Search

There is significant interest in using neuroimaging data to predict behavior. The predictive models are often interpreted by the computation of feature importance, which quantifies the predictive relevance of an imaging feature. Tian and Zalesky (2021) suggest that feature importance estimates exhibit low split-half reliability, as well as a trade-off between prediction accuracy and feature importance reliability across parcellation resolutions. However, it is unclear whether the trade-off between prediction accuracy and feature importance reliability is universal. Here, we demonstrate that, with a sufficient sample size, feature importance (operationalized as Haufe-transformed weights) can achieve fair to excellent split-half reliability. With a sample size of 2600 participants, Haufe-transformed weights achieve average intra-class correlation coefficients of 0.75, 0.57 and 0.53 for cognitive, personality and mental health measures respectively. Haufe-transformed weights are much more reliable than original regression weights and univariate FC-behavior correlations. Original regression weights are not reliable even with 2600 participants. Intriguingly, feature importance reliability is strongly positively correlated with prediction accuracy across phenotypes. Within a particular behavioral domain, there is no clear relationship between prediction performance and feature importance reliability across regression models. Furthermore, we show mathematically that feature importance reliability is necessary, but not sufficient, for low feature importance error. In the case of linear models, lower feature importance error is mathematically related to lower prediction error. Therefore, higher feature importance reliability might yield lower feature importance error and higher prediction accuracy. Finally, we discuss how our theoretical results relate with the reliability of imaging features and behavioral measures. Overall, the current study provides empirical and theoretical insights into the relationship between prediction accuracy and feature importance reliability., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2023 The Author(s). Published by Elsevier Inc. All rights reserved.)

Resting-state fMRI is commonly used to derive brain parcellations, which are widely used for dimensionality reduction and interpreting human neuroscience studies. We previously developed a model that integrates local and global approaches for estimating areal-level cortical parcellations. The resulting local-global parcellations are often referred to as the Schaefer parcellations. However, the lack of homotopic correspondence between left and right Schaefer parcels has limited their use for brain lateralization studies. Here, we extend our previous model to derive homotopic areal-level parcellations. Using resting-fMRI and task-fMRI across diverse scanners, acquisition protocols, preprocessing and demographics, we show that the resulting homotopic parcellations are as homogeneous as the Schaefer parcellations, while being more homogeneous than five publicly available parcellations. Furthermore, weaker correlations between homotopic parcels are associated with greater lateralization in resting network organization, as well as lateralization in language and motor task activation. Finally, the homotopic parcellations agree with the boundaries of a number of cortical areas estimated from histology and visuotopic fMRI, while capturing sub-areal (e.g., somatotopic and visuotopic) features. Overall, these results suggest that the homotopic local-global parcellations represent neurobiologically meaningful subdivisions of the human cerebral cortex and will be a useful resource for future studies. Multi-resolution parcellations estimated from 1479 participants are publicly available (https://github.com/ThomasYeoLab/CBIG/tree/master/stable_projects/brain_parcellation/Yan2023_homotopic)., (Copyright © 2023. Published by Elsevier Inc.)

An increasing number of studies have investigated the relationships between inter-individual variability in brain regions' connectivity and behavioral phenotypes, making use of large population neuroimaging datasets. However, the replicability of brain-behavior associations identified by these approaches remains an open question. In this study, we examined the cross-dataset replicability of brain-behavior association patterns for fluid cognition and openness predictions using a previously developed region-wise approach, as well as using a standard whole-brain approach. Overall, we found moderate similarity in patterns for fluid cognition predictions across cohorts, especially in the Human Connectome Project Young Adult, Human Connectome Project Aging, and Enhanced Nathan Kline Institute Rockland Sample cohorts, but low similarity in patterns for openness predictions. In addition, we assessed the generalizability of prediction models in cross-dataset predictions, by training the model in one dataset and testing in another. Making use of the region-wise prediction approach, we showed that first, a moderate extent of generalizability could be achieved with fluid cognition prediction, and that, second, a set of common brain regions related to fluid cognition across cohorts could be identified. Nevertheless, the moderate replicability and generalizability could only be achieved in specific contexts. Thus, we argue that replicability and generalizability in connectivity-based prediction remain limited and deserve greater attention in future studies., Competing Interests: Declaration of Competing Interest There is no conflict of interest., (Copyright © 2022. Published by Elsevier Inc.)

