Functional Connectivity Matrices Research Articles

Autism spectrum disorder identification with multi-site functional magnetic resonance imaging

Autism spectrum disorder (ASD) is a neurodevelopmental condition characterized by enduring difficulties in social interaction and communication. People analyzed with ASD may display repetitive behaviors and limited interests. Autism is classified as a spectrum disorder, implying that the symptom intensity might range from mild to severe depending on the individual. To detect ASD in this paper an attribute feature graph approach is designed by using the stastical dependencies features that necessarily accomplish the diagnosis of ASD. In the first phase the features extracted are designed based on the functional magnetic resonance imaging (fMRI) data, in the next-step the attribute feature graph layer learns the features of the node information of various nodes by ASD classification. Further, in the third step, it is employed to independently extract distinguishing features from the functional connectivity matrices of the brain that are derived from fMRI. The custom convolutional neural network (CNN) used in this study is trained on a comprehensive dataset comprising individuals diagnosed with ASD and typically developing individuals. In the fourth stage, a prototype learning is developed to augment the classification performance of the custom-CNN. The experimental analysis further carried out states that the proposed model works efficiently in comparison with the existing system.

Abnormal dynamic functional connectivity in young nondisabling intracerebral hemorrhage patients.

Previous resting-state functional magnetic resonance imaging studies on intracerebral hemorrhage patients have focused more on the static characteristics of brain activity, while the time-varying effects during scanning have received less attention. Therefore, the current study aimed to explore the dynamic functional network connectivity changes of intracerebral hemorrhage patients. Using independent component analysis, the sliding window approach, and the k-means clustering analysis method, different dynamic functional network connectivity states were detected from resting-state functional magnetic resonance imaging data of 37 intracerebral hemorrhage patients and 44 healthy controls. The inter-group differences in dynamic functional network connectivity patterns and temporal properties were investigated, followed by correlation analyses between clinical scales and abnormal functional indexes. Ten resting-state networks were identified, and the dynamic functional network connectivity matrices were clustered into four different states. The transition numbers were decreased in the intracerebral hemorrhage patients compared with healthy controls, which was associated with trail making test scores in patients. The cerebellar network and executive control network connectivity in State 1 was reduced in patients, and this abnormal dynamic functional connectivity was positively correlated with the animal fluency test scores of patients. The current study demonstrated the characteristics of dynamic functional network connectivity in intracerebral hemorrhage patients and revealed that abnormal temporal properties and functional connectivity may be related to the performance of different cognitive domains after ictus. These results may provide new insights into exploring the neurocognitive mechanisms of intracerebral hemorrhage.

Resting-state functional connectivity in children cooled for neonatal encephalopathy.

Therapeutic hypothermia improves outcomes following neonatal hypoxic-ischaemic encephalopathy, reducing cases of death and severe disability such as cerebral palsy compared with normothermia management. However, when cooled children reach early school-age, they have cognitive and motor impairments which are associated with underlying alterations to brain structure and white matter connectivity. It is unknown whether these differences in structural connectivity are associated with differences in functional connectivity between cooled children and healthy controls. Resting-state functional MRI has been used to characterize static and dynamic functional connectivity in children, both with typical development and those with neurodevelopmental disorders. Previous studies of resting-state brain networks in children with hypoxic-ischaemic encephalopathy have focussed on the neonatal period. In this study, we used resting-state fMRI to investigate static and dynamic functional connectivity in children aged 6-8 years who were cooled for neonatal hypoxic-ischaemic without cerebral palsy [n = 22, median age (interquartile range) 7.08 (6.85-7.52) years] and healthy controls matched for age, sex and socioeconomic status [n = 20, median age (interquartile range) 6.75 (6.48-7.25) years]. Using group independent component analysis, we identified 31 intrinsic functional connectivity networks consistent with those previously reported in children and adults. We found no case-control differences in the spatial maps of these intrinsic connectivity networks. We constructed subject-specific static functional connectivity networks by measuring pairwise Pearson correlations between component time courses and found no case-control differences in functional connectivity after false discovery rate correction. To study the time-varying organization of resting-state networks, we used sliding window correlations and deep clustering to investigate dynamic functional connectivity characteristics. We found k = 4 repetitively occurring functional connectivity states, which exhibited no case-control differences in dwell time, fractional occupancy or state functional connectivity matrices. In this small cohort, the spatiotemporal characteristics of resting-state brain networks in cooled children without severe disability were too subtle to be differentiated from healthy controls at early school-age, despite underlying differences in brain structure and white matter connectivity, possibly reflecting a level of recovery of healthy resting-state brain function. To our knowledge, this is the first study to investigate resting-state functional connectivity in children with hypoxic-ischaemic encephalopathy beyond the neonatal period and the first to investigate dynamic functional connectivity in any children with hypoxic-ischaemic encephalopathy.

