and symptom scores, highlighting the importance of multiscale analyses to identify potential biomarkers for schizophrenia. As such, we recommend such analyses as an important option for future functional connectivity studies.

Background: Electroconvulsive therapy (ECT) is one of the most effective treatments for major depressive disorder. Recently, there has been increasing attention to evaluate the effect of ECT on resting-state functional magnetic resonance imaging (rs-fMRI). This study aims to compare rs-fMRI of depressive disorder (DEP) patients with healthy participants, investigate whether pre-ECT dynamic functional network connectivity network (dFNC) estimated from patients rs-fMRI is associated with an eventual ECT outcome, and explore the effect of ECT on brain network states.Method: Resting-state functional magnetic resonance imaging (fMRI) data were collected from 119 patients with depression or depressive disorder (DEP) (76 females), and 61 healthy (HC) participants (34 females), with an age mean of 52.25 (N = 180) years old. The pre-ECT and post-ECT Hamilton Depression Rating Scale (HDRS) were 25.59 ± 6.14 and 11.48 ± 9.07, respectively. Twenty-four independent components from default mode (DMN) and cognitive control network (CCN) were extracted, using group-independent component analysis from pre-ECT and post-ECT rs-fMRI. Then, the sliding window approach was used to estimate the pre-and post-ECT dFNC of each subject. Next, k-means clustering was separately applied to pre-ECT dFNC and post-ECT dFNC to assess three distinct states from each participant. We calculated the amount of time each subject spends in each state, which is called “occupancy rate” or OCR. Next, we compared OCR values between HC and DEP participants. We also calculated the partial correlation between pre-ECT OCRs and HDRS change while controlling for age, gender, and site. Finally, we evaluated the effectiveness of ECT by comparing pre- and post-ECT OCR of DEP and HC participants.Results: The main findings include (1) depressive disorder (DEP) patients had significantly lower OCR values than the HC group in state 2, where connectivity between cognitive control network (CCN) and default mode network (DMN) was relatively higher than other states (corrected p = 0.015), (2) Pre-ECT OCR of state, with more negative connectivity between CCN and DMN components, is linked with the HDRS changes (R = 0.23 corrected p = 0.03). This means that those DEP patients who spent less time in this state showed more HDRS change, and (3) The post-ECT OCR analysis suggested that ECT increased the amount of time DEP patients spent in state 2 (corrected p = 0.03).Conclusion: Our finding suggests that dynamic functional network connectivity (dFNC) features, estimated from CCN and DMN, show promise as a predictive biomarker of the ECT outcome of DEP patients. Also, this study identifies a possible underlying mechanism associated with the ECT effect on DEP patients.

Bipolar disorder and schizophrenia are associated with brain morphometry alterations. This study investigates inter-individual variability in brain structural profiles, both within diagnostic groups and between patients and healthy individuals. Brain morphometric measures from three independent samples of patients with schizophrenia (n = 168), bipolar disorder (n = 122), and healthy individuals (n = 180) were modeled as single vectors to generated individualized profiles of subcortical volumes and regional cortical thickness. These profiles were then used to compute a person-based similarity index (PBSI) for subcortical volumes and for regional cortical thickness, to quantify the within-group similarity of the morphometric profile of each individual to that of the other participants in the same diagnostic group. There was no effect of diagnosis on the PBSI for subcortical volumes. In contrast, compared to healthy individuals, the PBSI for cortical thickness was lower in patients with schizophrenia (effect size = 0.4, p ≤ 0.0002), but not in patients with bipolar disorder. The results were robust and reproducible across samples. We conclude that disease mechanisms for these disorders produce modest inter-individual variations in brain morphometry that should be considered in future studies attempting to cluster patients in subgroups.

We introduce an extension of independent component analysis (ICA), called multiscale ICA, and design an approach to capture dynamic functional source interactions within and between multiple spatial scales. Multiscale ICA estimates functional sources at multiple spatial scales without imposing direct constraints on the size of functional sources, overcomes the limitation of using fixed anatomical locations, and eliminates the need for model-order selection in ICA analysis. We leveraged this approach to study sex-specific and sex-common connectivity patterns in schizophrenia. Results show dynamic reconfiguration and interaction within and between multi-spatial scales. Sex-specific differences occur (a) within the subcortical domain, (b) between the somatomotor and cerebellum domains, and (c) between the temporal domain and several others, including the subcortical, visual, and default mode domains. Most of the sex-specific differences belong to between-spatial-scale functional interactions and are associated with a dynamic state with strong functional interactions between the visual, somatomotor, and temporal domains and their anticorrelation patterns with the rest of the brain. We observed significant correlations between multi-spatial-scale functional interactions and symptom scores, highlighting the importance of multiscale analyses to identify potential biomarkers for schizophrenia. As such, we recommend such analyses as an important option for future functional connectivity studies.

