A review of group ICA for fMRI data and ICA for joint inference of imaging, genetic, and ERP data

Repeated Transcranial Photobiomodulation with Light-Emitting Diodes Improves Psychomotor Vigilance and EEG Networks of the Human Brain

Transcranial photobiomodulation (tPBM) has been suggested as a non-invasive neuromodulation tool. The repetitive administration of light-emitting diode (LED)-based tPBM for several weeks significantly improves human cognition. To understand the electrophysiological effects of LED-tPBM on the human brain, we investigated alterations by repeated tPBM in vigilance performance and brain networks using electroencephalography (EEG) in healthy participants. Active and sham LED-based tPBM were administered to the right forehead of young participants twice a week for four weeks. The participants performed a psychomotor vigilance task (PVT) during each tPBM/sham experiment. A 64-electrode EEG system recorded electrophysiological signals from each participant during the first and last visits in a 4-week study. Topographical maps of the EEG power enhanced by tPBM were statistically compared for the repeated tPBM effect. A new data processing framework combining the group's singular value decomposition (gSVD) with eLORETA was implemented to identify EEG brain networks. The reaction time of the PVT in the tPBM-treated group was significantly improved over four weeks compared to that in the sham group. We observed acute increases in EEG delta and alpha powers during a 10 min LED-tPBM while the participants performed the PVT task. We also found that the theta, beta, and gamma EEG powers significantly increased overall after four weeks of LED-tPBM. Combining gSVD with eLORETA enabled us to identify EEG brain networks and the corresponding network power changes by repeated 4-week tPBM. This study clearly demonstrated that a 4-week prefrontal LED-tPBM can neuromodulate several key EEG networks, implying a possible causal effect between modulated brain networks and improved psychomotor vigilance outcomes.

Extracting functional connectivity brain networks at the resting state from pulsed arterial spin labeling data

Introduction

Functional connectivity in the brain is often studied with blood oxygenation level dependent (BOLD) resting state functional magnetic resonance imaging (rsfMRI), but the BOLD signal is several steps removed from neuronal activity. Arterial spin labeling (ASL), particularly pulsed ASL (PASL), has also the capacity to measure the blood-flow changes in response to activity. In this paper, we investigated the feasibility of extracting major brain networks from PASL data, in contrast with rsfMRI analsyis.

Materials and methods

In this retrospective study, we analyzed a cohort dataset that consists of 21 mild traumatic brain injury (mTBI) patients and 29 healthy controls, which was collected in a previous study. By extracting 10 major brain networks from the data of both PASL and rsfMRI, we contrasted their similarities and differences in the 10 networks extracted from both modalities.

Results

Our data demonstrated that PASL could be used to extract all 10 major brain networks. Eight out of 10 networks demonstrated over 60 % similarity to rsfMRI data. Meanwhile, there are similar but not identical changes in networks detected between mTBI patients and healthy controls with both modalities. Notably, the PASL-extracted default mode network (DMN), other than the rsfMRI-extracted DMN, includes some regions known to be associated with the DMN in other studies. It demonstrated that PASL data can be analyzed to identify resting state networks with reasonable reliability, even without rsfMRI data.

Conclusion

Our analysis provides an opportunity to extract functional connectivity information in heritage datasets in which ASL but not BOLD was collected.

Leading basic modes of spontaneous activity drive individual functional connectivity organization in the resting human brain

Spontaneous activity of the human brain provides a window to explore intrinsic principles of functional organization. However, most studies have focused on interregional functional connectivity. The principles underlying rich repertoires of instantaneous activity remain largely unknown. We apply a recently proposed eigen-microstate analysis to three resting-state functional MRI datasets to identify basic modes that represent fundamental activity patterns that coexist over time. We identify five leading basic modes that dominate activity fluctuations. Each mode exhibits a distinct functional system-dependent coactivation pattern and corresponds to specific cognitive profiles. In particular, the spatial pattern of the first leading basis mode shows the separation of activity between the default-mode and primary and attention regions. Based on theoretical modelling, we further reconstruct individual functional connectivity as the weighted superposition of coactivation patterns corresponding to these leading basic modes. Moreover, these leading basic modes capture sleep deprivation-induced changes in brain activity and interregional connectivity, primarily involving the default-mode and task-positive regions. Our findings reveal a dominant set of basic modes of spontaneous activity that reflect multiplexed interregional coordination and drive conventional functional connectivity, furthering the understanding of the functional significance of spontaneous brain activity.

A deep learning method for autism spectrum disorder identification based on interactions of hierarchical brain networks

BACKGROUND

It has been recently shown that deep learning models exhibited remarkable performance of representing functional Magnetic Resonance Imaging (fMRI) data for the understanding of brain functional activities. With hierarchical structure, deep learning models can infer hierarchical functional brain networks (FBN) from fMRI. However, the applications of the hierarchical FBNs have been rarely studied.

METHODS

In this work, we proposed a hierarchical recurrent variational auto-encoder (HRVAE) to unsupervisedly model the fMRI data. The trained HRVAE encoder can predict hierarchical temporal features from its three hidden layers, and thus can be regarded as a hierarchical feature extractor. Then LASSO (least absolute shrinkage and selection operator) regression was applied to estimate the corresponding hierarchical FBNs. Based on the hierarchical FBNs from each subject, we constructed a novel classification framework for brain disorder identification and test it on the Autism Brain Imaging Data Exchange (ABIDE) dataset, a world-wide multi-site database of autism spectrum disorder (ASD). We analyzed the hierarchy organization of FBNs, and finally used the overlaps of hierarchical FBNs as features to differentiate ASD from typically developing controls (TDC).

RESULTS

The experimental results on 871 subjects from ABIDE dataset showed that the HRVAE model can effectively derive hierarchical FBNs including many well-known resting state networks (RSN). Moreover, the classification result improved the state-of-the-art by achieving a very high accuracy of 82.1 %.

CONCLUSIONS

This work presents a novel data-driven deep learning method using fMRI data for ASD identification, which could provide valuable reference for clinical diagnosis. The classification results suggest that the interactions of hierarchical FBNs have association with brain disorder, which promotes the understanding of FBN hierarchy and could be applied to other brain disorder analysis.

Broadening the scope: Multiple functional connectivity networks underlying threat and safety signaling

Introduction

Threat learning and extinction processes are thought to be foundational to anxiety and fear-related disorders. However, the study of these processes in the human brain has largely focused on a priori regions of interest, owing partly to the ease of translating between these regions in human and non-human animals. Moving beyond analyzing focal regions of interest to whole-brain dynamics during threat learning is essential for understanding the neuropathology of fear-related disorders in humans.

Methods

223 participants completed a 2-day Pavlovian threat conditioning paradigm while undergoing fMRI. Participants completed threat acquisition and extinction. Extinction recall was assessed 48 hours later. Using a data-driven group independent component analysis (ICA), we examined large-scale functional connectivity networks during each phase of threat conditioning. Connectivity networks were tested to see how they responded to conditional stimuli during early and late phases of threat acquisition and extinction and during early trials of extinction recall.

Results

A network overlapping with the default mode network involving hippocampus, vmPFC, and posterior cingulate was implicated in threat acquisition and extinction. Another network overlapping with the salience network involving dACC, mPFC, and inferior frontal gyrus was implicated in threat acquisition and extinction recall. Other networks overlapping with parts of the salience, somatomotor, visual, and fronto-parietal networks were involved in the acquisition or extinction of learned threat responses.