There is significant interest in pooling magnetic resonance image (MRI) data from multiple datasets to enable mega-analysis. Harmonization is typically performed to reduce heterogeneity when pooling MRI data across datasets. Most MRI harmonization algorithms do not explicitly consider downstream application performance during harmonization. However, the choice of downstream application might influence what might be considered as study-specific confounds. Therefore, ignoring downstream applications during harmonization might potentially limit downstream performance. Here we propose a goal-specific harmonization framework that utilizes downstream application performance to regularize the harmonization procedure. Our framework can be integrated with a wide variety of harmonization models based on deep neural networks, such as the recently proposed conditional variational autoencoder (cVAE) harmonization model. Three datasets from three different continents with a total of 2787 participants and 10,085 anatomical T1 scans were used for evaluation. We found that cVAE removed more dataset differences than the widely used ComBat model, but at the expense of removing desirable biological information as measured by downstream prediction of mini mental state examination (MMSE) scores and clinical diagnoses. On the other hand, our goal-specific cVAE (gcVAE) was able to remove as much dataset differences as cVAE, while improving downstream cross-sectional prediction of MMSE scores and clinical diagnoses., (Copyright © 2022. Published by Elsevier Inc.)

A fundamental goal across the neurosciences is the characterization of relationships linking brain anatomy, functioning, and behavior. Although various MRI modalities have been developed to probe these relationships, direct comparisons of their ability to predict behavior have been lacking. Here, we compared the ability of anatomical T1, diffusion and functional MRI (fMRI) to predict behavior at an individual level. Cortical thickness, area and volume were extracted from anatomical T1 images. Diffusion Tensor Imaging (DTI) and approximate Neurite Orientation Dispersion and Density Imaging (NODDI) models were fitted to the diffusion images. The resulting metrics were projected to the Tract-Based Spatial Statistics (TBSS) skeleton. We also ran probabilistic tractography for the diffusion images, from which we extracted the stream count, average stream length, and the average of each DTI and NODDI metric across tracts connecting each pair of brain regions. Functional connectivity (FC) was extracted from both task and resting-state fMRI. Individualized prediction of a wide range of behavioral measures were performed using kernel ridge regression, linear ridge regression and elastic net regression. Consistency of the results were investigated with the Human Connectome Project (HCP) and Adolescent Brain Cognitive Development (ABCD) datasets. In both datasets, FC-based models gave the best prediction performance, regardless of regression model or behavioral measure. This was especially true for the cognitive component. Furthermore, all modalities were able to predict cognition better than other behavioral components. Combining all modalities improved prediction of cognition, but not other behavioral components. Finally, across all behaviors, combining resting and task FC yielded prediction performance similar to combining all modalities. Overall, our study suggests that in the case of healthy children and young adults, behaviorally-relevant information in T1 and diffusion features might reflect a subset of the variance captured by FC., (Copyright © 2022. Published by Elsevier Inc.)

Individual differences in brain anatomy can be used to predict variations in cognitive ability. Most studies to date have focused on broad population-level trends, but the extent to which the observed predictive features are shared across sexes and age groups remains to be established. While it is standard practice to account for intracranial volume (ICV) using proportion correction in both regional and whole-brain morphometric analyses, in the context of brain-behavior predictions the possible differential impact of ICV correction on anatomical features and subgroups within the population has yet to be systematically investigated. In this work, we evaluate the effect of proportional ICV correction on sex-independent and sex-specific predictive models of individual cognitive abilities across multiple anatomical properties (surface area, gray matter volume, and cortical thickness) in healthy young adults (Human Connectome Project; n = 1013, 548 females) and typically developing children (Adolescent Brain Cognitive Development study; n = 1823, 979 females). We demonstrate that ICV correction generally reduces predictive accuracies derived from surface area and gray matter volume, while increasing predictive accuracies based on cortical thickness in both adults and children. Furthermore, the extent to which predictive models generalize across sexes and age groups depends on ICV correction: models based on surface area and gray matter volume are more generalizable without ICV correction, while models based on cortical thickness are more generalizable with ICV correction. Finally, the observed neuroanatomical features predictive of cognitive abilities are unique across age groups regardless of ICV correction, but whether they are shared or unique across sexes (within age groups) depends on ICV correction. These findings highlight the importance of considering individual differences in ICV, and show that proportional ICV correction does not remove the effects of cranial volume from anatomical measurements and can introduce ICV bias where previously there was none. ICV correction choices affect not just the strength of the relationships captured, but also the conclusions drawn regarding the neuroanatomical features that underlie those relationships., (Copyright © 2022. Published by Elsevier Inc.)