FNIRS based study of brain network characteristics in children with cerebral palsy during bilateral lower limb movement.

Motor dysfunctions in children with cerebral palsy (CP) are caused by nonprogressive brain damage. Understanding the functional characteristics of the brain is important for rehabilitation. This paper aimed to study the brain networks of children with CP during bilateral lower limb movement using functional near-infrared spectroscopy (fNIRS) and to explore effective fNIRS indices for reflecting functional brain activity. Using fNIRS, cerebral oxygenation signals in the bilateral prefrontal cortex (LPFC/RPFC) and motor cortex (LMC/RMC) were recorded from fifteen children with spastic CP and seventeen children with typical development (CTDs) in the resting state and during bilateral lower limb movement. Functional connectivity matrices based on phase-locking values (PLVs) were calculated using Hilbert transformation, and binary networks were constructed at different sparsity levels. Network metrics such as the clustering coefficient, global efficiency, local efficiency, and transitivity were calculated. Furthermore, the time-varying curves of network metrics during movement were obtained by dividing the time window and using sparse inverse covariance matrices. Finally, conditional Granger causality (GC) was used to explore the causal relationships between different brain regions. Compared to CTDs, the connectivity between RMC-RPFC (p=0.017) and RMC-LMC (p=0.002) in the brain network was decreased in children with CP, and the clustering coefficient (p=0.003), global efficiency (p=0.034), local efficiency (p=0.015), and transitivity (p=0.009) were significantly lower. The standard deviation of the changes in global efficiency of children with CP during motion was also greater than that of CTDs. Using GC, it was found that there was a significant increase in causal strength from the RMC to the RPFC (p=0.04) and from the RMC to the LMC (p=0.042) in children with CP during motion. Additionally, there were significant negative correlations between the PLV of LMC-RMC (p=0.002) and the Gross Motor Function Classification System (GMFCS) and between the GMFCS and the clustering coefficient (p=0.01). During rehabilitation training of the lower limbs, there were significant differences in brain network indices between children with CP and CTDs. The indicators proposed in this paper are effective at evaluating motor function and the real-time impact of rehabilitation training on the brain network and have great potential for application in guiding clinical motor function assessment and planning rehabilitation strategies.

Changes in Clinical and Network Functional Connectivity Parameters in Motor Networks and Cerebellum Based on Resting-State Functional Magnetic Resonance Imaging Data in Patients with Post-Stroke Hemiparesis Receiving Interactive Brain Stimulation Neurotherapy

Introduction. Interactive brain stimulation (IBS) neurotherapy is an advanced neurofeedback technology (NFB) that involves the organization of a feedback “target” based on signals recorded by functional magnetic resonance imaging (fMRI) and electroencephalography (EEG). The NFB allows patients to volitionally self-regulate their current brain activity and may therefore be a useful treatment option for diseases with altered activation and functional connectivity (FC) patterns.
 Our objective was to assess the effects of IBS on the FC changes in motor networks and correlations between clinical and network parameters in patients with post-stroke hand paresis.
 Materials and methods. Patients with a history of stroke 2 months were randomized into a main group (n = 7) and a control group (n = 7). All the patients followed the stroke physical rehabilitation for 3 weeks. The main group received IBS training, where the patients learned to imagine movements of the paretic hand trying to amplify the fMRI signal from the primary motor cortex (M1) and the supplementary motor area (SMA) on the lesion side with simultaneous desynchronizing the μ- and β-2 EEG rhythms in the central leads. Clinical tests and MRI were performed prior to and immediately after the treatment. FC matrices were constructed using CONN software based on resting-state fMRI data.
 Results. By the end of the training, M1–M1 functional connectivity in the control group weakened, while no changes were observed in the main group. The FC strength was positively correlated with the grip strength (ρ = 0.69; p 0.01) and with the results of the Box and Blocks test (BBT score, ρ = 0.72; p 0.01) and the Fugl-Meyer assessment for upper extremity (FM-UE score, ρ = 0.87; p 0.005). Ipsilesional SMA connectivity with contralesional cerebellum weakened (p 0.05 in the main group). Its strength was negatively correlated with the BBT and FM-UE scores (both tests ρ = –0.44; p 0.05).
 Conclusions. Volitional control of M1 and SMA activity in the lesion hemisphere during the post-stroke IBS training alters the architecture of the entire motor network affecting clinically significant FC types. We studied a possible mechanism of this technology and its potential use in treatment programs.