Conclusions

These findings help confirm previous investigations of specific brain regions in a model-free fashion and introduce new findings of spatially independent networks during threat and safety learning. Rather than being a single process in a core network of regions, threat learning involves multiple brain networks operating in parallel coordinating different functions at different timescales. Understanding the nature and interplay of these dynamics will be critical for comprehensive understanding of the multiple processes that may be at play in the neuropathology of anxiety and fear-related disorders.

brainlife.io: A decentralized and open source cloud platform to support neuroscience research

Neuroscience research has expanded dramatically over the past 30 years by advancing standardization and tool development to support rigor and transparency. Consequently, the complexity of the data pipeline has also increased, hindering access to FAIR data analysis to portions of the worldwide research community. brainlife.io was developed to reduce these burdens and democratize modern neuroscience research across institutions and career levels. Using community software and hardware infrastructure, the platform provides open-source data standardization, management, visualization, and processing and simplifies the data pipeline. brainlife.io automatically tracks the provenance history of thousands of data objects, supporting simplicity, efficiency, and transparency in neuroscience research. Here brainlife.io's technology and data services are described and evaluated for validity, reliability, reproducibility, replicability, and scientific utility. Using data from 4 modalities and 3,200 participants, we demonstrate that brainlife.io's services produce outputs that adhere to best practices in modern neuroscience research.

Subject-specific whole-brain parcellations of nodes and boundaries are modulated differently under 10 Hz rTMS

Repetitive transcranial magnetic stimulation (rTMS) has gained considerable importance in the treatment of neuropsychiatric disorders, including major depression. However, it is not yet understood how rTMS alters brain's functional connectivity. Here we report changes in functional connectivity captured by resting state functional magnetic resonance imaging (rsfMRI) within the first hour after 10 Hz rTMS. We apply subject-specific parcellation schemes to detect changes (1) in network nodes, where the strongest functional connectivity of regions is observed, and (2) in network boundaries, where functional transitions between regions occur. We use support vector machine (SVM), a widely used machine learning algorithm that is robust and effective, for the classification and characterization of time intervals of changes in node and boundary maps. Our results reveal that changes in connectivity at the boundaries are slower and more complex than in those observed in the nodes, but of similar magnitude according to accuracy confidence intervals. These results were strongest in the posterior cingulate cortex and precuneus. As network boundaries are indeed under-investigated in comparison to nodes in connectomics research, our results highlight their contribution to functional adjustments to rTMS.

Identifiability in Functional Connectivity May Unintentionally Inflate Prediction Results

Functional magnetic resonance (fMRI) is an invaluable tool in studying cognitive processes in vivo. Many recent studies use functional connectivity (FC), partial correlation connectivity (PC), or fMRI-derived brain networks to predict phenotypes with results that sometimes cannot be replicated. At the same time, FC can be used to identify the same subject from different scans with great accuracy. In this paper, we show a method by which one can unknowingly inflate classification results from 61% accuracy to 86% accuracy by treating longitudinal or contemporaneous scans of the same subject as independent data points. Using the UK Biobank dataset, we find one can achieve the same level of variance explained with 50 training subjects by exploiting identifiability as with 10,000 training subjects without double-dipping. We replicate this effect in four different datasets: the UK Biobank (UKB), the Philadelphia Neurodevelopmental Cohort (PNC), the Bipolar and Schizophrenia Network for Intermediate Phenotypes (BSNIP), and an OpenNeuro Fibromyalgia dataset (Fibro). The unintentional improvement ranges between 7% and 25% in the four datasets. Additionally, we find that by using dynamic functional connectivity (dFC), one can apply this method even when one is limited to a single scan per subject. One major problem is that features such as ROIs or connectivities that are reported alongside inflated results may confuse future work. This article hopes to shed light on how even minor pipeline anomalies may lead to unexpectedly superb results.

Changes of creative ability and underlying brain network connectivity throughout the lifespan

Creativity, or divergent thinking, is essential to and supported by cognitive functions necessary for everyday tasks. The current study investigates divergent thinking and its neural mechanisms from adolescence to late adulthood. To do this, 180 healthy participants completed a creativity task called the egg task including 86 adolescents (mean age (SD) = 13.62 (1.98)), 52 young adults (24.92 (3.60), and 42 older adults (62.84 (7.02)). Additionally, a subsample of 111 participants completed a resting-state fMRI scan. After investigating the impact of age on different divergent thinking metrics, we investigated the impact of age on the association between divergent thinking and resting-state functional connectivity within and between major resting-state brain networks associated with creative thinking: the DMN, ECN, and SN. Adolescents tended to be less creative than both young and older adults in divergent thinking scores related to expansion creativity, and not in persistent creativity, while young and older adults performed relatively similar. We found that adolescents' functional integrity of the executive control network (ECN) was positively associated with expansion creativity, which was significantly different from the negative association in both the young and older adults. These results suggest that creative performance and supporting brain networks change throughout the lifespan.

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.

Disordered network structure and function in dystonia: pathological connectivity vs. adaptive responses

Primary dystonia is thought to emerge through abnormal functional relationships between basal ganglia and cerebellar motor circuits. These interactions may differ across disease subtypes and provide a novel biomarker for diagnosis and treatment. Using a network mapping algorithm based on resting-state functional MRI (rs-fMRI), a method that is readily implemented on conventional MRI scanners, we identified similar disease topographies in hereditary dystonia associated with the DYT1 or DYT6 mutations and in sporadic patients lacking these mutations. Both networks were characterized by contributions from the basal ganglia, cerebellum, thalamus, sensorimotor areas, as well as cortical association regions. Expression levels for the two networks were elevated in hereditary and sporadic dystonia, and in non-manifesting carriers of dystonia mutations. Nonetheless, the distribution of abnormal functional connections differed across groups, as did metrics of network organization and efficiency in key modules. Despite these differences, network expression correlated with dystonia motor ratings, significantly improving the accuracy of predictions based on thalamocortical tract integrity obtained with diffusion tensor MRI (DTI). Thus, in addition to providing unique information regarding the anatomy of abnormal brain circuits, rs-fMRI functional networks may provide a widely accessible method to help in the objective evaluation of new treatments for this disorder.

Multivariate associations between dopamine receptor availability and risky investment decision-making across adulthood

Enhancing dopamine increases financial risk taking across adulthood but it is unclear whether baseline individual differences in dopamine function are related to risky financial decisions. Here, thirty-five healthy adults completed an incentive-compatible risky investment decision task and a PET scan at rest using [11C]FLB457 to assess dopamine D2-like receptor availability. Participants made choices between a safe asset (bond) and a risky asset (stock) with either an expected value less than the bond ("bad stock") or expected value greater than the bond ("good stock"). Five measures of behavior (choice inflexibility, risk seeking, suboptimal investment) and beliefs (absolute error, optimism) were computed and D2-like binding potential was extracted from four brain regions of interest (midbrain, amygdala, anterior cingulate, insula). We used canonical correlation analysis to evaluate multivariate associations between decision-making and dopamine function controlling for age. Decomposition of the first dimension (r = 0.76) revealed that the strongest associations were between measures of choice inflexibility, incorrect choice, optimism, amygdala binding potential, and age. Follow-up univariate analyses revealed that amygdala binding potential and age were both independently associated with choice inflexibility. The findings suggest that individual differences in dopamine function may be associated with financial risk taking in healthy adults.