What dynamic processes underly functional brain networks? Functional connectivity (FC) and functional connectivity dynamics (FCD) are used to represent the patterns and dynamics of functional brain networks. FC(D) is related to the synchrony of brain activity: when brain areas oscillate in a coordinated manner this yields a high correlation between their signal time series. To explain the processes underlying FC(D) we review how synchronized oscillations emerge from coupled neural populations in brain network models (BNMs). From detailed spiking networks to more abstract population models, there is strong support for the idea that the brain operates near critical instabilities that give rise to multistable or metastable dynamics that in turn lead to the intermittently synchronized slow oscillations underlying FC(D). We explore further consequences from these fundamental mechanisms and how they fit with reality. We conclude by highlighting the need for integrative brain models that connect separate mechanisms across levels of description and spatiotemporal scales and link them with cognitive function., (Copyright © 2022. Published by Elsevier Inc.)

Cognitive abilities and affective experience are key human traits that are interrelated in behavior and brain. Individual variation of cognitive and affective traits, as well as brain structure, has been shown to partly underlie genetic effects. However, to what extent affect and cognition have a shared genetic relationship with local brain structure is incompletely understood. Here we studied phenotypic and genetic correlations of cognitive and affective traits in behavior and brain structure (cortical thickness, surface area and subcortical volumes) in the pedigree-based Human Connectome Project sample (N = 1091). Both cognitive and affective trait scores were highly heritable and showed significant phenotypic correlation on the behavioral level. Cortical thickness in the left superior frontal cortex showed a phenotypic association with both affect and cognition. Decomposing the phenotypic correlations into genetic and environmental components showed that the associations were accounted for by shared genetic effects between the traits. Quantitative functional decoding of the left superior frontal cortex further indicated that this region is associated with cognitive and emotional functioning. This study provides a multi-level approach to study the association between affect and cognition and suggests a convergence of both in superior frontal cortical thickness., Competing Interests: Declaration of Competing Interest The authors declare no competing interests., (Copyright © 2021. Published by Elsevier Inc.)

Null models are useful for assessing whether a dataset exhibits a non-trivial property of interest. These models have recently gained interest in the neuroimaging community as means to explore dynamic properties of functional Magnetic Resonance Imaging (fMRI) time series. Interpretation of null-model testing in this context may not be straightforward because (i) null hypotheses associated to different null models are sometimes unclear and (ii) fMRI metrics might be 'trivial', i.e. preserved under the null hypothesis, and still be useful in neuroimaging applications. In this commentary, we review several commonly used null models of fMRI time series and discuss the interpretation of the corresponding tests. We argue that, while null-model testing allows for a better characterization of the statistical properties of fMRI time series and associated metrics, it should not be considered as a mandatory validation step to assess their relevance in representing brain functional dynamics., (Copyright © 2021 The Author(s). Published by Elsevier Inc. All rights reserved.)

Modern approaches to investigate complex brain dynamics suggest to represent the brain as a functional network of brain regions defined by a brain atlas, while edges represent the structural or functional connectivity among them. This approach is also utilized for mathematical modeling of the resting-state brain dynamics, where the applied brain parcellation plays an essential role in deriving the model network and governing the modeling results. There is however no consensus and empirical evidence on how a given brain atlas affects the model outcome, and the choice of parcellation is still rather arbitrary. Accordingly, we explore the impact of brain parcellation on inter-subject and inter-parcellation variability of model fitting to empirical data. Our objective is to provide a comprehensive empirical evidence of potential influences of parcellation choice on resting-state whole-brain dynamical modeling. We show that brain atlases strongly influence the quality of model validation and propose several variables calculated from empirical data to account for the observed variability. A few classes of such data variables can be distinguished depending on their inter-subject and inter-parcellation explanatory power., Competing Interests: Declaration of Competing Interest The authors have declared that no competing interests exist., (Copyright © 2021. Published by Elsevier Inc.)

Connectomes are typically mapped at low resolution based on a specific brain parcellation atlas. Here, we investigate high-resolution connectomes independent of any atlas, propose new methodologies to facilitate their mapping and demonstrate their utility in predicting behavior and identifying individuals. Using structural, functional and diffusion-weighted MRI acquired in 1000 healthy adults, we aimed to map the cortical correlates of identity and behavior at ultra-high spatial resolution. Using methods based on sparse matrix representations, we propose a computationally feasible high-resolution connectomic approach that improves neural fingerprinting and behavior prediction. Using this high-resolution approach, we find that the multimodal cortical gradients of individual uniqueness reside in the association cortices. Furthermore, our analyses identified a striking dichotomy between the facets of a person's neural identity that best predict their behavior and cognition, compared to those that best differentiate them from other individuals. Functional connectivity was one of the most accurate predictors of behavior, yet resided among the weakest differentiators of identity; whereas the converse was found for morphological properties, such as cortical curvature. This study provides new insights into the neural basis of personal identity and new tools to facilitate ultra-high-resolution connectomics., Competing Interests: Declaration of Competing Interest The authors declare no competing interests., (Copyright © 2021 The Author(s). Published by Elsevier Inc. All rights reserved.)