Use of 3D-CAPSNET and RNN models for 4D fMRI-based Alzheimer’s Disease Pre-detection

Predicting Alzheimer's disease (AD) at an early stage can assist more successfully prevent cognitive decline. Numerous investigations have focused on utilizing various convolutional neural network (CNN)-based techniques for automated diagnosis of AD through resting-state functional magnetic resonance imaging (rs-fMRI). Two main constraints face the methodologies presented in these studies. First, overfitting occurs due to the small size of fMRI datasets. Second, an effective modeling of the 4D information from fMRI sessions is required. In order to represent the 4D information, some studies used the deep learning techniques on functional connectivity matrices created from fMRI data, or on fMRI data as distinct 2D slices or 3D volumes. However, this results in information loss in both types of methods. In order to model the spatiotemporal (4D) information of fMRI data for AD diagnosis, a new model based on the capsule network (CapsNet) and recurrent neural network (RNN) is proposed in this study. To assess the suggested model's effectiveness, experiments were run. The findings show that the suggested model could classify AD against normal control (NC) and late mild cognitive impairment (lMCI) against early mild cognitive impairment (eMCI) with accuracy rates of 94.5% and 61.8%, respectively.

Pre-acquired Functional Connectivity Predicts Choice Inconsistency.

Economic choice theories usually assume that humans maximize utility in their choices. However, studies have shown that humans make inconsistent choices, leading to suboptimal behavior, even without context-dependent manipulations. Previous studies showed that activation in value and motor networks are associated with inconsistent choices at the moment of choice. Here, we investigated if the neural predispositions, measured before a choice task, can predict choice inconsistency in a later risky choice task. Using functional connectivity (FC) measures from resting-state functional magnetic resonance imaging (rsfMRI), derived before any choice was made, we aimed to predict subjects' inconsistency levels in a later-performed choice task. We hypothesized that rsfMRI FC measures extracted from value and motor brain areas would predict inconsistency. Forty subjects (21 females) completed a rsfMRI scan before performing a risky choice task. We compared models that were trained on FC that included only hypothesized value and motor regions with models trained on whole-brain FC. We found that both model types significantly predicted inconsistency levels. Moreover, even the whole-brain models relied mostly on FC between value and motor areas. For external validation, we used a neural network pretrained on FC matrices of 37,000 subjects and fine-tuned it on our data and again showed significant predictions. Together, this shows that the tendency for choice inconsistency is predicted by predispositions of the nervous system and that synchrony between the motor and value networks plays a crucial role in this tendency.

Effect of blood oxygen-level-dependent signal data trimming on functional connectivity metrics

In the big data era, with a lack of comparable functional neuroimaging data, researchers try to combine heterogeneous data of different lengths, trimming those to the same number of timepoints (NTPs). However, the effects of trimming blood-oxygen-level dependent (BOLD) signal data on functional connectivity (FC) are still poorly understood.Resting-state functional magnetic resonance imaging data from thirty healthy subjects were pre-processed for five different NTPs, from which FC matrices were computed. These BOLD signal correlation matrices were binarized for several thresholds, excluding weak correlations. Graph metrics were computed to study FC differences between different NTPs. The study included node degree analysis for each brain region and assessment of small-worldness coefficients (σ and ω), whereas in small-world networks, characteristic values are σ > 1 and ω≈ 0, indicating a balance between high clustering coefficients and short characteristic path lengths.A tendency of decreasing global network degrees for higher NTPs was observed, translating the loss of stronger correlations with longer BOLD signals. Trimming such data affects brain regions differently, probably due to brain network dynamics. Regarding small-worldness, we observed that σ was greater than 1 for all the different NTPs, showing an increasing trend for higher NTPs (median value: σBRAIN= 3.05). In addition, ω consistently remained greater than 0 for all NTPs, gradually approaching 0 as the NTPs increased (median value ωBRAIN= 0.20). As such, the results suggest a tendency for an increase of small-world properties for increasing NTPs. Nonetheless, the overall properties of brain networks almost remain constant. In conclusion, trimming BOLD signal data leads to small differences in FC.

Coupling Between Human Brain Cortical Thickness and Glucose Metabolism from Regional to Connective Level: A Positron Emission Tomography/Magnetic Resonance Imaging Study.