Objective: A graph theoretical approach provides a powerful framework for discovering potential biomarkers of psychotic disorders. Comparing the brain graphs of the control and patient groups can help us to discover changes in mental disorders in a more convenient way. In this paper, we propose a novel tool to identify missing links associated with blocked paths (segregation) and new links associated with additional paths (abnormal integration) in estimated patient group’s time-varying graphs. We highlight the approach in an example application to the resting-state functional magnetic resonance imaging (fMRI) data of schizophrenia (SZ) patients. Methods: We first estimated whole-brain functional connectivity dynamics using a combination of spatial independent component analysis (ICA), sliding time window, and k-means clustering of windowed correlation matrices on resting-state fMRI data. The clusters are regarded as functional connectivity states, and each of them includes time intervals exhibiting similar connectivity patterns. We then estimated a Gaussian graphical model (GGM) for each state for both groups. To evaluate this approach, we compared different paths between brain components (nodes) of SZ and control groups’ graphs within each state by using the concept of connected components in graph theory. Results: We identified missing edges associated with disconnectivity (disconnectors) and showed there are additional edges in the SZ group graph that contribute to creating new paths in brain graphs. Conclusion: The proposed approach provides a tool for extracting time-varying graphs and identifying disconnectors associated with disconnectivity (absence of paths) and also connectors associated with abnormal integration (additional paths) in patient group graphs. Significance: We detected several missing links in SZ, both within and between functional domains, in particular within the subcortical (3 links) and somatomotor (4 links) domains. Interestingly, our proposed method identified new links within the somatomotor domain which may be related to a compensatory response in patients that warrants future study.

Understanding the associations of the structural and functional patterns of the brain is vital. Recent studies have focused on utilizing this information within and across the different functional and anatomical domains (i.e., groups of brain networks) using neuroimaging data. In this work, we use a Bayesian optimization-based method known as the Tree Parzen Estimator (TPE) to identify variation in the nature of information encoded by different functional magnetic resonance imaging (fMRI) sub-domains of the brain. We show by repeated cross-validation on a schizophrenia classification task that specific sub-domains may require more sophisticated learning architectures to contribute optimally to classification, while others require less complicated ones. Our findings reveal the need for adaptive, hierarchical learning frameworks catering to features from different sub-domains to optimally identify features enabling the prediction of the outcome of interest.

Background Dysfunctional reward processing is implicated in multiple mental disorders. Novelty seeking (NS) assesses preference for seeking novel experiences, which is linked to sensitivity to reward environmental cues. Methods A subset of 14-year-old adolescents (IMAGEN) with the top 20% ranked high-NS scores was used to identify high-NS–associated multimodal components by supervised fusion. These features were then used to longitudinally predict five different risk scales for the same and unseen subjects (an independent dataset of subjects at 19 years of age that was not used in predictive modeling training at 14 years of age) (within IMAGEN, n ≈ 1100) and even for the corresponding symptom scores of five types of patient cohorts (non-IMAGEN), including drinking (n = 313), smoking (n = 104), attention-deficit/hyperactivity disorder (n = 320), major depressive disorder (n = 81), and schizophrenia (n = 147), as well as to classify different patient groups with diagnostic labels. Results Multimodal biomarkers, including the prefrontal cortex, striatum, amygdala, and hippocampus, associated with high NS in 14-year-old adolescents were identified. The prediction models built on these features are able to longitudinally predict five different risk scales, including alcohol drinking, smoking, hyperactivity, depression, and psychosis for the same and unseen 19-year-old adolescents and even predict the corresponding symptom scores of five types of patient cohorts. Furthermore, the identified reward-related multimodal features can classify among attention-deficit/hyperactivity disorder, major depressive disorder, and schizophrenia with an accuracy of 87.2%. Conclusions Adolescents with higher NS scores can be used to reveal brain alterations in the reward-related system, implicating potential higher risk for subsequent development of multiple disorders. The identified high-NS–associated multimodal reward-related signatures may serve as a transdiagnostic neuroimaging biomarker to predict disease risks or severity.

Thalamocortical dysrhythmia is a key pathology of chronic neuropathic pain, but few studies have investigated thalamocortical networks in chronic low back pain (cLBP) given its non-specific etiology and complexity. Using fMRI, we propose an analytical pipeline to identify abnormal thalamocortical network dynamics in cLBP patients and validate the findings in two independent cohorts. We first identify two reoccurring dynamic connectivity states and their associations with chronic and temporary pain. Further analyses show that cLBP patients have abnormal connectivity between the ventral lateral/posterolateral nucleus (VL/VPL) and postcentral gyrus (PoCG) and between the dorsal/ventral medial nucleus and insula in the less frequent connectivity state, and temporary pain exacerbation alters connectivity between the VL/VPL and PoCG and the default mode network in the more frequent connectivity state. These results extend current findings on thalamocortical dysfunction and dysrhythmia in chronic pain and demonstrate that cLBP pathophysiology and clinical pain intensity are associated with distinct thalamocortical network dynamics., Thalamocortical dysrhythmia is a key pathology of chronic pain. Here, the authors propose an analytical pipeline to study dynamic fMRI brain networks and demonstrate that chronic low back pain pathophysiology and clinical pain intensity are associated with distinct thalamocortical network dynamics.