While the function of most biological systems is tightly constrained by their structure, current evidence suggests that coupling between the structure and function of brain networks is relatively modest. We aimed to investigate whether the modest coupling between connectome structure and function is a fundamental property of nervous systems or a limitation of current brain network models. We developed a new deep learning framework to predict an individual's brain function from their structural connectome, achieving prediction accuracies that substantially exceeded state-of-the-art biophysical models (group: R=0.9±0.1, individual: R=0.55±0.1). Crucially, brain function predicted from an individual's structural connectome explained significant inter-individual variation in cognitive performance. Our results suggest that structure-function coupling in human brain networks is substantially tighter than previously suggested. We establish the margin by which current brain network models can be improved and demonstrate how deep learning can facilitate investigation of relations between brain function and behavior., (Copyright © 2020. Published by Elsevier Inc.)

Early identification of individuals at risk of developing Alzheimer's disease (AD) dementia is important for developing disease-modifying therapies. In this study, given multimodal AD markers and clinical diagnosis of an individual from one or more timepoints, we seek to predict the clinical diagnosis, cognition and ventricular volume of the individual for every month (indefinitely) into the future. We proposed and applied a minimal recurrent neural network (minimalRNN) model to data from The Alzheimer's Disease Prediction Of Longitudinal Evolution (TADPOLE) challenge, comprising longitudinal data of 1677 participants (Marinescu et al., 2018) from the Alzheimer's Disease Neuroimaging Initiative (ADNI). We compared the performance of the minimalRNN model and four baseline algorithms up to 6 years into the future. Most previous work on predicting AD progression ignore the issue of missing data, which is a prevalent issue in longitudinal data. Here, we explored three different strategies to handle missing data. Two of the strategies treated the missing data as a "preprocessing" issue, by imputing the missing data using the previous timepoint ("forward filling") or linear interpolation ("linear filling). The third strategy utilized the minimalRNN model itself to fill in the missing data both during training and testing ("model filling"). Our analyses suggest that the minimalRNN with "model filling" compared favorably with baseline algorithms, including support vector machine/regression, linear state space (LSS) model, and long short-term memory (LSTM) model. Importantly, although the training procedure utilized longitudinal data, we found that the trained minimalRNN model exhibited similar performance, when using only 1 input timepoint or 4 input timepoints, suggesting that our approach might work well with just cross-sectional data. An earlier version of our approach was ranked 5th (out of 53 entries) in the TADPOLE challenge in 2019. The current approach is ranked 2nd out of 63 entries as of June 3rd, 2020., (Copyright © 2020. Published by Elsevier Inc.)

There is considerable interest in elucidating the cluster structure of brain networks in terms of modules, blocks or clusters of similar nodes. However, it is currently challenging to handle data on multiple subjects since most of the existing methods are applicable only on a subject-by-subject basis or for analysis of an average group network. The main limitation of per-subject models is that there is no obvious way to combine the results for group comparisons, and of group-averaged models that they do not reflect the variability between subjects. Here, we propose two new extensions of the classical Stochastic Blockmodel (SBM) that use a mixture model to estimate blocks or clusters of connected nodes, combined with a regression model to capture the effects of subject-level covariates on individual differences in cluster structure. The proposed Multi-Subject Stochastic Blockmodels (MS-SBMs) can flexibly account for between-subject variability in terms of homogeneous or heterogeneous covariate effects on connectivity using subject demographics such as age or diagnostic status. Using synthetic data, representing a range of block sizes and cluster structures, we investigate the accuracy of the estimated MS-SBM parameters as well as the validity of inference procedures based on the Wald, likelihood ratio and permutation tests. We show that the proposed multi-subject SBMs recover the true cluster structure of synthetic networks more accurately and adaptively than standard methods for modular decomposition (i.e. the Fast Louvain and Newman Spectral algorithms). Permutation tests of MS-SBM parameters were more robustly valid for statistical inference and Type I error control than tests based on standard asymptotic assumptions. Applied to analysis of multi-subject resting-state fMRI networks (13 healthy volunteers; 12 people with schizophrenia; n=268 brain regions), we show that Heterogeneous Stochastic Blockmodel (Het-SBM) identifies a range of network topologies simultaneously, including modular and core structures., (Copyright © 2020. Published by Elsevier Inc.)