How the body remembers: Examining the default mode and sensorimotor networks during moral injury autobiographical memory retrieval in PTSD

Neural representations of sensory percepts and motor responses constitute key elements of autobiographical memory. However, these representations may remain as unintegrated sensory and motor fragments in traumatic memory, thus contributing toward re-experiencing and reliving symptoms in trauma-related conditions such as post-traumatic stress disorder (PTSD). Here, we investigated the sensorimotor network (SMN) and posterior default mode network (pDMN) using a group independent component analysis (ICA) by examining their functional connectivity during a script-driven memory retrieval paradigm of (potentially) morally injurious events in individuals with PTSD and healthy controls. Moral injury (MI), where an individual acts or fails to act in a morally aligned manner, is examined given its inherent ties to disrupted motor planning and thus sensorimotor mechanisms. Our findings revealed significant differences in functional network connectivity across the SMN and pDMN during MI retrieval in participants with PTSD (n = 65) as compared to healthy controls (n = 25). No such significant group-wise differences emerged during retrieval of a neutral memory. PTSD-related alterations included hyperconnectivity between the SMN and pDMN, enhanced within-network connectivity of the SMN with premotor areas, and increased recruitment of the supramarginal gyrus into both the SMN and the pDMN during MI retrieval. In parallel with these neuroimaging findings, a positive correlation was found between PTSD severity and subjective re-experiencing intensity ratings after MI retrieval. These results suggest a neural basis for traumatic re-experiencing, where reliving and/or re-enacting a past morally injurious event in the form of sensory and motor fragments occurs in place of retrieving a complete, past-contextualized narrative as put forth by Brewin and colleagues (1996) and Conway and Pleydell-Pearce (2000). These findings have implications for bottom-up treatments targeting directly the sensory and motoric elements of traumatic experiences.

Differences in the Default Mode Network of Temporal Lobe Epilepsy Patients Detected by Hilbert-Huang Transform Based Dynamic Functional Connectivity

Resting-state functional connectivity, constructed via functional magnetic resonance imaging, has become an essential tool for exploring brain functions. Aside from the methods focusing on the static state, investigating dynamic functional connectivity can better uncover the fundamental properties of brain networks. Hilbert-Huang transform (HHT) is a novel time-frequency technique that can adapt to both non-linear and non-stationary signals, which may be an effective tool for investigating dynamic functional connectivity. To perform the present study, we investigated time-frequency dynamic functional connectivity among 11 brain regions of the default mode network by first projecting the coherence into the time and frequency domains, and subsequently by identifying clusters in the time-frequency domain using k-means clustering. Experiments on 14 temporal lobe epilepsy (TLE) patients and 21 age and sex-matched healthy controls were performed. The results show that functional connections in the brain regions of the hippocampal formation, parahippocampal gyrus, and retrosplenial cortex (Rsp) were reduced in the TLE group. However, the connections in the brain regions of the posterior inferior parietal lobule, ventral medial prefrontal cortex, and the core subsystem could hardly be detected in TLE patients. The findings not only demonstrate the feasibility of utilizing HHT in dynamic functional connectivity for epilepsy research, but also indicate that TLE may cause damage to memory functions, disorders of processing self-related tasks, and impairment of constructing a mental scene.

Identification of overlapping and interacting networks reveals intrinsic spatiotemporal organization of the human brain

The human brain is a complex network that exhibits dynamic fluctuations in activity across space and time. Depending on the analysis method, canonical brain networks identified from resting-state fMRI (rs-fMRI) are typically constrained to be either orthogonal or statistically independent in their spatial and/or temporal domains. We avoid imposing these potentially unnatural constraints through the combination of a temporal synchronization process ("BrainSync") and a three-way tensor decomposition method ("NASCAR") to jointly analyze rs-fMRI data from multiple subjects. The resulting set of interacting networks comprises minimally constrained spatiotemporal distributions, each representing one component of functionally coherent activity across the brain. We show that these networks can be clustered into six distinct functional categories and naturally form a representative functional network atlas for a healthy population. This functional network atlas could help explore group and individual differences in neurocognitive function, as we demonstrate in the context of ADHD and IQ prediction.

Federated Analysis in COINSTAC Reveals Functional Network Connectivity and Spectral Links to Smoking and Alcohol Consumption in Nearly 2,000 Adolescent Brains

With the growth of decentralized/federated analysis approaches in neuroimaging, the opportunities to study brain disorders using data from multiple sites has grown multi-fold. One such initiative is the Neuromark, a fully automated spatially constrained independent component analysis (ICA) that is used to link brain network abnormalities among different datasets, studies, and disorders while leveraging subject-specific networks. In this study, we implement the neuromark pipeline in COINSTAC, an open-source neuroimaging framework for collaborative/decentralized analysis. Decentralized exploratory analysis of nearly 2000 resting-state functional magnetic resonance imaging datasets collected at different sites across two cohorts and co-located in different countries was performed to study the resting brain functional network connectivity changes in adolescents who smoke and consume alcohol. Results showed hypoconnectivity across the majority of networks including sensory, default mode, and subcortical domains, more for alcohol than smoking, and decreased low frequency power. These findings suggest that global reduced synchronization is associated with both tobacco and alcohol use. This proof-of-concept work demonstrates the utility and incentives associated with large-scale decentralized collaborations spanning multiple sites.

Brain connectomics: time for a molecular imaging perspective?

In the past two decades brain connectomics has evolved into a major concept in neuroscience. However, the current perspective on brain connectivity and how it underpins brain function relies mainly on the hemodynamic signal of functional magnetic resonance imaging (MRI). Molecular imaging provides unique information inaccessible to MRI-based and electrophysiological techniques. Thus, positron emission tomography (PET) has been successfully applied to measure neural activity, neurotransmission, and proteinopathies in normal and pathological cognition. Here, we position molecular imaging within the brain connectivity framework from the perspective of timeliness, validity, reproducibility, and resolution. We encourage the neuroscientific community to take an integrative approach whereby MRI-based, electrophysiological techniques, and molecular imaging contribute to our understanding of the brain connectome.

Tensorial independent component analysis reveals social and reward networks associated with major depressive disorder

Major depressive disorder (MDD) has been associated with changes in functional brain connectivity. Yet, typical analyses of functional connectivity, such as spatial independent components analysis (ICA) for resting-state data, often ignore sources of between-subject variability, which may be crucial for identifying functional connectivity patterns associated with MDD. Typically, methods like spatial ICA will identify a single component to represent a network like the default mode network (DMN), even if groups within the data show differential DMN coactivation. To address this gap, this project applies a tensorial extension of ICA (tensorial ICA)-which explicitly incorporates between-subject variability-to identify functionally connected networks using functional MRI data from the Human Connectome Project (HCP). Data from the HCP included individuals with a diagnosis of MDD, a family history of MDD, and healthy controls performing a gambling and social cognition task. Based on evidence associating MDD with blunted neural activation to rewards and social stimuli, we predicted that tensorial ICA would identify networks associated with reduced spatiotemporal coherence and blunted social and reward-based network activity in MDD. Across both tasks, tensorial ICA identified three networks showing decreased coherence in MDD. All three networks included ventromedial prefrontal cortex, striatum, and cerebellum and showed different activation across the conditions of their respective tasks. However, MDD was only associated with differences in task-based activation in one network from the social task. Additionally, these results suggest that tensorial ICA could be a valuable tool for understanding clinical differences in relation to network activation and connectivity.