Resting-state fMRI (rs-fMRI) data are typically filtered at different frequency bins between 0.008∼0.2 Hz (varies across the literature) prior to analysis to mitigate nuisance variables (e.g., drift, motion, cardiac, and respiratory) and maximize the sensitivity to neuronal-mediated BOLD signal. However, multiple lines of evidence suggest meaningful BOLD signal may also be parsed at higher frequencies. To test this notion, a functional network connectivity (FNC) analysis based on a spatially informed independent component analysis (ICA) was performed at seven different bandpass frequency bins to examine FNC matrices across spectra. Further, eyes open (EO) vs. eyes closed (EC) resting-state acquisitions from the same participants were compared across frequency bins to examine if EO vs. EC FNC matrices and randomness estimations of FNC matrices are distinguishable at different frequencies.Results show that FNCs in higher-frequency bins display modular FNC similar to the lowest frequency bin, while r-to-z FNC and FNC-based measures indicating matrix non-randomness were highest in the 0.31-0.46 Hz range relative to all frequency bins above and below this range. As such, the FNC within this range appears to be the most temporally correlated, but the mechanisms facilitating this coherence require further analyses. Compared to EO, EC displayed greater FNC (involved in visual, cognitive control, somatomotor, and auditory domains) and randomness values at lower frequency bins, but this phenomenon flipped (EO > EC) at frequency bins greater than 0.46 Hz, particularly within visual regions.While the effect sizes range from small to large specific to frequency range and resting state (EO vs. EC), with little influence from common artifacts. These differences indicate that unique information can be derived from FNC between BOLD signals at different frequencies relative to a given restingstate acquisition and support the hypothesis meaningful BOLD signal is present at higher frequency ranges.

Previous successes of deep learning (DL) approaches on several complex tasks have hugely inflated expectations of their power to learn subtle properties of complex brain imaging data, and scale to large datasets. Perhaps as a reaction to this inflation, recent critical commentaries unfavorably compare DL with standard machine learning (SML) approaches for the analysis of brain imaging data. Yet, their conclusions are based on pre-engineered features which deprives DL of its main advantage: representation learning. Here we evaluate this and show the importance of representation learning for DL performance on brain imaging data. We report our findings from a large-scale systematic comparison of SML approaches versus DL profiled in a ten-way age and gender-based classification task on 12,314 structural MRI images. Results show that DL methods, if implemented and trained following the prevalent DL practices, have the potential to substantially improve compared to SML approaches. We also show that DL approaches scale particularly well presenting a lower asymptotic complexity in relative computational time, despite being more complex. Our analysis reveals that the performance improvement saturates as the training sample size grows, but shows significantly higher performance throughout. We also show evidence that the superior performance of DL is primarily due to the excellent representation learning capabilities and that SML methods can perform equally well when operating on representations produced by the trained DL models. Finally, we demonstrate that DL embeddings span a comprehensible projection spectrum and that DL consistently localizes discriminative brain biomarkers, providing an example of the robustness of prediction relevance estimates. Our findings highlight the presence of non-linearities in brain imaging data that DL frameworks can exploit to generate superior predictive representations for characterizing the human brain, even with currently available data sizes.

Our recent findings show that functional organizations evolve spatially over time, highlighting the importance of considering within-subject spatial variations and dynamic functional parcellations in brain functional analyses. Meanwhile, a considerable level of multi-functionality suggests the need for overlapping brain parcellations. In this work, we used ultra-high-order ICA to identify fine overlapping functional dynamic parcellations of the brain. The preliminary result of this work was presented at the organization for human brain mapping workshop (OHBM 2019)1.

Schizophrenia (SZ) is a severe neuropsychiatric disorder with a hallmark of functional dysconnectivity between numerous brain regions. With an implicit assumption of stationary brain interactions during the scanning period, most of the resting-state functional magnetic resonance imaging (fMRI) studies are conducted on static functional network connectivity (sFNC). Dynamic functional network connectivity (dFNC) that explores temporal patterns of functional connectivity (FC) might provide additional information to its static counterpart. In this work, we first estimate latent features (called connectivity states) by applying k-means clustering on dFNC. Next, using the estimated latent features, we trained and tested a classifier, which can differentiate SZ from healthy control (HC) subjects with 71% accuracy. Using a feature selection method embedded in the classifier, we have highlighted the role of transition probabilities between states as potential biomarkers and identified the role of lightly modularized transient connectivity state in pulling healthy subjects out of both highly modularized and very disconnected states. This will offer some new understandings about the way the healthy brain shifts between the most and the least connected states of whole brain connectivity.

Background: Major depressive disorder characterized by abnormalities of cognitive functions and mood is one of the leading causes of disability. Electroconvulsive therapy (ECT), which involves a brief electrical stimulation of the anesthesia brain, is the most potent treatment used in patients with depressive episodes (DEP) due to its rapid efficacy. Methods: In this work, we investigated how the human brain dynamically responds to ECT and whether the dynamic responses are associated with treatment outcomes and side-effects in 110 patients with DEP and 60 healthy controls (HCs) via a combination of a fully automated independent component analysis (ICA)-based pipeline and the clustering analysis. Findings: Five reoccurring connectivity states were identified, among which DEP had fewer occurrences in one antagonism state (state 1) with strong positive and negative connectivity. DEP patients increased the occupancy of two states (states 3 and 4) after ECT, resulting in significantly different occurrences of one additional state (state 3) compared to HC. We further found that DEP patients had diminished global meta-state dynamism, two of which recovered to normal after ECT. Interestingly, the changes in dynamic connectivity characteristics were associated with the changes in memory recall and Hamilton Depression Rating Scale of DEP after ECT. Interpretation: These converging results extend current findings on subcortical-cortical dysfunction and dysrhythmia in DEP and demonstrate that ECT might cause a remodeling of brain functional dynamics that underlies antidepressant symptom response and memory function of patients. Funding Statement: This work was supported National Institutes of Health (No. R01MH118695, R01EB020407 and R01MH117107), and China Natural Science Foundation (No. 61773380), Beijing Municipal Science and Technology Commission (No. Z181100001518005). Declaration of Interests: The authors declare no competing interests. Ethics Approval Statement: Written informed consent was obtained from all participants under protocols approved by the Institutional Review Board (IRB) of UNM and UCLA.