Local cortical architecture is highly heritable and distinct genes are associated with specific cortical regions. Total surface area has been shown to be genetically correlated with complex cognitive capacities, suggesting cortical brain structure is a viable endophenotype linking genes to behavior. However, to what extend local brain structure has a genetic association with cognitive and emotional functioning is incompletely understood. Here, we study the genetic correlation between personality traits and local cortical structure in a large-scale twin sample (Human Connectome Project, n ​= ​1102, 22-37y) and we evaluated whether observed associations reflect generalizable relationships between personality and local brain structure two independent age-matched samples (Brain Genomics Superstructure Project: n ​= ​925, age ​= ​19-35y, enhanced Nathan Kline Institute dataset: n ​= ​209, age: 19-39y). We found a genetic overlap between personality traits and local cortical structure in 10 of 18 observed phenotypic associations in predominantly frontal cortices. However, we only observed evidence in favor of replication for the negative association between surface area in medial prefrontal cortex and Neuroticism in both replication samples. Quantitative functional decoding indicated this region is implicated in emotional and socio-cognitive functional processes. In sum, our observations suggest that associations between local brain structure and personality are, in part, under genetic control. However, associations are weak and only the relation between frontal surface area and Neuroticism was consistently observed across three independent samples of young adults., Competing Interests: Declaration of competing interest The authors declare no competing interests., (Copyright © 2020 The Authors. Published by Elsevier Inc. All rights reserved.)

Attention is a critical cognitive function, allowing humans to select, enhance, and sustain focus on information of behavioral relevance. Attention contains dissociable neural and psychological components. Nevertheless, some brain networks support multiple attentional functions. In this study, we used the visual attentional blink (VAB) as a test of the functional generalizability of the brain's attentional networks. In a VAB task, attention devoted to a target often causes a subsequent item to be missed. Although frequently attributed to limitations in attentional capacity or selection, VAB deficits attenuate when participants are distracted or deploy attention diffusely. The VAB is also behaviorally and theoretically dissociable from other attention tasks. Here we used Connectome-based Predictive Models (CPMs), which associate individual differences in task performance with functional connectivity patterns, to test their ability to predict performance for multiple attentional tasks. We constructed visual attentional blink (VAB) CPMs, and then used them and a sustained attention network model (saCPM; Rosenberg et al., 2016a) to predict performance. The latter model had been previously shown to successfully predict performance across tasks involving selective attention, inhibitory control, and even reading recall. Participants (n ​= ​73; 24 males) underwent fMRI while performing the VAB task and while resting. Outside the scanner, they completed other cognitive tasks over several days. A vabCPM constructed from VAB task data (behavior and fMRI) successfully predicted VAB performance. Strikingly, the network edges that predicted better VAB performance (positive edges) predicted worse performance for selective and sustained attention tasks, and vice versa. Predictions from applying the saCPM to the data mirrored these results, with the network's negative edges predicting better VAB performance. The vabCPM's positive edges partially yet significantly overlapped with the saCPM's negative edges, and vice versa. Many positive edges from the vabCPM involved the default mode network, whereas many negative edges involved the salience/ventral attention network. We conclude that the vabCPM and saCPM networks reflect general attentional functions that influence performance on many tasks. The networks may indicate an individual's propensity to deploy attention in a more diffuse or a more focused manner., Competing Interests: Declaration of competing interest None., (Copyright © 2020 The Authors. Published by Elsevier Inc. All rights reserved.)

The impact of in-scanner motion on functional magnetic resonance imaging (fMRI) data has a notorious reputation in the neuroimaging community. State-of-the-art guidelines advise to scrub out excessively corrupted frames as assessed by a composite framewise displacement (FD) score, to regress out models of nuisance variables, and to include average FD as a covariate in group-level analyses. Here, we studied individual motion time courses at time points typically retained in fMRI analyses. We observed that even in this set of putatively clean time points, motion exhibited a very clear spatio-temporal structure, so that we could distinguish subjects into separate groups of movers with varying characteristics. Then, we showed that this spatio-temporal motion cartography tightly relates to a broad array of anthropometric and cognitive factors. Convergent results were obtained from two different analytical perspectives: univariate assessment of behavioural differences across mover subgroups unraveled defining markers, while subsequent multivariate analysis broadened the range of involved factors and clarified that multiple motion/behaviour modes of covariance overlap in the data. Our results demonstrate that even the smaller episodes of motion typically retained in fMRI analyses carry structured, behaviourally relevant information. They call for further examinations of possible biases in current regression-based motion correction strategies., (Copyright © 2020 The Authors. Published by Elsevier Inc. All rights reserved.)