Cognitive domain-independent aberrant frontoparietal network strength in individuals with excessive smartphone use

Excessive smartphone use (ESU) may fulfill criteria for addictive behavior. In contrast to other related behavioral addictions, particularly Internet Gaming Disorder, little is known about the neural correlates underlying ESU. In this study, we used functional magnetic resonance imaging (fMRI) to acquire task data from three distinct behavioral paradigms, i.e. cue-reactivity, inhibition, and working memory, in individuals with psychometrically defined ESU (n = 19) compared to controls (n-ESU; n = 20). The Smartphone Addiction Inventory (SPAI) was used to quantify ESU-severity according to a novel five-factor model (SPAI-I). A multivariate data fusion approach, i.e. joint Independent Component Analysis (jICA) was employed to analyze fMRI-data derived from three separate experimental conditions, but analyzed jointly to detect converging and domain-independent neural signatures that differ between persons with vs. those without ESU. Across the three functional tasks, jICA identified a predominantly frontoparietal system that showed lower network strength in individuals with ESU compared to n-ESU (p < 0.05 FDR-corrected). Furthermore, significant associations between frontoparietal network strength and SPAI-I's dimensions "time spent" and "craving" were found. The data suggest a frontoparietal cognitive control network as cognitive domain-independent neural signature of excessive and potentially addictive smartphone use.

Brain Network Analysis: A Review on Multivariate Analytical Methods

Despite the explosive growth of neuroimaging studies aimed at analyzing the brain as a complex system, critical methodological gaps remain to be addressed. Most tools currently used for analyzing network data of the brain are univariate in nature and are based on assumptions borne out of previous techniques not directly related to the big and complex data of the brain. Although graph-based methods have shown great promise, the development of principled multivariate models to address inherent limitations of graph-based methods, such as their dependence on network size and degree distributions, and to allow assessing the effects of multiple phenotypes on the brain and simulating brain networks has largely lagged behind. Although some studies have been made in developing multivariate frameworks to fill this gap, in the absence of a "gold-standard" method or guidelines, choosing the most appropriate method for each study can be another critical challenge for investigators in this multidisciplinary field. Here, we briefly introduce important multivariate methods for brain network analyses in two main categories: data-driven and model-based methods. We discuss whether/how such methods are suited for examining connectivity (edge-level), topology (system-level), or both. This review will aid in choosing an appropriate multivariate method with respect to variables such as network type, number of subjects and brain regions included, and the interest in connectivity, topology, or both. This review is aimed to be accessible to investigators from different backgrounds, with a focus on applications in brain network studies, though the methods may be applicable in other areas too. Impact statement As the U.S. National Institute of Health notes, the rich biomedical data can greatly improve our knowledge of human health if new analytical tools are developed, and their applications are broadly disseminated. A major challenge in analyzing the brain as a complex system is about developing parsimonious multivariate methods, and particularly choosing the most appropriate one among the existing methods with respect to the study variables in this multidisciplinary field. This study provides a review on the most important multivariate methods to aid in helping the most appropriate ones with respect to the desired variables for each study.

Multimodal fusion of multiple rest fMRI networks and MRI gray matter via multilink joint ICA reveals highly significant function/structure coupling in Alzheimer's disease

In this paper we focus on estimating the joint relationship between structural MRI (sMRI) gray matter (GM) and multiple functional MRI (fMRI) intrinsic connectivity networks (ICN) using a novel approach called multi-link joint independent component analysis (ml-jICA). The proposed model offers several improvements over the existing joint independent component analysis (jICA) model. We assume a shared mixing matrix for both the sMRI and fMRI modalities, while allowing for different mixing matrices linking the sMRI data to the different ICNs. We introduce the model and then apply this approach to study the differences in resting fMRI and sMRI data from patients with Alzheimer's disease (AD) versus controls. The results yield significant differences with large effect sizes that include regions in overlapping portions of default mode network, and also hippocampus and thalamus. Importantly, we identify two joint components with partially overlapping regions which show opposite effects for Alzheimer's disease versus controls, but were able to be separated due to being linked to distinct functional and structural patterns. This highlights the unique strength of our approach and multimodal fusion approaches generally in revealing potentially biomarkers of brain disorders that would likely be missed by a unimodal approach. These results represent the first work linking multiple fMRI ICNs to gray matter components within a multimodal data fusion model and challenges the typical view that brain structure is more sensitive to AD than fMRI.

The Mediating Role of Brain Structural Imaging Markers in Connecting Adverse Childhood Experiences and Psychological Resilience

The impact of adverse childhood experiences (ACEs) on brain structure has been noticed. Resilience has been considered a protective characteristic from being mentally ill; however, the link between ACEs, psychological resilience, and brain imaging remains untested. A total of 108 participants (mean age 22.92 ± 2.43 years) completed the ACEs questionnaire and the Resilience Scale for Adults (RSA), with five subscales: personal strength (RSA_ps), family cohesion (RSA_fc), social resources (RSA_sr), social competence (RSA_sc), and future structured style (RSA_fss), and Magnetic Resonance Imaging (MRI) to acquire imaging data, and the fusion-independent component analysis was employed to determine multimodal imaging components. The results showed a significantly negative association between ACE subscales and RSA_total score (ps < 0.05). The parallel mediation model showed significant indirect mediation of mean gray matter volumes in the regions of the middle frontal gyrus, superior frontal gyrus, posterior cingulate, superior temporal gyrus, middle temporal gyrus, postcentral gyrus, middle temporal gyrus, and precuneus between childhood maltreatment and RSA_sr and RSA_sc. (ps < 0.05). This study highlighted the ACEs effect on gray matter volumes in the regions of the middle frontal gyrus, superior frontal gyrus, posterior cingulate, superior temporal gyrus, middle temporal gyrus, postcentral gyrus, middle temporal gyrus, and precuneus leading to decreased psychological resilience.

Functional brain networks in the evaluation of patients with neurodegenerative disorders

Network analytical tools are increasingly being applied to brain imaging maps of resting metabolic activity (PET) or blood oxygenation-dependent signals (functional MRI) to characterize the abnormal neural circuitry that underlies brain diseases. This approach is particularly valuable for the study of neurodegenerative disorders, which are characterized by stereotyped spread of pathology along discrete neural pathways. Identification and validation of disease-specific brain networks facilitate the quantitative assessment of pathway changes over time and during the course of treatment. Network abnormalities can often be identified before symptom onset and can be used to track disease progression even in the preclinical period. Likewise, network activity can be modulated by treatment and might therefore be used as a marker of efficacy in clinical trials. Finally, early differential diagnosis can be achieved by simultaneously measuring the activity levels of multiple disease networks in an individual patient's scans. Although these techniques were originally developed for PET, over the past several years analogous methods have been introduced for functional MRI, a more accessible non-invasive imaging modality. This advance is expected to broaden the application of network tools to large and diverse patient populations.

Parkinson's Disease Diagnosis beyond Clinical Features: A Bio-marker using Topological Machine Learning of Resting-state Functional Magnetic Resonance Imaging

Parkinson's disease (PD) is one of the leading causes of neurological disability, and its prevalence is expected to increase rapidly in the following few decades. PD diagnosis heavily depends on clinical features using the patient's symptoms. Therefore, an accurate, robust, and non-invasive bio-marker is of critical clinical importance for PD. This study proposes to develop a new bio-marker for PD diagnosis using resting-state functional Magnetic Resonance Imaging (rs-fMRI). Unlike most existing rs-fMRI data analytics using correlational analysis, a Topological Machine Learning approach is proposed to construct the bio-marker. The default functional network is identified first using rs-fMRI. Next, rs-fMRI's high dimensional spatial-temporal data structure is mapped on a Riemann Manifold using topological dimensional reduction. Following the topological dimensional reduction, machine learning is used for classification and sensitivity analysis. The proposed methodology is applied to three open fMRI databases for demonstration and validation. The PD diagnosis accuracy can reach 96.4% when the proposed methodology is used. Thus, rs-fMRI and topological machine learning provide a quantifiable and verifiable bio-marker for future PD early detection and treatment evaluation.