Behavioral changes are the earliest signs of a mental disorder, but arguably, the dynamics of brain function gets affected even earlier. Subsequently, spatio-temporal structure of disorder-specific dynamics is crucial for early diagnosis and understanding the disorder mechanism. A common way of learning discriminatory features relies on training a classifier and evaluating feature importance. Classical classifiers, based on handcrafted features are quite powerful, but suffer the curse of dimensionality when applied to large input dimensions of spatio-temporal data. Deep learning algorithms could handle the problem and a model introspection could highlight discriminatory spatio-temporal regions but need way more samples to train. In this paper we present a novel self supervised training schema which reinforces whole sequence mutual information local to context (whole MILC). We pre-train the whole MILC model on unlabeled and unrelated healthy control data. We test our model on three different disorders (i) Schizophrenia (ii) Autism and (iii) Alzheimers and four different studies. Our algorithm outperforms existing self-supervised pre-training methods and provides competitive classification results to classical machine learning algorithms. Importantly, whole MILC enables attribution of subject diagnosis to specific spatio-temporal regions in the fMRI signal.

Dimension reduction is essential in fMRI decoding, but the complex relationship between fMRI data and class labels is often unknown or not well modeled so that the most effective dimension reduction (e.d.r.) directions can hardly be identified. In the present study, we introduce a novel fMRI decoding approach based on an effective and general dimension reduction method, namely sliced inverse regression (SIR), which can exploit class information for estimating e.d.r. directions even when the relationship between fMRI data and class labels is not explicitly known. We incorporate singular value decomposition (SVD) into SIR to overcome SIR's limitation in dealing with ultra-high-dimensional data, and integrate SVD-SIR into a pattern classifier to enable quantification of the contributions of fMRI voxels to class labels. The resultant new SIR decoding analysis (SIR-DA) approach is capable of decoding behavioral responses and identifying predictive fMRI patterns. Simulation results showed that SIR-DA can more accurately detect e.d.r. directions and achieve higher classification accuracy than decoding approaches based on conventional dimension reduction methods. Further, we applied SIR-DA on real-world pain-evoked fMRI data to decode the level of pain perception and showed that SIR-DA can achieve higher accuracy in pain prediction than conventional methods. These results suggest that SIR-DA is an effective data-driven technique to decode behavioral or cognitive states from fMRI data and to uncover unknown brain patterns associated with behavior or cognitive responses.

Accumulating evidence shows that Coronavirus Disease 19 (COVID-19) survivors may encounter prolonged mental issues, especially post-traumatic stress symptoms (PTSS). Despite manifesting a plethora of behavioral or mental issues in COVID-19 survivors, previous studies illustrated that static brain functional networks of these survivors remain intact. The insignificant results could be due to the conventional statistic network analysis was unable to reveal information that can vary considerably in different temporal scales. In contrast, time-varying characteristics of the dynamic functional networks may help reveal important brain abnormalities in COVID-19 survivors. To test this hypothesis, we assessed PTSS and collected functional magnetic resonance imaging (fMRI) with COVID-19 survivors discharged from hospitals and matched controls. Results showed that COVID-19 survivors self-reported a significantly higher PTSS than controls. Tapping into the moment-to-moment variations of the fMRI data, we captured the dynamic functional network connectivity (dFNC) states, and three discriminative reoccurring brain dFNC states were identified. First of all, COVID-19 survivors showed an increased occurrence of a dFNC state with heterogeneous patterns between sensorimotor and visual networks. More importantly, the occurrence rate of this state was significantly correlated with the severity of PTSS. Finally, COVID-19 survivors demonstrated decreased topological organizations in this dFNC state than controls, including the node strength, degree, and local efficiency of the supplementary motor area. To conclude, our findings revealed the altered temporal characteristics of functional networks and their associations with PTSS due to COVID- 19. The current results highlight the importance of evaluating dynamic functional network changes with COVID-19 survivors.

Background Resting-state fMRI (rs-fMRI) is employed to assess “functional connections” of signal between brain regions. However, multiple rs-fMRI paradigms and data-filtering strategies have been used, highlighting the need to explore BOLD signal across the spectrum. Rs-fMRI data is typically filtered at frequencies ranging between 0.008∼0.2 Hz to mitigate nuisance signal (e.g. cardiac, respiratory) and maximize neuronal BOLD signal. However, some argue neuronal BOLD signal may be parsed at higher frequencies. New method To assess the contributions of rs-fMRI along the BOLD spectra on functional network connectivity (FNC) matrices, a spatially constrained independent component analysis (ICA) was performed at seven different frequency “bins”, after which FNC values and FNC measures of matrix-randomness were assessed using linear mixed models. Results Results show FNCs at higher-frequency bins display similar randomness to those from the typical frequency bins (0.01−0.15), while the largest values are in the 0.31−0.46 Hz bin. Eyes open (EO) vs eyes closed (EC) comparison found EC was less random than EO across most frequency bins. Further, FNC was greater in EC across auditory and cognitive control pairings while EO values were greater in somatomotor, visual, and default mode FNC. Comparison with existing methods Effect sizes of frequency and resting-state paradigm vary from small to large, but the most notable results are specific to frequency ranges and resting-state paradigm with artifacts like motion displaying negligible effect sizes. Conclusions These suggest unique information may be derived from FNC matrices across frequencies and paradigms, but additional data is necessary prior to any definitive conclusions.