There is significant interest in the development and application of deep neural networks (DNNs) to neuroimaging data. A growing literature suggests that DNNs outperform their classical counterparts in a variety of neuroimaging applications, yet there are few direct comparisons of relative utility. Here, we compared the performance of three DNN architectures and a classical machine learning algorithm (kernel regression) in predicting individual phenotypes from whole-brain resting-state functional connectivity (RSFC) patterns. One of the DNNs was a generic fully-connected feedforward neural network, while the other two DNNs were recently published approaches specifically designed to exploit the structure of connectome data. By using a combined sample of almost 10,000 participants from the Human Connectome Project (HCP) and UK Biobank, we showed that the three DNNs and kernel regression achieved similar performance across a wide range of behavioral and demographic measures. Furthermore, the generic feedforward neural network exhibited similar performance to the two state-of-the-art connectome-specific DNNs. When predicting fluid intelligence in the UK Biobank, performance of all algorithms dramatically improved when sample size increased from 100 to 1000 subjects. Improvement was smaller, but still significant, when sample size increased from 1000 to 5000 subjects. Importantly, kernel regression was competitive across all sample sizes. Overall, our study suggests that kernel regression is as effective as DNNs for RSFC-based behavioral prediction, while incurring significantly lower computational costs. Therefore, kernel regression might serve as a useful baseline algorithm for future studies., (Copyright © 2019 Elsevier Inc. All rights reserved.)

Individuals with Alzheimer's disease (AD) dementia exhibit significant heterogeneity across clinical symptoms, atrophy patterns, and spatial distribution of Tau deposition. Most previous studies of AD heterogeneity have focused on atypical clinical subtypes, defined subtypes with a single modality, or restricted their analyses to a priori brain regions and cognitive tests. Here, we considered a data-driven hierarchical Bayesian model to identify latent factors from atrophy patterns and cognitive deficits simultaneously, thus exploiting the rich dimensionality within each modality. Unlike most previous studies, our model allows each factor to be expressed to varying degrees within an individual, in order to reflect potential multiple co-existing pathologies. By applying our model to ADNI-GO/2 AD dementia participants, we found three atrophy-cognitive factors. The first factor was associated with medial temporal lobe atrophy, episodic memory deficits and disorientation to time/place ("MTL-Memory"). The second factor was associated with lateral temporal atrophy and language deficits ("Lateral Temporal-Language"). The third factor was associated with atrophy in posterior bilateral cortex, and visuospatial executive function deficits ("Posterior Cortical-Executive"). While the MTL-Memory and Posterior Cortical-Executive factors were discussed in previous literature, the Lateral Temporal-Language factor is novel and emerged only by considering atrophy and cognition jointly. Several analyses were performed to ensure generalizability, replicability and stability of the estimated factors. First, the factors generalized to new participants within a 10-fold cross-validation of ADNI-GO/2 AD dementia participants. Second, the factors were replicated in an independent ADNI-1 AD dementia cohort. Third, factor loadings of ADNI-GO/2 AD dementia participants were longitudinally stable, suggesting that these factors capture heterogeneity across patients, rather than longitudinal disease progression. Fourth, the model outperformed canonical correlation analysis at capturing associations between atrophy patterns and cognitive deficits. To explore the influence of the factors early in the disease process, factor loadings were estimated in ADNI-GO/2 mild cognitively impaired (MCI) participants. Although the associations between the atrophy patterns and cognitive profiles were weak in MCI compared to AD, we found that factor loadings were associated with inter-individual regional variation in Tau uptake. Taken together, these results suggest that distinct atrophy-cognitive patterns exist in typical Alzheimer's disease, and are associated with distinct patterns of Tau depositions before clinical dementia emerges., (Copyright © 2019 Elsevier Inc. All rights reserved.)

Coordinate-based meta-analysis can provide important insights into mind-brain relationships. A popular approach for curated small-scale meta-analysis is activation likelihood estimation (ALE), which identifies brain regions consistently activated across a selected set of experiments, such as within a functional domain or mental disorder. ALE can also be utilized in meta-analytic co-activation modeling (MACM) to identify brain regions consistently co-activated with a seed region. Therefore, ALE aims to find consensus across experiments, treating heterogeneity across experiments as noise. However, heterogeneity within an ALE analysis of a functional domain might indicate the presence of functional sub-domains. Similarly, heterogeneity within a MACM analysis might indicate the involvement of a seed region in multiple co-activation patterns that are dependent on task contexts. Here, we demonstrate the use of the author-topic model to automatically determine if heterogeneities within ALE-type meta-analyses can be robustly explained by a small number of latent patterns. In the first application, the author-topic modeling of experiments involving self-generated thought (N = 179) revealed cognitive components fractionating the default network. In the second application, the author-topic model revealed that the left inferior frontal junction (IFJ) participated in multiple task-dependent co-activation patterns (N = 323). Furthermore, the author-topic model estimates compared favorably with spatial independent component analysis in both simulation and real data. Overall, the results suggest that the author-topic model is a flexible tool for exploring heterogeneity in ALE-type meta-analyses that might arise from functional sub-domains, mental disorder subtypes or task-dependent co-activation patterns. Code for this study is publicly available (https://github.com/ThomasYeoLab/CBIG/tree/master/stable_projects/meta-analysis/Ngo2019_AuthorTopic)., (Copyright © 2019 Elsevier Inc. All rights reserved.)