Background: Balance between brain structure and function is implicated in aging and many brain disorders. This study aimed to investigate the coupling between brain structure and function using 18F-fludeoxyglucose positron emission tomography (PET)/magnetic resonance imaging (MRI). Methods: One hundred thirty-eight subjects who underwent brain 18F-FDG PET/MRI were recruited. The structural and functional coupling at the regional level was explored by calculating within-subject Spearman's correlation between glucose metabolism (GluM) and cortical thickness (CTh) across the cortex for each subject, which was then correlated with age to explore its physiological effects. Then, subjects were divided into groups of middle-aged and young adults and older adults (OAs); structural connectivity (SC) based on CTh and functional connectivity (FC) based on GluM were constructed for the two groups, respectively, followed by exploring the connective-level structural and functional coupling on SC and FC matrices. The global and local efficiency values of the brain SC and FC were also evaluated. Results: Of the subjects, 97.83% exhibited a significant negative correlation between regional CTh and GluM (r = -0.24 to -0.71, p < 0.05, FDR correction), and this CTh-GluM correlation was negatively correlated with age (R = -0.35, p < 0.001). For connectivity matrices, many regions showed positive correlation between SC and FC, especially in the OA group. Besides, FC exhibited denser connections than SC, resulting in both higher global and local efficiency, but lower global efficiency when the network size was corrected. Conclusions: This study found couplings between CTh and GluM at both regional and connective levels, which reflected the aging progress, and might provide new insight into brain disorders. Impact statement The intricate interplay between brain structures and functions plays a pivotal role in unraveling the complexities inherent in the aging process and the pathogenesis of neurological disorders. This study revealed that 97.83% subjects showed negative correlation between the brain's regional cortical thickness and glucose metabolism, while at the connective level, many regions showed positive correlations between structural and functional connectivity. The observed coupling at the regional and connective levels reflected physiological progress, such as aging, and provides insights into the brain mechanisms and potential implications for the diagnosis and treatment of brain disorders.

Residual graph transformer for autism spectrum disorder prediction

Brain functional connectivity (FC) based on resting-state functional magnetic resonance imaging (rs-fMRI) has been in vogue to predict Autism Spectrum Disorder (ASD), which is a neuropsychiatric disease up the plight of locating latent biomarkers for clinical diagnosis. Albeit massive endeavors have been made, most studies are fed up with several chronic issues, such as the intractability of harnessing the interaction flourishing within brain regions, the astriction of representation due to vanishing gradient within deeper network architecture, and the poor interpretability leading to unpersuasive diagnosis. To ameliorate these issues, a FC-learned Residual Graph Transformer Network, namely RGTNet, is proposed. Specifically, we design a Graph Encoder to extract temporal-related features with long-range dependencies, from which interpretable FC matrices would be modeled. Besides, the residual trick is introduced to deepen the GCN architecture, thereby learning the higher-level information. Moreover, a novel Graph Sparse Fitting followed by weighted aggregation is proposed to ease dimensionality explosion. Empirically, the results on two types of ABIDE data sets demonstrate the meliority of RGTNet. Notably, the achieved ACC metric reaches 73.4%, overwhelming most competitors with merely 70.9% on the AAL atlas using a five-fold cross-validation policy. Moreover, the investigated biomarkers concord closely with the authoritative medical knowledge, paving a viable way for ASD-clinical diagnosis. Our code is available at https://github.com/CodeGoat24/RGTNet.

Sex based differences in functional connectivity during a working memory task: an fNIRS study.

Differences in corticocerebral structure and function between males and females and their effects on behavior and the prevalence of various neuropsychiatric disorders have been considered as a fundamental topic in various fields of neuroscience. Recent studies on working memory (WM) reported the impact of sex on brain connectivity patterns, which reflect the important role of functional connectivity in the sex topic. Working memory, one of the most important cognitive tasks performed by regions of the PFC, can provide evidence regarding the presence of a difference between males and females. The present study aimed to assess sex differences in brain functional connectivity during working memory-related tasks by using functional near-infrared spectroscopy (fNIRS). In this regard, nine males and nine females completed a dual n-back working memory task with two target inputs of color and location stimuli in three difficulty levels (n = 0, 1, 2). Functional connectivity matrices were extracted for each subject for each memory load level. Females made less errors than males while spending more time performing the task for all workload levels except in 0-back related to the color stimulus, where the reaction time of females was shorter than males. The results of functional connectivity reveal the inverse behavior of two hemispheres at different memory workload levels between males and females. In the left hemisphere, males exhibited stronger connectivity compared to the females, while stronger connectivity was observed in the females' right hemisphere. Furthermore, an inverse trend was detected in the channel pairs with significant connectivity in the right hemisphere of males (falling) and females (rising) by enhancing working memory load level. Considering both behavioral and functional results for two sexes demonstrated a better performance in females due to the more effective use of the brain. The results indicate that sex affects functional connectivity between different areas in both hemispheres of the brain during cognitive tasks of varying difficulty levels although the general impression is that spatial capabilities are considered as a performance of the brain's right hemisphere. These results reinforce the presence of a sex effect in the functional imaging studies of hemodynamic function and emphasize the importance of evaluating brain network connectivity for achieving a better scientific understanding of sex differences.