Spatial independent component analysis (sICA) has become an integral part of functional MRI (fMRI) studies, particularly with resting-state fMRI. Early work used low-order ICA with between 20 and 45 components, which has led to the identification of around a dozen reproducible, distributed, large-scale brain networks. While regions within each largescale network are fairly temporally coherent, later studies have shown that each distributed network can be split into a group of spatially granular, and temporally covarying functional parcels. Thus, higher model order ICAs (75~150 components) have been employed to identify functional units known as intrinsic connectivity networks (ICNs). Our recent work suggests that an ICA framework can identify even more granular and functionally homogeneous brain functional units, and has the potential to provide more precise estimates of ICNs. In this study, we adopted an ICA with 1000 components (1k-ICA) to parcellate the brain into fine-grain sparse but overlapping ICNs and evaluated their properties and reliability in various ways. Our findings show that ultra-high-order ICA approaches like 1k-ICA can provide reliable, spatially-sparse ICNs.

Background Bipolar disorder (BD) patients show morphological abnormalities in gray matter (GM) and white matter (WM), which can be revealed by structure MRI (sMRI) and diffusion tensor imaging (DTI) respectively. However, previous studies on BD mainly relied on separated analysis of single neuroimaging modality, and it remains unclear how GM and WM covary to the abnormal brain structures of BD patients. Methods We recorded multimodal sMRI-DTI data of 35 BD patients and 30 healthy controls (HC) and used multimodal canonical component analysis and joint independent component analysis (mCCA-jICA) to identify altered covariant structures in GM and WM of BD. Group-discriminative and joint group-discriminative independent components (ICs) were identified between BD and HC. Correlation analysis was performed between the mixing coefficients and behavioral index. Results For BD patients, experiments results revealed that the GM atrophy in inferior frontal gyrus, right anterior cingulate gyrus and left superior frontal gyrus are associated with the WM integrity reduction in corticospinal tract and superior longitudinal fasciculus. Further, compared with HC, different correlation between mixing coefficients of ICs and age was observed for BD patients. Limitations The number of participants needs to be increased to more rigorously validate the results of this study, ideally from multiple sites. Functional imaging data could be utilized to explore structural-functional covariant pattern in BD. Possible confounding effect of medication. Conclusions We performed fusion analysis of sMRI and DTI and revealed covariant (GM-WM) structural patterns of BD patients. This study could be useful for developing more reliable neural biomarkers of BD.

ObjectiveTo identify and validate an fMRI-based neural marker for migraine without aura (MwoA) and to examine its association with treatment response.MethodsWe conducted cross-sectional studies with resting-state fMRI data from 230 participants and machine learning analyses. In studies 1 through 3, we identified, cross-validated, independently validated, and cross-sectionally validated an fMRI-based neural marker for MwoA. In study 4, we assessed the relationship between the neural marker and treatment responses in migraineurs who received a 4-week real or sham acupuncture treatment, or were waitlisted, in a registered clinical trial.ResultsIn study 1 (n = 116), we identified a neural marker with abnormal functional connectivity within the visual, default mode, sensorimotor, and frontal-parietal networks that could discriminate migraineurs from healthy controls (HCs) with 93% sensitivity and 89% specificity. In study 2 (n = 38), we investigated the generalizability of the marker by applying it to an independent cohort of migraineurs and HCs and achieved 84% sensitivity and specificity. In study 3 (n = 76), we verified the specificity of the marker with new datasets of migraineurs and patients with other chronic pain disorders (chronic low back pain and fibromyalgia) and demonstrated 78% sensitivity and 76% specificity for discriminating migraineurs from nonmigraineurs. In study 4 (n = 116), we found that the changes in the marker responses showed significant correlation with the changes in headache frequency in response to real acupuncture.ConclusionWe identified an fMRI-based neural marker that captures distinct characteristics of MwoA and can link disease pattern changes to brain changes.

There is growing evidence that rather than using a single brain imaging modality to study its association with physiological or symptomatic features, the field is paying more attention to fusion of multimodal information. However, most current multimodal fusion approaches that incorporate functional magnetic resonance imaging (fMRI) are restricted to second-level 3D features, rather than the original 4D fMRI data. This trade-off is that the valuable temporal information is not utilized during the fusion step. Here we are motivated to propose a novel approach called “parallel group ICA+ICA” that incorporates temporal fMRI information from GICA into a parallel ICA framework, aiming to enable direct fusion of first-level fMRI features with other modalities (e.g.structural MRI), which thus can detect linked functional network variability and structural covariations. Simulation results show that the proposed method yields accurate inter-modality linkage detection regardless of whether it is strong or weak. When applied to real data, we identified one pair of significantly associated fMRI-sMRI components that show group difference between schizophrenia and controls in both modalities. Finally, multiple cognitive domain scores can be predicted by the features identified in the linked component pair by our proposed method. We also show these multimodal brain features can predict multiple cognitive scores in an independent cohort. Overall, results demonstrate the ability of parallel GICA+ICA to estimate joint information from 4D and 3D data without discarding much of the available information up front, and the potential for using this approach to identify imaging biomarkers to study brain disorders.