Rapid Interactions of Widespread Brain Networks Characterize Semantic Cognition

Language comprehension requires the rapid retrieval and integration of contextually appropriate concepts ("semantic cognition"). Current neurobiological models of semantic cognition are limited by the spatial and temporal restrictions of single-modality neuroimaging and lesion approaches. This is a major impediment given the rapid sequence of processing steps that have to be coordinated to accurately comprehend language. Through the use of fused functional magnetic resonance imaging and electroencephalography analysis in humans (n = 26 adults; 15 females), we elucidate a temporally and spatially specific neurobiological model for real-time semantic cognition. We find that semantic cognition in the context of language comprehension is supported by trade-offs between widespread neural networks over the course of milliseconds. Incorporation of spatial and temporal characteristics, as well as behavioral measures, provide convergent evidence for the following progression: a hippocampal/anterior temporal phonological semantic retrieval network (peaking at ∼300 ms after the sentence final word); a frontotemporal thematic semantic network (∼400 ms); a hippocampal memory update network (∼500 ms); an inferior frontal semantic syntactic reappraisal network (∼600 ms); and nodes of the default mode network associated with conceptual coherence (∼750 ms). Additionally, in typical adults, mediatory relationships among these networks are significantly predictive of language comprehension ability. These findings provide a conceptual and methodological framework for the examination of speech and language disorders, with additional implications for the characterization of cognitive processes and clinical populations in other cognitive domains.SIGNIFICANCE STATEMENT The present study identifies a real-time neurobiological model of the meaning processes required during language comprehension (i.e., "semantic cognition"). Using a novel application of fused magnetic resonance imaging and electroencephalography in humans, we found that semantic cognition during language comprehension is supported by a rapid progression of widespread neural networks related to meaning, meaning integration, memory, reappraisal, and conceptual cohesion. Relationships among these systems were predictive of individuals' language comprehension efficiency. Our findings are the first to use fused neuroimaging analysis to elucidate language processes. In so doing, this study provides a new conceptual and methodological framework in which to characterize language processes and guide the treatment of speech and language deficits/disorders.

Increased top-down control of emotions during symptom provocation working memory tasks following a RCT of alpha-down neurofeedback in PTSD

BACKGROUND

Posttraumatic stress disorder (PTSD) has been found to be associated with emotion under-modulation from the prefrontal cortex and a breakdown of the top-down control of cognition and emotion. Novel adjunct therapies such as neurofeedback (NFB) have been shown to normalize aberrant neural circuits that underlie PTSD psychopathology at rest. However, little evidence exists for NFB-linked neural improvements under emotionally relevant cognitive load. The current study sought to address this gap by examining the effects of alpha-down NFB in the context of an emotional n-back task.

METHODS

We conducted a 20-week double-blind randomized, sham-controlled trial of alpha-down NFB and collected neuroimaging data before and after the NFB protocol. Participants performed an emotional 1-back and 2-back working memory task, with interleaved trauma-neutral and trauma-relevant cues in the fMRI scanner. Data from 35 participants with a primary diagnosis of PTSD were analyzed in this study (n = 18 in the experimental group undergoing alpha-down NFB, n = 17 in the sham-control group).

RESULTS

Firstly, within-group analyses showed clinically significant reductions in PTSD symptom severity scores at the post-intervention timepoint and 3-month follow-up for the experimental group, and not for the sham-control group. The neuroimaging analyses revealed that alpha-down NFB enhanced engagement of top-down cognitive and emotional control centers, such as the dorsolateral prefrontal cortex (dlPFC), and improved integration of the anterior and posterior parts of the default mode network (DMN). Finally, our results also indicate that increased alpha-down NFB performance correlated with increased activity in brain regions involved in top-down control and bodily consciousness/embodied processing of self (TPJ and posterior insula).

CONCLUSION

This is the first study to provide mechanistic insights into how NFB may normalize dysfunctional brain activity and connectivity in PTSD under cognitive load with simultaneous symptom provocation, adding to a growing body of evidence supporting the therapeutic neuromodulatory effects of NFB. This preliminary study highlights the benefits of alpha-down NFB training as an adjunctive therapy for PTSD and warrants further investigation into its therapeutic effects on cognitive and emotion control in those with PTSD.

A Review of Fusion Methods for Omics and Imaging Data

The development of omics data and biomedical images has greatly advanced the progress of precision medicine in diagnosis, treatment, and prognosis. The fusion of omics and imaging data, i.e., omics-imaging fusion, offers a new strategy for understanding complex diseases. However, due to a variety of issues such as the limited number of samples, high dimensionality of features, and heterogeneity of different data types, efficiently learning complementary or associated discriminative fusion information from omics and imaging data remains a challenge. Recently, numerous machine learning methods have been proposed to alleviate these problems. In this review, from the perspective of fusion levels and fusion methods, we first provide an overview of preprocessing and feature extraction methods for omics and imaging data, and comprehensively analyze and summarize the basic forms and variations of commonly used and newly emerging fusion methods, along with their advantages, disadvantages and the applicable scope. We then describe public datasets and compare experimental results of various fusion methods on the ADNI and TCGA datasets. Finally, we discuss future prospects and highlight remaining challenges in the field.

Multi-Subject Analysis for Brain Developmental Patterns Discovery via Tensor Decomposition of MEG Data

Identification of informative signatures from electrophysiological signals is important for understanding brain developmental patterns, where techniques such as magnetoencephalography (MEG) are particularly useful. However, less attention has been given to fully utilizing the multidimensional nature of MEG data for extracting components that describe these patterns. Tensor factorizations of MEG yield components that encapsulate the data's multidimensional nature, providing parsimonious models identifying latent brain patterns for meaningful summarization of neural processes. To address the need for meaningful MEG signatures for studies of pediatric cohorts, we propose a tensor-based approach for extracting developmental signatures of multi-subject MEG data. We employ the canonical polyadic (CP) decomposition for estimating latent spatiotemporal components of the data, and use these components for group level statistical inference. Using CP decomposition along with hierarchical clustering, we were able to extract typical early and late latency event-related field (ERF) components that were discriminative of high and low performance groups ([Formula: see text]) and significantly correlated with major cognitive domains such as attention, episodic memory, executive function, and language comprehension. We demonstrate that tensor-based group level statistical inference of MEG can produce signatures descriptive of the multidimensional MEG data. Furthermore, these features can be used to study group differences in brain patterns and cognitive function of healthy children. We provide an effective tool that may be useful for assessing child developmental status and brain function directly from electrophysiological measurements and facilitate the prospective assessment of cognitive processes.

The Role of fMRI in Drug Development: An Update

Functional magnetic resonance imaging (fMRI) of the brain is a technology that holds great potential for increasing the efficiency of drug development for the central nervous system (CNS). In preclinical studies and both early- and late-phase human trials, fMRI has the potential to improve cross-species translation of drug effects, help to de-risk compounds early in development, and contribute to the portfolio of evidence for a compound's efficacy and mechanism of action. However, to date, the utilization of fMRI in the CNS drug development process has been limited. The purpose of this chapter is to explore this mismatch between potential and utilization. This chapter provides introductory material related to fMRI and drug development, describes what is required of fMRI measurements for them to be useful in a drug development setting, lists current capabilities of fMRI in this setting and challenges faced in its utilization, and ends with directions for future development of capabilities in this arena. This chapter is the 5-year update of material from a previously published workshop summary (Carmichael et al., Drug DiscovToday 23(2):333-348, 2018).

Changes in Functional Connectivity Between Default Mode Network and Attention Network in Response to Changes in Aerobic Exercise Intensity

OBJECTIVE

Aerobic exercise may be associated with changes in brain activity within the default mode network (DMN) and dorsal attention network (DAN). We hypothesized that changes in functional connectivity (FC) within the DMN and DAN might be most effectively activated by moderate-intensity exercise.