A Method to Estimate Longitudinal Change Patterns in Functional Network Connectivity of the Developing Brain Relevant to Psychiatric Problems, Cognition, and Age.

Aim: To develop an approach to evaluate multiple overlapping brain functional change patterns (FCPs) in functional network connectivity (FNC) and apply to study developmental changes in brain function. Introduction: FNC, the network analog of functional connectivity (FC), is commonly used to capture the intrinsic functional relationships among brain networks. Ongoing research on longitudinal changes of intrinsic FC across whole-brain functional networks has proven useful for characterizing age-related changes, but to date, there has been little focus on capturing multivariate patterns of FNC change with brain development. Methods: In this article, we introduce a novel approach to evaluate multiple overlapping FCPs by utilizing FNC matrices. We computed FNC matrices from the large-scale Adolescent Brain Cognitive Development data using fully automated spatially constrained independent component analysis (ICA). We next evaluated changes in these patterns for a 2-year period using a second-level ICA on the FNC change maps. Results: Our proposed approach reveals several highly structured (modular) FCPs and significant results including strong brain FC between visual and sensorimotor domains that increase with age. We also find several FCPs that are associated with longitudinal changes of psychiatric problems, cognition, and age in the developing brain. Interestingly, FCP cross-covariation, reflecting coupling between maximally independent FCPs, also shows significant differences between upper and lower quartile loadings for longitudinal changes in age, psychiatric problems, and cognition scores, as well as baseline age in the developing brain. FCP patterns and results were also found to be highly reliable based on analysis of data collected in a separate scan session. Conclusion: In sum, our results show evidence of consistent multivariate patterns of functional change in emerging adolescents and the proposed approach provides a useful and general tool to evaluate covarying patterns of whole-brain functional changes in longitudinal data. Impact statement In this article, we introduce a novel approach utilizing functional network connectivity (FNC) matrices to estimate multiple overlapping brain functional change patterns (FCPs). The findings demonstrate several well-structured FCPs that exhibit significant changes for a 2-year period, particularly in the functional connectivity between the visual and sensorimotor domains. In addition, we discover several FCPs that are associated with psychopathology, cognition, and age. Finally, our proposed approach for studying age-related FCPs represents a pioneering method that provides a valuable tool for assessing interconnected patterns of whole-brain functional changes in longitudinal data and may be useful to study change over time with applicability to many other areas, including the study of longitudinal changes within diagnostic groups, treatment effects, aging effects, and more.

Intra-Atlas Node Size Effects on Graph Metrics in fMRI Data: Implications for Alzheimer's Disease and Cognitive Impairment.

Network neuroscience, a multidisciplinary field merging insights from neuroscience and network theory, offers a profound understanding of neural network intricacies. However, the impact of varying node sizes on computed graph metrics in neuroimaging data remains underexplored. This study addresses this gap by adopting a data-driven methodology to delineate functional nodes and assess their influence on graph metrics. Using the Neuromark framework, automated independent component analysis is applied to resting state fMRI data, capturing functional network connectivity (FNC) matrices. Global and local graph metrics reveal intricate connectivity patterns, emphasizing the need for nuanced analysis. Notably, node sizes, computed based on voxel counts, contribute to a novel metric termed 'node-metric coupling' (NMC). Correlations between graph metrics and node dimensions are consistently observed. The study extends its analysis to a dataset comprising Alzheimer's disease, mild cognitive impairment, and control subjects, showcasing the potential of NMC as a biomarker for brain disorders. The two key outcomes underscore the interplay between node sizes and resultant graph metrics within a given atlas, shedding light on an often-overlooked source of variability. Additionally, the study highlights the utility of NMC as a valuable biomarker, emphasizing the necessity of accounting for node sizes in future neuroimaging investigations. This work contributes to refining comparative studies employing diverse atlases and advocates for thoughtful consideration of intra-atlas node size in shaping graph metrics, paving the way for more robust neuroimaging research.

Static and dynamic fMRI-derived functional connectomes represent largely similar information.