Conventionally, brain function is inferred from the magnitude data of the complex-valued fMRI output. Since the fMRI phase image (unwrapped) provides a representation of brain internal magnetic fieldmap (by a constant scale difference), it can also be used to study brain function while providing a more direct representation of the brain's magnetic state. In this study, we collected a cohort of resting-state fMRI magnitude and phase data pairs from 600 subjects (age from 10 to 76, 346 males), decomposed the phase data by group independent component analysis (pICA), calculated the functional network connectivity (pFNC). In comparison with the magnitude-based brain function analysis (mICA and mFNC), we find that the pFNC matrix contains fewer significant functional connections (with p-value thresholding) than the mFNC matrix, which are sparsely distributed across the whole brain with near/far interconnections and positive/negative correlations in rough balance. We also find a few of brain rest sub-networks within the phase data, primarily in subcortical, cerebellar, and visual regions. Overall, our findings offer new insights into brain function connectivity in the context of a focus on the brain's internal magnetic state.

Subcortical ischemic vascular disease (SIVD) is a major subtype of vascular dementia with features that overlap clinically with Alzheimer's disease (AD), confounding diagnosis. Neuroimaging is a more specific and biologically based approach for detecting brain changes and thus may help to distinguish these diseases. There is still a lack of knowledge regarding the shared and specific functional brain abnormalities, especially functional connectivity changes in relation to AD and SIVD. In this study, we investigated both static functional network connectivity (sFNC) and dynamic FNC (dFNC) between 54 intrinsic connectivity networks in 19 AD patients, 19 SIVD patients, and 38 age-matched healthy controls. The results show that both patient groups have increased sFNC between the visual and cerebellar (CB) domains but decreased sFNC between the cognitive-control and CB domains. SIVD has specifically decreased sFNC within the sensorimotor domain while AD has specifically altered sFNC between the default-mode and CB domains. In addition, SIVD has more occurrences and a longer dwell time in the weakly connected dFNC states, but with fewer occurrences and a shorter dwell time in the strongly connected dFNC states. AD has both similar and opposite changes in certain dynamic features. More importantly, the dynamic features are found to be associated with cognitive performance. Our findings highlight similar and distinct functional connectivity alterations in AD and SIVD from both static and dynamic perspectives and indicate dFNC to be a more important biomarker for dementia since its progressively altered patterns can better track cognitive impairment in AD and SIVD.

Brain functional networks identified from fMRI data can provide potential biomarkers for brain disorders. Group independent component analysis (GICA) is popular for extracting brain functional networks from multiple subjects. In GICA, different strategies exist for reconstructing subject-specific networks from the group-level networks. However, it is unknown whether these strategies have different sensitivities to group differences and abilities in distinguishing patients. Among GICA, spatio-temporal regression (STR) and spatially constrained ICA approaches such as group information guided ICA (GIG-ICA) can be used to propagate components (indicating networks) to a new subject that is not included in the original subjects. In this study, based on the same a priori network maps, we reconstructed subject-specific networks using these two methods separately from resting-state fMRI data of 151 schizophrenia patients (SZs) and 163 healthy controls (HCs). We investigated group differences in the estimated functional networks and the functional network connectivity (FNC) obtained by each method. The networks were also used as features in a cross-validated support vector machine (SVM) for classifying SZs and HCs. We selected features using different strategies to provide a comprehensive comparison between the two methods. GIG-ICA generally showed greater sensitivity in statistical analysis and better classification performance (accuracy 76.45 ± 8.9%, sensitivity 0.74 ± 0.11, specificity 0.79 ± 0.11) than STR (accuracy 67.45 ± 8.13%, sensitivity 0.65 ± 0.11, specificity 0.71 ± 0.11). Importantly, results were also consistent when applied to an independent dataset including 82 HCs and 82 SZs. Our work suggests that the functional networks estimated by GIG-ICA are more sensitive to group differences, and GIG-ICA is promising for identifying image-derived biomarkers of brain disease., Highlights • We investigate which group ICA method is more sensitive to group differences in networks and FNC. • We evaluate which group ICA method can result in better classification performance. • We compare the spatio-temporal regression (STR) and GIG-ICA using different feature selection and model building strategies and independent samples. • We use different feature selection and model building strategies and classification using independent samples. • Using fMRI data including schizophrenia patients, GIG-ICA shows better performance than the traditional STR method.

Thalamocortical dysrhythmia is a key pathology of chronic neuropathic pain, but few studies have investigated thalamocortical networks in chronic low back pain (cLBP) given its non-specific etiology and complexity. Using fMRI, we propose an analytical pipeline to identify abnormal thalamocortical network dynamics in cLBP patients and validate the findings in two independent cohorts. We first identify two reoccurring dynamic connectivity states and their associations with chronic and temporary pain. Further analyses show that cLBP patients have abnormal connectivity between the ventral lateral/posterolateral nucleus (VL/VPL) and postcentral gyrus (PoCG) and between the dorsal/ventral medial nucleus and insula in the less frequent connectivity state, and temporary pain exacerbation alters connectivity between the VL/VPL and PoCG and the default mode network in the more frequent connectivity state. These results extend current findings on thalamocortical dysfunction and dysrhythmia in chronic pain and demonstrate that cLBP pathophysiology and clinical pain intensity are associated with distinct thalamocortical network dynamics.