Global signal regression (GSR) is one of the most debated preprocessing strategies for resting-state functional MRI. GSR effectively removes global artifacts driven by motion and respiration, but also discards globally distributed neural information and introduces negative correlations between certain brain regions. The vast majority of previous studies have focused on the effectiveness of GSR in removing imaging artifacts, as well as its potential biases. Given the growing interest in functional connectivity fingerprinting, here we considered the utilitarian question of whether GSR strengthens or weakens associations between resting-state functional connectivity (RSFC) and multiple behavioral measures across cognition, personality and emotion. By applying the variance component model to the Brain Genomics Superstruct Project (GSP), we found that behavioral variance explained by whole-brain RSFC increased by an average of 47% across 23 behavioral measures after GSR. In the Human Connectome Project (HCP), we found that behavioral variance explained by whole-brain RSFC increased by an average of 40% across 58 behavioral measures, when GSR was applied after ICA-FIX de-noising. To ensure generalizability, we repeated our analyses using kernel regression. GSR improved behavioral prediction accuracies by an average of 64% and 12% in the GSP and HCP datasets respectively. Importantly, the results were consistent across methods. A behavioral measure with greater RSFC-explained variance (using the variance component model) also exhibited greater prediction accuracy (using kernel regression). A behavioral measure with greater improvement in behavioral variance explained after GSR (using the variance component model) also enjoyed greater improvement in prediction accuracy after GSR (using kernel regression). Furthermore, GSR appeared to benefit task performance measures more than self-reported measures. Since GSR was more effective at removing motion-related and respiratory-related artifacts, GSR-related increases in variance explained and prediction accuracies were unlikely the result of motion-related or respiratory-related artifacts. However, it is worth emphasizing that the current study focused on whole-brain RSFC, so it remains unclear whether GSR improves RSFC-behavioral associations for specific connections or networks. Overall, our results suggest that at least in the case for young healthy adults, GSR strengthens the associations between RSFC and most (although not all) behavioral measures. Code for the variance component model and ridge regression can be found here: https://github.com/ThomasYeoLab/CBIG/tree/master/stable_projects/preprocessing/Li2019_GSR., (Copyright © 2019 Elsevier Inc. All rights reserved.)

Component analysis is a powerful tool to identify dominant patterns of interactions in multivariate datasets. In the context of fMRI data, methods such as principal component analysis or independent component analysis have been used to identify the brain networks shaping functional connectivity (FC). Importantly, these approaches are static in the sense that they ignore the temporal information contained in fMRI time series. Therefore, the corresponding components provide a static characterization of FC. Building upon recent findings suggesting that FC dynamics encode richer information about brain functional organization, we use a dynamic extension of component analysis to identify dynamic modes (DMs) of fMRI time series. We demonstrate the feasibility and relevance of this approach using resting-state and motor-task fMRI data of 730 healthy subjects of the Human Connectome Project (HCP). In resting-state, dominant DMs have strong resemblance with classical resting-state networks, with an additional temporal characterization of the networks in terms of oscillatory periods and damping times. In motor-task conditions, dominant DMs reveal interactions between several brain areas, including but not limited to the posterior parietal cortex and primary motor areas, that are not found with classical activation maps. Finally, we identify two canonical components linking the temporal properties of the resting-state DMs with 158 behavioral and demographic HCP measures. Altogether, these findings illustrate the benefits of the proposed dynamic component analysis framework, making it a promising tool to characterize the spatio-temporal organization of brain activity., (Copyright © 2019 Elsevier Inc. All rights reserved.)

One of the most specific but also challenging properties of the brain is its topographic organization into distinct modules or cortical areas. In this paper, we first review the concept of topographic organization and its historical development. Next, we provide a critical discussion of the current definition of what constitutes a cortical area, why the concept has been so central to the field of neuroimaging and the challenges that arise from this view. A key aspect in this discussion is the issue of spatial scale and hierarchy in the brain. Focusing on in-vivo brain parcellation as a rapidly expanding field of research, we highlight potential limitations of the classical concept of cortical areas in the context of multi-modal parcellation and propose a revised interpretation of cortical areas building on the concept of neurobiological atoms that may be aggregated into larger units within and across modalities. We conclude by presenting an outlook on the implication of this revised concept for future mapping studies and raise some open questions in the context of brain parcellation., (Copyright © 2017 Elsevier Inc. All rights reserved.)