Functional connectivity (FC) of blood oxygen level-dependent (BOLD) fMRI time series can be estimated using methods that differ in sensitivity to the temporal order of time points (static vs. dynamic) and the number of regions considered in estimating a single edge (bivariate vs. multivariate). Previous research suggests that dynamic FC explains variability in FC fluctuations and behavior beyond static FC. Our aim was to systematically compare methods on both dimensions. We compared five FC methods: Pearson's/full correlation (static, bivariate), lagged correlation (dynamic, bivariate), partial correlation (static, multivariate), and multivariate AR model with and without self-connections (dynamic, multivariate). We compared these methods by (i) assessing similarities between FC matrices, (ii) by comparing node centrality measures, and (iii) by comparing the patterns of brain-behavior associations. Although FC estimates did not differ as a function of sensitivity to temporal order, we observed differences between the multivariate and bivariate FC methods. The dynamic FC estimates were highly correlated with the static FC estimates, especially when comparing group-level FC matrices. Similarly, there were high correlations between the patterns of brain-behavior associations obtained using the dynamic and static FC methods. We conclude that the dynamic FC estimates represent information largely similar to that of the static FC.

Fusing multi-scale fMRI features using a brain-inspired multi-channel graph neural network for major depressive disorder diagnosis

Depression stands as one of the most pernicious mental disorders in contemporary society, characterized by a highly intricate pathological mechanism. Specifically, individuals suffering from Major Depressive Disorder (MDD) exhibit heightened vulnerability to suicidal tendencies. Currently, healthcare practitioners often encounter challenges related to the misdiagnosis and underdiagnosis of depression during clinical assessments. Consequently, it is of paramount importance to develop highly accurate auxiliary diagnostic tools for depression. Unfortunately, traditional machine learning and deep learning methodologies frequently neglect the integration of multi-source data and disregard the intricate topological structure and high-order attributes of brain networks. In this study, a multi-scale feature fusion classification framework is proposed to distinguish between MDD patients and healthy controls. Within the proposed model, a novel method, the Cross-Level High-Order Interaction (CLHOI), is introduced and implemented on a low-order functional connectivity (LOFC) matrix to derive two distinct high-order functional connectivity (HOFC) matrices. Subsequently, a Multi-Channel Fusion Graph Convolutional Network (MFGCN) is trained by integrating high-order and low-order brain graph features along with phenotypic information. The results of 10-fold cross-validation experiments conducted on the publicly available REST-meta-MDD dataset indicate that the fusion of multi-scale features improves the average accuracy by approximately 3%, resulting in an accuracy rate of 77.6%. Simultaneously, the hypothesis asserting the existence of intricate information interactions at various levels within brain connectivity networks is validated. Moreover, our model exhibits strong explanatory capabilities, effectively identifying brain regions closely associated with MDD, including the Precentral gyrus, Superior frontal gyrus, Cuneus, Lingual gyrus, and Fusiform gyrus. In comparison to numerous advanced studies within the same domain, our approach has produced competitive results. Furthermore, our proposed method can be readily extended to facilitate the diagnosis of various neurological diseases.

Discriminative analysis of schizophrenia patients using an integrated model combining 3D CNN with 2D CNN: A multimodal MR image and connectomics analysis

ObjectiveFew studies have applied deep learning to the discriminative analysis of schizophrenia (SZ) patients using the fusional features of multimodal MRI data. Here, we proposed an integrated model combining a 3D convolutional neural network (CNN) with a 2D CNN to classify SZ patients. MethodStructural MRI (sMRI) and resting-state functional MRI (rs-fMRI) data were acquired for 140 SZ patients and 205 normal controls. We computed structural connectivity (SC) from the sMRI data as well as functional connectivity (FC), amplitude of low-frequency fluctuation (ALFF), and regional homogeneity (ReHo) from the rs-fMRI data. The 3D images of T1, ReHo, and ALFF were used as the inputs for the 3D CNN model, while the SC and FC matrices were used as the inputs for the 2D CNN model. Moreover, we added squeeze and excitation blocks (SE-blocks) to each layer of the integrated model and used a support vector machine (SVM) to replace the softmax classifier. ResultsThe integrated model proposed in this study, using the fusional features of the T1 images, and the matrices of FC, showed the best performance. The use of the SE-blocks and SVM classifiers significantly improved the performance of the integrated model, in which the accuracy, sensitivity, specificity, area under the curve, and F1-score were 89.86%, 86.21%, 92.50%, 89.35%, and 87.72%, respectively. ConclusionsOur findings indicated that an integrated model combining 3D CNN with 2D CNN is a promising method to improve the classification performance of SZ patients and has potential for the clinical diagnosis of psychiatric diseases.