METHODS

Resting-state functional magnetic resonance imaging scans and visuospatial attention tests after resting were performed before and after each of moderate- and high-intensity aerobic exercises (10 min each) in 15 healthy male volunteers.

RESULTS

The reaction time during the attention test increased significantly, and the rate of correct responses decreased from moderate-intensity exercise condition to high-intensity exercise condition. FC within the DMN under high-intensity exercise condition was higher than that under pre-exercise and moderate-intensity exercise conditions. FC within the DAN under moderate-intensity exercise condition was the highest, whereas FC between the DMN and DAN under moderate-intensity exercise condition was the lowest. Changes in cognitive domain functions were associated with changes in FC between the DMN and DAN.

CONCLUSION

Our results support the inverted-U hypothesis of maximum arousal efficacy during moderate exercise. Both cognitive domains, namely, the attention system and brain activity domains, may be better under moderate-intensity exercise than under high-intensity exercise.

Understanding the neurodynamic process of decision-making for mobile application downloading

In this article, we try to explore and understand the neurodynamics of the decision-making process for mobile application downloading. We begin the model development in a rather unorthodox fashion. Patterns of brain activation regions are identified, across participants, at different time instance of the decision-making process. Region-wise activation knowledge from previous studies is used to put together the entire process model like a cognitive jigsaw puzzle. We find that there are indeed a common dynamic set of activation patterns that are consistent across people and apps. That is to say that not only are there consistent patterns of activation there is a consistent change from one pattern to another across time as people make the app adoption decision. Moreover, this pattern is clearly different for decisions that end in adoption than for decisions that end with no adoption.

Clusterwise Independent Component Analysis (C-ICA): Using fMRI resting state networks to cluster subjects and find neurofunctional subtypes

BACKGROUND

FMRI resting state networks (RSNs) are used to characterize brain disorders. They also show extensive heterogeneity across patients. Identifying systematic differences between RSNs in patients, i.e. discovering neurofunctional subtypes, may further increase our understanding of disease heterogeneity. Currently, no methodology is available to estimate neurofunctional subtypes and their associated RSNs simultaneously.

NEW METHOD

We present an unsupervised learning method for fMRI data, called Clusterwise Independent Component Analysis (C-ICA). This enables the clustering of patients into neurofunctional subtypes based on differences in shared ICA-derived RSNs. The parameters are estimated simultaneously, which leads to an improved estimation of subtypes and their associated RSNs.

RESULTS

In five simulation studies, the C-ICA model is successfully validated using both artificially and realistically simulated data (N = 30-40). The successful performance of the C-ICA model is also illustrated on an empirical data set consisting of Alzheimer's disease patients and elderly control subjects (N = 250). C-ICA is able to uncover a meaningful clustering that partially matches (balanced accuracy = .72) the diagnostic labels and identifies differences in RSNs between the Alzheimer and control cluster.

COMPARISON WITH OTHER METHODS

Both in the simulation study and the empirical application, C-ICA yields better results compared to competing clustering methods (i.e., a two step clustering procedure based on single subject ICA's and a Group ICA plus dual regression variant thereof) that do not simultaneously estimate a clustering and associated RSNs. Indeed, the overall mean adjusted Rand Index, a measure for cluster recovery, equals 0.65 for C-ICA and ranges from 0.27 to 0.46 for competing methods.

CONCLUSIONS

The successful performance of C-ICA indicates that it is a promising method to extract neurofunctional subtypes from multi-subject resting state-fMRI data. This method can be applied on fMRI scans of patient groups to study (neurofunctional) subtypes, which may eventually further increase understanding of disease heterogeneity.

Inter-individual differences in baseline dynamic functional connectivity are linked to cognitive aftereffects of tDCS

Transcranial direct current stimulation (tDCS) has the potential to modulate cognitive training in healthy aging; however, results from various studies have been inconsistent. We hypothesized that inter-individual differences in baseline brain state may contribute to the varied results. We aimed to explore whether baseline resting-state dynamic functional connectivity (rs-dFC) and/or conventional resting-state static functional connectivity (rs-sFC) may be related to the magnitude of cognitive aftereffects of tDCS. To achieve this aim, we used data from our double-blind randomized sham-controlled cross-over tDCS trial in 25 healthy seniors in which bifrontal tDCS combined with cognitive training had induced significant behavioral aftereffects. We performed a backward regression analysis including rs-sFC/rs-dFC measures to explain the variability in the magnitude of tDCS-induced improvements in visual object-matching task (VOMT) accuracy. Rs-dFC analysis revealed four rs-dFC states. The occurrence rate of a rs-dFC state 4, characterized by a high correlation between the left fronto-parietal control network and the language network, was significantly associated with tDCS-induced VOMT accuracy changes. The rs-sFC measure was not significantly associated with the cognitive outcome. We show that flexibility of the brain state representing readiness for top-down control of object identification implicated in the studied task is linked to the tDCS-enhanced task accuracy.

Mapping the effects of pregnancy on resting state brain activity, white matter microstructure, neural metabolite concentrations and grey matter architecture

While animal studies have demonstrated a unique reproduction-related neuroplasticity, little is known on the effects of pregnancy on the human brain. Here we investigated whether pregnancy is associated with changes to resting state brain activity, white matter microstructure, neural metabolite concentrations and grey matter architecture using a comprehensive pre-conception cohort study. We show that pregnancy leads to selective and robust changes in neural architecture and neural network organization, which are most pronounced in the Default Mode Network. These neural changes correlated with pregnancy hormones, primarily third-trimester estradiol, while no associations were found with other factors such as osmotic effects, stress and sleep. Furthermore, the changes related to measures of maternal-fetal bonding, nesting behavior and the physiological responsiveness to infant cues, and predicted measures of mother-infant bonding and bonding impairments. These findings suggest there are selective pregnancy-related modifications in brain structure and function that may facilitate peripartum maternal processes of key relevance to the mother-infant dyad.

Modern views of machine learning for precision psychiatry

In light of the National Institute of Mental Health (NIMH)'s Research Domain Criteria (RDoC), the advent of functional neuroimaging, novel technologies and methods provide new opportunities to develop precise and personalized prognosis and diagnosis of mental disorders. Machine learning (ML) and artificial intelligence (AI) technologies are playing an increasingly critical role in the new era of precision psychiatry. Combining ML/AI with neuromodulation technologies can potentially provide explainable solutions in clinical practice and effective therapeutic treatment. Advanced wearable and mobile technologies also call for the new role of ML/AI for digital phenotyping in mobile mental health. In this review, we provide a comprehensive review of ML methodologies and applications by combining neuroimaging, neuromodulation, and advanced mobile technologies in psychiatry practice. We further review the role of ML in molecular phenotyping and cross-species biomarker identification in precision psychiatry. We also discuss explainable AI (XAI) and neuromodulation in a closed human-in-the-loop manner and highlight the ML potential in multi-media information extraction and multi-modal data fusion. Finally, we discuss conceptual and practical challenges in precision psychiatry and highlight ML opportunities in future research.