Graphical abstract, Highlights • Propose a new pipeline to link brain changes among different datasets, studies, and disorders. • Identify reproducible biomarkers in schizophrenia using independent data. • Find both common and unique brain impairments in schizophrenia and autism. • Reveal gradual changes from healthy controls to mild cognitive impairment to Alzheimer’s disease. • Obtain high classification accuracy (~90%) between bipolar disorder and major depressive disorder., Many mental illnesses share overlapping or similar clinical symptoms, confounding the diagnosis. It is important to systematically characterize the degree to which unique and similar changing patterns are reflective of brain disorders. Increasing sharing initiatives on neuroimaging data have provided unprecedented opportunities to study brain disorders. However, it is still an open question on replicating and translating findings across studies. Standardized approaches for capturing reproducible and comparable imaging markers are greatly needed. Here, we propose a pipeline based on the priori-driven independent component analysis, NeuroMark, which is capable of estimating brain functional network measures from functional magnetic resonance imaging (fMRI) data that can be used to link brain network abnormalities among different datasets, studies, and disorders. NeuroMark automatically estimates features adaptable to each individual subject and comparable across datasets/studies/disorders by taking advantage of the reliable brain network templates extracted from 1828 healthy controls as guidance. Four studies including 2442 subjects were conducted spanning six brain disorders (schizophrenia, autism spectrum disorder, mild cognitive impairment, Alzheimer’s disease, bipolar disorder, and major depressive disorder) to evaluate validity of the proposed pipeline from different perspectives (replication of brain abnormalities, cross-study comparison, identification of subtle brain changes, and multi-disorder classification using identified biomarkers). Our results highlight that NeuroMark effectively identified replicated brain network abnormalities of schizophrenia across different datasets; revealed interesting neural clues on the overlap and specificity between autism and schizophrenia; demonstrated brain functional impairments present to varying degrees in mild cognitive impairments and Alzheimer's disease; and captured biomarkers that achieved good performance in classifying bipolar disorder and major depressive disorder.

Brain functional imaging data, especially functional magnetic resonance imaging (fMRI) data, have been employed to reflect functional integration of the brain. Alteration in brain functional connectivity (FC) is expected to provide potential biomarkers for classifying or predicting brain disorders. In this paper, we present a comprehensive review in order to provide guidance about the available brain FC measures and typical classification strategies. We survey the state-of-the-art FC analysis methods including widely used static functional connectivity (SFC) and more recently proposed dynamic functional connectivity (DFC). Temporal correlations among regions of interest (ROIs), data-driven spatial network and functional network connectivity (FNC) are often computed to reflect SFC from different angles. SFC can be extended to DFC using a sliding-window framework, and intrinsic connectivity states along the time-varying connectivity patterns are typically extracted using clustering or decomposition approaches. We also briefly summarize window-less DFC approaches. Subsequently, we highlight various strategies for feature selection including the filter, wrapper and embedded methods. In terms of model building, we include traditional classifiers as well as more recently applied deep learning methods. Moreover, we review representative applications with remarkable classification accuracy for psychosis and mood disorders, neurodevelopmental disorder, and neurological disorders using fMRI data. Schizophrenia, bipolar disorder, autism spectrum disorder (ASD), attention deficit hyperactivity disorder (ADHD), Alzheimer's disease and mild cognitive impairment (MCI) are discussed. Finally, challenges in the field are pointed out with respect to the inaccurate diagnosis labeling, the abundant number of possible features and the difficulty in validation. Some suggestions for future work are also provided.

Discriminating Alzheimer's disease (AD) from its prodromal form, mild cognitive impairment (MCI), is a significant clinical problem that may facilitate early diagnosis and intervention, in which a more challenging issue is to classify MCI subtypes, i.e., those who eventually convert to AD (cMCI) versus those who do not (MCI). To solve this difficult 4-way classification problem (AD, MCI, cMCI and healthy controls), a competition was hosted by Kaggle to invite the scientific community to apply their machine learning approaches on pre-processed sets of T1-weighted magnetic resonance images (MRI) data and the demographic information from the international Alzheimer's disease neuroimaging initiative (ADNI) database. This paper summarizes our competition results. We first proposed a hierarchical process by turning the 4-way classification into five binary classification problems. A new feature selection technology based on relative importance was also proposed, aiming to identify a more informative and concise subset from 426 sMRI morphometric and 3 demographic features, to ensure each binary classifier to achieve its highest accuracy. As a result, about 2% of the original features were selected to build a new feature space, which can achieve the final four-way classification with a 54.38% accuracy on testing data through hierarchical grouping, higher than several alternative methods in comparison. More importantly, the selected discriminative features such as hippocampal volume, parahippocampal surface area, and medial orbitofrontal thickness, etc. as well as the MMSE score, are reasonable and consistent with those reported in AD/MCI deficits. In summary, the proposed method provides a new framework for multi-way classification using hierarchical grouping and precise feature selection.

Studies of time-varying or dynamic brain connectivity (BC) using functional magnetic resonance imaging (fMRI) are crucial to understand the relationship between different brain regions. This paper presents a novel method for estimating dynamic BC using a time-varying multivariate autoregressive (AR) model with spatial sparsity and temporal continuity constraints. The problem is formulated as a maximum a posterior probability (MAP) estimation problem and solved as a least square problem with Li-regularization for imposing the constraints. The Limited-memory Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) method is employed to estimate the model parameters for making inference of dynamic BC. The proposed method was evaluated using synthetic data and visual checkerboard task experiment fMRI data. The results show that the method can effectively capture transient information transfer among visual-related brain regions whereas controlled areas not related to the process remain inactive. These verify the effectiveness and reduced variance of the proposed method for investigating dynamic task-related BC from fMRI data.