Cross‐scale functional connectivity patterns of the aging brain learned from the multi‐cohort iSTAGing study

AbstractBackgroundBrain functional connectivity (FC) measures derived from resting‐state fMRI (rsfMRI) data have advanced our understanding of the brain organization. However, most existing studies investigate the brain functional networks (FNs) at a specific scale, ignoring potentially complementary FC information at different levels of spatial granularity as well as the cross‐scale information. Given that the brain is a multi‐scale, hierarchical system, FCs between FNs at different spatial scales may provide complementary information to those measured within‐scales for characterizing the aging brain.MethodWe computed multiscale FNs to capture their FC measures across multiple scales and investigated the associations between cross‐scale FCs and the brain aging based on rsfMRI data of a diverse cohort (n = 4186, aged 22 to 97 years) from seven sites of the iSTAGING study [1]. For each subject, we computed the FNs at seven scales, yielding seven sets of k (k = 17, 25, 50, 75, 100, 125 and 150) FNs, same as our previous study [2‐3]. We estimated FC measures between the 542 FNs using Pearson’s correlation, yielding the cross‐scale FC matrices that were harmonized across sites in the tangent space using ComBat‐GAM [4] with sex and age as covariates. The harmonized FC measures were used as features to build an age prediction model using Ridge regression as illustrated in Figure 1. The brain age prediction model was compared with that built on harmonized FC measures between FNs of the same scales, with the performance estimated with a nested five‐fold cross‐validation.ResultThe cross‐scale FC measures performed significantly better than the within‐scale FC measures for the brain age prediction, as summarized in Table 1. Figure 2 shows the most informative FC measures for the brain age prediction. Specifically, FC measures computed both cross‐ and within‐scales between the visual, somatomotor, and frontoparietal networks provided relatively more useful information for the brain age prediction than FC measures between other FNs.ConclusionOur study demonstrated that the cross‐scale FC measures provided complementary information to the within‐scale FC measures for predicting the brain age, facilitating enhanced understanding of aging related changes of the brain functional organization.

Anterior and posterior default mode network functional connectivity and segregation in aging

AbstractBackgroundThe default mode network (DMN) supports memory performance and has been shown to be vulnerable to the effects of aging. The posterior DMN (pDMN) is an early region of amyloid‐β (Aβ) accumulation, which may lead to disrupted communication within/between subnetworks of the DMN. Using resting state functional connectivity (FC), we tested whether communication within and between the pDMN and anterior DMN (aDMN) were altered by aging and Aβ pathology and related to memory performance. We also tested segregation FC between the aDMN and pDMN.MethodData from 27 cognitively normal adults (69.6 years, 59% females) with resting state fMRI (rsfMRI), 18F‐florbetapir (FBP) PET to measure Aβ, and the Rey Auditory Verbal LearningTest (RAVLT) delayed recall to assess memory were analyzed. The rsfMRI data were preprocessed using CONN toolbox. ROI‐to‐ROI FC matrices were constructed for regions included in the aDMN and pDMN. We calculated: (1) the average FC within each network (Zw), (2) average FC between the aDMN and pDMN (Zb), and (3) network segregation [(Zw‐Zb)/Zw] for each DMN network. Global FBP SUVR was calculated as the mean across association cortical regions, and a cutoff of 1.11 was used to determine Aβ positivity. Analyses included the FC variables and their associations with age, Aβ, and memory.ResultOlder age was associated with reduced network segregation of the pDMN (r = ‐0.53, p = 0.004), and a trend level association for reduced within‐pDMN FC (r = ‐0.37, p = 0.057). Global FBP SUVR was not associated with FC or network segregation, and there was no significant difference in FC between Aβ groups (controlling for age). There was a trend‐level association between better memory performance and stronger FC between aDMN and pDMN (r = 0.37, p = 0.066; controlling for age and education).ConclusionResults demonstrated that aging is associated with specific FC disruptions of pDMN, leading to reduced specialization of processing as indexed by network segregation. Moreover, increased communication between DMN subnetworks may support memory in aging. Future analyses in a larger sample will further test the association between Alzheimer’s pathology and functional changes to DMN subnetworks.