Evidence for independent representational contents in inhibitory control subprocesses associated with frontoparietal cortices

Inhibitory control processes have intensively been studied in cognitive science for the past decades. Even though the neural dynamics underlying these processes are increasingly better understood, a critical open question is how the representational dynamics of the inhibitory control processes are modulated when engaging in response inhibition in a relatively automatic or a controlled mode. Against the background of an overarching theory of perception-action integration, we combine temporal and spatial EEG signal decomposition methods with multivariate pattern analysis and source localization to obtain fine-grained insights into the neural dynamics of the representational content of response inhibition. For this purpose, we used a sample of N = 40 healthy adult participants. The behavioural data suggest that response inhibition was better in a more controlled than a more automated response execution mode. Regarding neural dynamics, effects of response inhibition modes relied on a concomitant coding of stimulus-related information and rules of how stimulus information is related to the appropriate motor programme. Crucially, these fractions of information, which are encoded at the same time in the neurophysiological signal, are based on two independent spatial neurophysiological activity patterns, also showing differences in the temporal stability of the representational content. Source localizations revealed that the precuneus and inferior parietal cortex regions are more relevant than prefrontal areas for the representation of stimulus-response selection codes. We provide a blueprint how a concatenation of EEG signal analysis methods, capturing distinct aspects of neural dynamics, can be connected to cognitive science theory on the importance of representations in action control.

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.

The expanding horizons of network neuroscience: From description to prediction and control

The field of network neuroscience has emerged as a natural framework for the study of the brain and has been increasingly applied across divergent problems in neuroscience. From a disciplinary perspective, network neuroscience originally emerged as a formal integration of graph theory (from mathematics) and neuroscience (from biology). This early integration afforded marked utility in describing the interconnected nature of neural units, both structurally and functionally, and underscored the relevance of that interconnection for cognition and behavior. But since its inception, the field has not remained static in its methodological composition. Instead, it has grown to use increasingly advanced graph-theoretic tools and to bring in several other disciplinary perspectives-including machine learning and systems engineering-that have proven complementary. In doing so, the problem space amenable to the discipline has expanded markedly. In this review, we discuss three distinct flavors of investigation in state-of-the-art network neuroscience: (i) descriptive network neuroscience, (ii) predictive network neuroscience, and (iii) a perturbative network neuroscience that draws on recent advances in network control theory. In considering each area, we provide a brief summary of the approaches, discuss the nature of the insights obtained, and highlight future directions.

Decoding Task-Based fMRI Data with Graph Neural Networks, Considering Individual Differences

Task fMRI provides an opportunity to analyze the working mechanisms of the human brain during specific experimental paradigms. Deep learning models have increasingly been applied for decoding and encoding purposes study to representations in task fMRI data. More recently, graph neural networks, or neural networks models designed to leverage the properties of graph representations, have recently shown promise in task fMRI decoding studies. Here, we propose an end-to-end graph convolutional network (GCN) framework with three convolutional layers to classify task fMRI data from the Human Connectome Project dataset. We compared the predictive performance of our GCN model across four of the most widely used node embedding algorithms-NetMF, RandNE, Node2Vec, and Walklets-to automatically extract the structural properties of the nodes in the functional graph. The empirical results indicated that our GCN framework accurately predicted individual differences (0.978 and 0.976) with the NetMF and RandNE embedding methods, respectively. Furthermore, to assess the effects of individual differences, we tested the classification performance of the model on sub-datasets divided according to gender and fluid intelligence. Experimental results indicated significant differences in the classification predictions of gender, but not high/low fluid intelligence fMRI data. Our experiments yielded promising results and demonstrated the superior ability of our GCN in modeling task fMRI data.

The multiple indicator multiple cause model for cognitive neuroscience: An analytic tool which emphasizes the behavior in brain–behavior relationships

Cognitive neuroscience has inspired a number of methodological advances to extract the highest signal-to-noise ratio from neuroimaging data. Popular techniques used to summarize behavioral data include sum-scores and item response theory (IRT). While these techniques can be useful when applied appropriately, item dimensionality and the quality of information are often left unexplored allowing poor performing items to be included in an itemset. The purpose of this study is to highlight how the application of two-stage approaches introduces parameter bias, differential item functioning (DIF) can manifest in cognitive neuroscience data and how techniques such as the multiple indicator multiple cause (MIMIC) model can identify and remove items with DIF and model these data with greater sensitivity for brain–behavior relationships. This was performed using a simulation and an empirical study. The simulation explores parameter bias across two separate techniques used to summarize behavioral data: sum-scores and IRT and formative relationships with those estimated from a MIMIC model. In an empirical study participants performed an emotional identification task while concurrent electroencephalogram data were acquired across 384 trials. Participants were asked to identify the emotion presented by a static face of a child across four categories: happy, neutral, discomfort, and distress. The primary outcomes of interest were P200 event-related potential (ERP) amplitude and latency within each emotion category. Instances of DIF related to correct emotion identification were explored with respect to an individual’s neurophysiology; specifically an item’s difficulty and discrimination were explored with respect to an individual’s average P200 amplitude and latency using a MIMIC model. The MIMIC model’s sensitivity was then compared to popular two-stage approaches for cognitive performance summary scores, including sum-scores and an IRT model framework and then regressing these onto the ERP characteristics. Here sensitivity refers to the magnitude and significance of coefficients relating the brain to these behavioral outcomes. The first set of analyses displayed instances of DIF within all four emotions which were then removed from all further models. The next set of analyses compared the two-stage approaches with the MIMIC model. Only the MIMIC model identified any significant brain–behavior relationships. Taken together, these results indicate that item performance can be gleaned from subject-specific biomarkers, and that techniques such as the MIMIC model may be useful tools to derive complex item-level brain–behavior relationships.

Learning brain representation using recurrent Wasserstein generative adversarial net

BACKGROUND AND OBJECTIVE

To understand brain cognition and disorders, modeling the mapping between mind and brain has been of great interest to the neuroscience community. The key is the brain representation, including functional brain networks (FBN) and their corresponding temporal features. Recently, it has been proven that deep learning models have superb representation power on functional magnetic resonance imaging (fMRI) over traditional machine learning methods. However, due to the lack of high-quality data and labels, deep learning models tend to suffer from overfitting in the training process.

METHODS

In this work, we applied a recurrent Wasserstein generative adversarial net (RWGAN) to learn brain representation from volumetric fMRI data. Generative adversarial net (GAN) is widely used in natural image generation and is able to capture the distribution of the input data, which enables the extraction of generalized features from fMRI and thus relieves the overfitting issue. The recurrent layers in RWGAN are designed to better model the local temporal features of the fMRI time series. The discriminator of RWGAN works as a deep feature extractor. With LASSO regression, the RWGAN model can decompose the fMRI data into temporal features and spatial features (FBNs). Furthermore, the generator of RWGAN can generate high-quality new data for fMRI augmentation.

RESULTS

The experimental results on seven tasks from the HCP dataset showed that the RWGAN can learn meaningful and interpretable temporal features and FBNs, compared to HCP task designs and general linear model (GLM) derived networks. Besides, the results on different training datasets showed that the RWGAN performed better on small datasets than other deep learning models. Moreover, we used the generator of RWGAN to yield fake subjects. The result showed that the fake data can also be used to learn meaningful representation compared to those learned from real data.

CONCLUSIONS

To our best knowledge, this work is among the earliest attempts of applying generative deep learning for modeling fMRI data. The proposed RWGAN offers a novel methodology for learning brain representation from fMRI, and it can generate high-quality fake data for the potential use of fMRI data augmentation.