Resting-state functional connectivity is a powerful tool for studying human functional brain networks. Temporal fluctuations in functional connectivity, i.e., dynamic functional connectivity (dFC), are thought to reflect dynamic changes in brain organization and non-stationary switching of discrete brain states. However, recent studies have suggested that dFC might be attributed to sampling variability of static FC. Despite this controversy, a detailed exposition of stationarity and statistical testing of dFC is lacking in the literature. This article seeks an in-depth exploration of these statistical issues at a level appealing to both neuroscientists and statisticians. We first review the statistical notion of stationarity, emphasizing its reliance on ensemble statistics. In contrast, all FC measures depend on sample statistics. An important consequence is that the space of stationary signals is much broader than expected, e.g., encompassing hidden markov models (HMM) widely used to extract discrete brain states. In other words, stationarity does not imply the absence of brain states. We then expound the assumptions underlying the statistical testing of dFC. It turns out that the two popular frameworks - phase randomization (PR) and autoregressive randomization (ARR) - generate stationary, linear, Gaussian null data. Therefore, statistical rejection can be due to non-stationarity, nonlinearity and/or non-Gaussianity. For example, the null hypothesis can be rejected for the stationary HMM due to nonlinearity and non-Gaussianity. Finally, we show that a common form of ARR (bivariate ARR) is susceptible to false positives compared with PR and an adapted version of ARR (multivariate ARR). Application of PR and multivariate ARR to Human Connectome Project data suggests that the stationary, linear, Gaussian null hypothesis cannot be rejected for most participants. However, failure to reject the null hypothesis does not imply that static FC can fully explain dFC. We find that first order AR models explain temporal FC fluctuations significantly better than static FC models. Since first order AR models encode both static FC and one-lag FC, this suggests the presence of dynamical information beyond static FC. Furthermore, even in subjects where the null hypothesis was rejected, AR models explain temporal FC fluctuations significantly better than a popular HMM, suggesting the lack of discrete states (as measured by resting-state fMRI). Overall, our results suggest that AR models are not only useful as a means for generating null data, but may be a powerful tool for exploring the dynamical properties of resting-state fMRI. Finally, we discuss how apparent contradictions in the growing dFC literature might be reconciled., (Copyright © 2017 Elsevier Inc. All rights reserved.)

Neuroscience is undergoing faster changes than ever before. Over 100 years our field qualitatively described and invasively manipulated single or few organisms to gain anatomical, physiological, and pharmacological insights. In the last 10 years neuroscience spawned quantitative datasets of unprecedented breadth (e.g., microanatomy, synaptic connections, and optogenetic brain-behavior assays) and size (e.g., cognition, brain imaging, and genetics). While growing data availability and information granularity have been amply discussed, we direct attention to a less explored question: How will the unprecedented data richness shape data analysis practices? Statistical reasoning is becoming more important to distill neurobiological knowledge from healthy and pathological brain measurements. We argue that large-scale data analysis will use more statistical models that are non-parametric, generative, and mixing frequentist and Bayesian aspects, while supplementing classical hypothesis testing with out-of-sample predictions., (Copyright © 2017 The Authors. Published by Elsevier Inc. All rights reserved.)

Graph theoretical analysis has become an important tool in the examination of brain dysconnectivity in neurological and psychiatric brain disorders. A common analysis step in the construction of the functional graph or network involves "thresholding" of the connectivity matrix, selecting the set of edges that together form the graph on which network organization is evaluated. To avoid systematic differences in absolute number of edges, studies have argued against the use of an "absolute threshold" in case-control studies and have proposed the use of "proportional thresholding" instead, in which a pre-defined number of strongest connections are selected as network edges, ensuring equal network density across datasets. Here, we systematically studied the effect of proportional thresholding on the construction of functional matrices and subsequent graph analysis in patient-control functional connectome studies. In a few simple experiments we show that differences in overall strength of functional connectivity (FC) - as often observed between patients and controls - can have predictable consequences for between-group differences in network organization. In individual networks with lower overall FC the proportional thresholding algorithm has to select more edges based on lower correlations, which have (on average) a higher probability of being spurious, and thus introduces a higher degree of randomness in the resulting network. We show across both empirical and artificial patient-control datasets that lower levels of overall FC in either the patient or control group will most often lead to differences in network efficiency and clustering, suggesting that differences in FC across subjects will be artificially inflated or translated into differences in network organization. Based on the presented case-control findings we inform about the caveats of proportional thresholding in patient-control studies in which groups show a between-group difference in overall FC. We make recommendations on how to examine, report and to take into account overall FC effects in future patient-control functional connectome studies., (Copyright © 2017 The Authors. Published by Elsevier Inc. All rights reserved.)