Longitudinal robustness of resting‐state functional connectivity signatures across different stages of the Alzheimer’s disease continuum

AbstractBackgroundA growing number of resting‐state fMRI (rs‐fMRI) studies report changes in network activity to be a frequent and often early pathological feature in many neurodegenerative conditions. The majority of studies have primarily focused on changes in group averages, failing to account for individual functional connectivity (FC) idiosyncrasies that have been shown to provide a unique identifier or “fingerprint” in healthy individuals (Finn et al., 2015). It thus remains unclear to what extent FC‐signatures are stable in the context of cognitive decline and Alzheimer’s Disease (AD) pathology (e.g., β‐amyloid [Aβ] and tau). We studied the robustness of subject‐specific FC in a longitudinal cohort covering the AD spectrum.Method275 participants with longitudinal rs‐fMRI data—baseline plus a follow up after approximately two‐years—were selected from the Swedish BioFINDER‐2 cohort, split into six groups: cognitively normal (CN) Aβ‐, subjective cognitive decline (SCD) Aβ‐, SCD Aβ+, mild cognitive decline (MCI) Aβ‐, MCI Aβ+, and AD Aβ+. FC matrices were derived from each subject rs‐fMRI session using the Schaefer cortical parcellation (200 and 400 ROIs), and consequently, a 275×275 identifiability matrix (IM) was derived; see Fig. 1B. For each subject, two fingerprinting measures were extracted from the IM: self identifiability (I‐self) and differential identifiability (I‐diff); see Fig.1B.ResultI‐self and I‐diff show varied distributions across the six groups, but values remain notably high even within the AD Aβ+ (Fig.2A). Both I‐self and I‐diff show significant differences between CN Aβ‐ and AD Aβ+ (Fig.2B). SCD Aβ‐ not only show a greater difference to AD Aβ+, but also to SCD Aβ+ and MCI Aβ+; see Fig.2B. Baseline tau‐PET was associated with I‐self (r = ‐0.1928,p = 0.0014) and to a lesser degree with I‐diff (r = ‐0.1293, p = 0.0334), whereas baseline MMSE was only found associated to I‐self (r = 0.1671,p = 0.0055); neither I‐self nor I‐diff were associated with age (Fig.3).ConclusionOur results suggest that resting‐state FC is mostly stable under cognitive decline and presence of AD pathology. FC was most notably differentiable for SCD Aβ‐, suggesting it as a potential signature to decipher underlying cognitive status and to predict likelihood of future conversion to AD.

Spatial gradients of gene expression, amyloid PET and functional connectome for predicting regional rates of tau‐PET increase: towards tau‐PET outcome ROIs in clinical trials.

AbstractBackgroundIn clinical trials, an important question is which brain regions show high tau accumulation within 1‐2 years. Building upon our previous epicenter‐connectivity‐based distance model for predicting the progression of tau‐PET accumulation from initial seed regions of high tau (epicenters) to other regions (Franzmeier et al. Sci Adv. 2020), we incorporate here epicenter‐based distance within spatial gradients of gene expression and amyloid‐PET to enhance the predictive power.MethodWe included 351 participants encompassing Ab(+) non‐demented subjects (controls &amp; MCI), characterized by cross‐sectional 11‐F florbetaben amyloid‐PET and longitudinal 11‐F flortaucipir tau‐PET (longitudinal, n = 172, mean/range = 1,7 yrs (0.7 – 5.2 yrs)) from ADNI. For transcriptomic gene expression maps obtained from the Allen Human Brain Atlas, we computed for each of the 18,686 genes maps spatial correlations with the average tau‐PET map in a randomly selected training sample (n = 106), and retained the top 5% of positively/inversely tau‐correlating gene maps. For each imaging modality, spatial gradients were subsequently computed across amyloid‐PET scans in the training sample, the subset of gene expression maps, and functional connectivity (FC) matrices from Human Connectome project. In the test sample (n = 245), tau‐PET epicenters at baseline (Franzmeier et al. Sci Adv 2020) were projected onto the gradient maps and epicenter‐based gradient distance was computed. In step‐wise spatial regression analyses, we tested which gradient distances predicted the spatial patterns of cross‐sectional and longitudinal tau‐PET accumulation, accounting for spatial autocorrelations.ResultFor cross‐sectional tau‐PET (Fig. 1A), a total of 54.6% (p &lt; 0.0001) of the spatial tau pattern could be explained, where an amyloid‐PET gradients and a FC gradient (spanning the transmodal axis from primary sensory to DMN regions) showed the strongest effects (Fig. 2, upper row &amp; Fig. 3A). For longitudinal tau‐PET (Fig. 1B), a total of 47.8% (p &lt; 0.0001) of tau increases could be explained, where gene‐expression gradients showed the strongest effects (Fig. 2 lower row &amp; Fig. 3B).ConclusionEpicenter‐based gradient distances of connectivity and molecular brain properties show additive, high predictive accuracy for predicting regions of high future tau‐accumulation, providing an attractive approach to define tau‐PET outcome ROIs in clinical trials.