SVR-based Multimodal Active Subspace Analysis for the Brain using Neuroimaging Data.. [DOI: 10.1101/2022.07.28.501879] [Citation(s) in RCA: 0] [Impact Index Per Article: 0]

Understanding the patterns of changes in brain function and structure due to various disorders and diseases is of utmost importance. There have been numerous efforts toward successful biomarker discovery for complex brain disorders by evaluating neuroimaging datasets with novel analytical frameworks. However, due to the multi-faceted nature of the disorders involving a wide and overlapping range of symptoms as well as complex changes in structural and functional brain networks, it is increasingly important to devise computational frameworks that can consider the underlying patterns of heterogeneous changes with specific target assessments, at the same time producing a summarizing output from the high-dimensional neuroimaging data. While various machine learning approaches focus on diagnostic prediction, many learning frameworks analyze important features at the level of brain regions involved in prediction using supervised methods. Unsupervised learning methods have also been utilized to break down the neuroimaging features into lower dimensional components. However, most learning frameworks either do not consider the target assessment information while extracting brain subspaces, or can extract only higher dimensional importance associations as an ordered list of involved features, making manual interpretation at the level of subspaces difficult. We present a novel multimodal active subspace learning framework to understand various subspaces within the brain that are associated with changes in particular biological and cognitive traits. For a given cognitive or biological trait, our framework performs a decomposition of the feature importances to extract robust multimodal subspaces that define the most significant change in the given trait. Through a rigorous cross-validation procedure on an Alzheimer’s disease (AD) dataset, we show that our framework can extract subspaces covering both functional and structural modalities, which are specific to a given clinical assessment (like memory and other cognitive skills) and also retain predictive performance in standard machine learning algorithms. We show that our framework not only uncovers AD-related brain regions (e.g., hippocampus, entorhinal cortex) in the associated brain subspaces, but also enables an automated identification of multiple underlying structural and functional sub-systems of the brain that collectively characterize changes in memory and cognitive skill proficiency related to brain disorders like AD.

Aberrant brain connectivity is associated with childhood maltreatment in individuals with major depressive disorder

Although childhood maltreatment confers a high risk for the development of major depressive disorder, the neurobiological mechanisms underlying this connection remain unknown. The present study sought to identify the specific resting-state networks associated with childhood maltreatment. We recruited major depressive disorder patients with and without a history of childhood maltreatment (n = 31 and n = 30, respectively) and healthy subjects (n = 80). We used independent component analysis to compute inter- and intra- network connectivity. We found that individuals with major depressive disorder and childhood maltreatment could be characterized by the following network disconnectivity model relative to healthy subjects: (i) decreased intra-network connectivity in the left frontoparietal network and increased intra-network connectivity in the right frontoparietal network, (ii) decreased inter-network connectivity in the posterior default mode network-auditory network, posterior default mode network-limbic system, posterior default mode network-anterior default mode network, auditory network-medial visual network, lateral visual network - medial visual network, medial visual network-sensorimotor network, medial visual network - anterior default mode network, occipital pole visual network-dorsal attention network, and posterior default mode network-anterior default mode network, and (iii) increased inter-network connectivity in the sensorimotor network-ventral attention network, and dorsal attention network-ventral attention network. Moreover, we found significant correlations between the severity of childhood maltreatment and the intra-network connectivity of the frontoparietal network. Our study demonstrated that childhood maltreatment is integrally associated with aberrant network architecture in patients with major depressive disorder.

Transient brain-wide coactivations and structured transitions revealed in hemodynamic imaging data

Brain-wide patterns in resting human brains, as either structured functional connectivity (FC) or recurring brain states, have been widely studied in the neuroimaging literature. In particular, resting-state FCs estimated over windowed timeframe neuroimaging data from sub-minutes to minutes using correlation or blind source separation techniques have reported many brain-wide patterns of significant behavioral and disease correlates. The present pilot study utilized a novel whole-head cap-based high-density diffuse optical tomography (DOT) technology, together with data-driven analysis methods, to investigate recurring transient brain-wide patterns in spontaneous fluctuations of hemodynamic signals at the resolution of single timeframes from thirteen healthy adults in resting conditions. Our results report that a small number, i.e., six, of brain-wide coactivation patterns (CAPs) describe major spatiotemporal dynamics of spontaneous hemodynamic signals recorded by DOT. These CAPs represent recurring brain states, showing spatial topographies of hemispheric symmetry, and exhibit highly anticorrelated pairs. Moreover, a structured transition pattern among the six brain states is identified, where two CAPs with anterior-posterior spatial patterns are significantly involved in transitions among all brain states. Our results further elucidate two brain states of global positive and negative patterns, indicating transient neuronal coactivations and co-deactivations, respectively, over the entire cortex. We demonstrate that these two brain states are responsible for the generation of a subset of peaks and troughs in global signals (GS), supporting the recent reports on neuronal relevance of hemodynamic GS. Collectively, our results suggest that transient neuronal events (i.e., CAPs), global brain activity, and brain-wide structured transitions co-exist in humans and these phenomena are closely related, which extend the observations of similar neuronal events recently reported in animal hemodynamic data. Future studies on the quantitative relationship among these transient events and their relationships to windowed FCs along with larger sample size are needed to understand their changes with behaviors and diseased conditions.

Alterations in Functional and Structural Connectivity of Basal Ganglia Network in Patients with Obesity

Obesity is related to overconsumption of high-calorie (HiCal) food, which is modulated by brain reward and inhibitory control circuitries. The basal ganglia (BG) are a key set of nuclei within the reward circuitry, but obesity-associated functional and structural abnormalities of BG have not been well studied. Resting-state functional MRI with independent component analysis (ICA) and probabilistic tractography were employed to investigate differences in BG-related functional-(FC) and structural connectivity (SC) between 32 patients with obesity (OB) and 35 normal-weight (NW) participants. Compared to NW, OB showed significantly lower FC strength in the caudate nucleus within the BG network, and seed-based FC analysis showed lower FC between caudate and dorsolateral prefrontal cortex (DLPFC), which was negatively correlated with craving for HiCal food cues. Further SC analysis revealed that OB showed lower SC than NW between left caudate and left DLPFC as measured with fractional anisotropy (FA). Alterations in FC and SC between caudate and DLPFC in obese patients, which highlights the role of BG network in modulating the balance between reward and inhibitory-control.

Longitudinal manganese-enhanced magnetic resonance imaging of neural projections and activity

Manganese-enhanced magnetic resonance imaging (MEMRI) holds exceptional promise for preclinical studies of brain-wide physiology in awake-behaving animals. The objectives of this review are to update the current information regarding MEMRI and to inform new investigators as to its potential. Mn(II) is a powerful contrast agent for two main reasons: (1) high signal intensity at low doses; and (2) biological interactions, such as projection tracing and neural activity mapping via entry into electrically active neurons in the living brain. High-spin Mn(II) reduces the relaxation time of water protons: at Mn(II) concentrations typically encountered in MEMRI, robust hyperintensity is obtained without adverse effects. By selectively entering neurons through voltage-gated calcium channels, Mn(II) highlights active neurons. Safe doses may be repeated over weeks to allow for longitudinal imaging of brain-wide dynamics in the same individual across time. When delivered by stereotactic intracerebral injection, Mn(II) enters active neurons at the injection site and then travels inside axons for long distances, tracing neuronal projection anatomy. Rates of axonal transport within the brain were measured for the first time in "time-lapse" MEMRI. When delivered systemically, Mn(II) enters active neurons throughout the brain via voltage-sensitive calcium channels and clears slowly. Thus behavior can be monitored during Mn(II) uptake and hyperintense signals due to Mn(II) uptake captured retrospectively, allowing pairing of behavior with neural activity maps for the first time. Here we review critical information gained from MEMRI projection mapping about human neuropsychological disorders. We then discuss results from neural activity mapping from systemic Mn(II) imaged longitudinally that have illuminated development of the tonotopic map in the inferior colliculus as well as brain-wide responses to acute threat and how it evolves over time. MEMRI posed specific challenges for image data analysis that have recently been transcended. We predict a bright future for longitudinal MEMRI in pursuit of solutions to the brain-behavior mystery.