Dimension reduction is essential for identifying a small set of discriminative features that are predictive of behavior or cognition from high-dimensional functional magnetic resonance imaging (fMRI) data. However, conventional linear dimension reduction techniques cannot reduce the dimension effectively if the relationship between imaging data and behavioral parameters are nonlinear. In the paper, we proposed a novel supervised dimension reduction technique, named PC-SIR (Principal Component - Sliced Inverse Regression), for analyzing high-dimensional fMRI data. The PC-SIR method is an important extension of the renowned SIR method, which can achieve the effective dimension reduction (e.d.r.) directions even the relationship between class labels and predictors is nonlinear but is unable to handle high-dimensional data. By using PCA prior to SIR to orthogonalize and reduce the predictors, PC-SIR can overcome the limitation of SIR and thus can be used for fMRI data. Simulation showed that PC-SIR can result in a more accurate identification of brain activation as well as better prediction than support vector regression (SVR) and partial least square regression (PLSR). Then, we applied PC-SIR on real fMRI data recorded in a pain stimulation experiment to identify pain-related brain regions and predict the pain perception. Results on 32 subjects showed that PC-SIR can lead to significantly higher prediction accuracy than SVR and PLSR. Therefore, PC-SIR could be a promising dimension reduction technique for multivariate pattern analysis of fMRI.

Time-varying covariance is an important metric to measure the statistical dependence between non-stationary biological processes. Time-varying covariance is conventionally estimated from short-time data segments within a window having a certain bandwidth, but it is difficult to choose an appropriate bandwidth to estimate covariance with different degrees of non-stationarity. This paper introduces a local polynomial regression (LPR) method to estimate time-varying covariance and performs an asymptotic analysis of the LPR covariance estimator to show that both the estimation bias and variance are functions of the bandwidth and there exists an optimal bandwidth to minimize the mean square error (MSE) locally. A data-driven variable bandwidth selection method, namely the intersection of confidence intervals (ICI), is adopted in LPR for adaptively determining the local optimal bandwidth that minimizes the MSE. Experimental results on simulated signals show that the LPR-ICI method can achieve robust and reliable performance in estimating time-varying covariance with different degrees of variations and under different noise scenarios, making it a powerful tool to study the dynamic relationship between non-stationary biomedical signals. Further, we apply the LPR-ICI method to estimate time-varying covariance of functional magnetic resonance imaging (fMRI) signals in a visual task for the inference of dynamic functional brain connectivity. The results show that the LPR-ICI method can effectively capture the transient connectivity patterns from fMRI.

Exploration of the dynamics of functional brain connectivity based on the correlation coefficients of functional magnetic resonance imaging (fMRI) data is important for understanding the brain mechanisms. Because fMRI data are time-varying in nature, the functional connectivity shows substantial fluctuations and dynamic characteristics. However, an effective method for estimating time-varying functional connectivity is lacking, which is mainly due to the difficulty in choosing an appropriate window to localize the time-varying correlation coefficients (TVCC). This paper introduces a novel method for adaptively estimating the TVCC of non-stationary signals and studies its application to infer dynamic functional connectivity of fMRI data in a visual task. The proposed method employs a sliding window having a certain bandwidth to estimate the TVCC locally and the window bandwidths are selected adaptively by a local plug-in rule to minimize the mean squared error. The results show that the functional connectivity changes in the visual task are transient, which suggests that simply assuming sustained connectivity changes during task period might not be sufficient to capture dynamic connectivity changes induced by tasks.

Early prediction of diseased brain conditions is critical for curing illness and preventing irreversible neuronal dysfunction and loss. Generically regarding the different neuroimaging modalities as filtered, complementary insights of brain’s anatomical and functional organization, multimodal data fusion could be hypothesized to enhance the predictive power as compared to a unimodal prediction of disease progression. More recently, deep learning (DL) based methods on structural MRI (sMRI) data have outperformed classical machine learning approaches in several neuroimaging applications including diagnostic classification and prediction. Similarly, functional MRI (fMRI) features estimated using a dynamic (i.e. time-varying) functional connectivity (FC) approach have been found to be more discriminative and predictive of the clinical diagnosis than those based on the static FC approach. Motivated by this, we introduce a novel multimodal data fusion framework featuring deep residual learning of non-linear sMRI features and dynamic FC (dFC) based extraction of fMRI features to predict the subset of individuals with mild cognitive impairments who would progress to Alzheimer’s disease within a time-period of three years from the baseline scanning sessions. Our cross-validated results from the developed multimodal (sMRI-fMRI) data fusion framework demonstrate a significant improvement in performance over the unimodal prediction analyses with the fMRI (p = 7.03 x 10-7) and sMRI (p = 6.72 x 10-4) modalities. As such, the findings in this work highlight the benefits of combining multiple neuroimaging data modalities via data fusion, corroborate the predictive value of the tested DL and dFC features and argue in favor of exploration of similar approaches to learn neuroanatomical and functional alterations in the neuroimaging data.