Dimensionality estimation for optimal detection of functional networks in BOLD fMRI data

Revealing the Relevant Spatiotemporal Scale Underlying Whole-Brain Dynamics

The brain rapidly processes and adapts to new information by dynamically transitioning between whole-brain functional networks. In this whole-brain modeling study we investigate the relevance of spatiotemporal scale in whole-brain functional networks. This is achieved through estimating brain parcellations at different spatial scales (100-900 regions) and time series at different temporal scales (from milliseconds to seconds) generated by a whole-brain model fitted to fMRI data. We quantify the richness of the dynamic repertoire at each spatiotemporal scale by computing the entropy of transitions between whole-brain functional networks. The results show that the optimal relevant spatial scale is around 300 regions and a temporal scale of around 150 ms. Overall, this study provides much needed evidence for the relevant spatiotemporal scales and recommendations for analyses of brain dynamics.

Resolution-based spectral clustering for brain parcellation using functional MRI

BACKGROUND

Brain parcellation is important for exploiting neuroimaging data. Variability in physiology between individuals has led to the need for data-driven approaches to parcellation, with recent research focusing on simultaneously estimating and partitioning the network structure of the brain.

NEW METHOD

We view data preprocessing, parcellation, and parcel validation from the perspective of predictive modeling. The goal is to identify parcels in a way that best generalizes to unseen data. We utilize an uncertainty quantification approach from image science to define parcels as groups of unresolvable variables in the predictive model. Model parameters are chosen via cross-validation. Parcellation results are compared based on both their repeatability as well as their ability to describe held-out data.

RESULTS

The approach provides insight and strategies for open questions such as the choice of evaluation metrics, selection of model order, and the optimal tuning of preprocessing steps for functional imaging data.

COMPARISON WITH EXISTING METHODS

We compare new and established approaches using functional imaging data, where we find the proposed approach produces parcellations which are both more accurate and more repeatable than the current baseline clustering method. The metrics also demonstrate potential problems with overfitting for the baseline method, and with bias for other methods.

CONCLUSIONS

Clustering of resolution provides a principled and robust approach to brain parcellation which improves on the current baseline.

A method to compare the discriminatory power of data-driven methods: Application to ICA and IVA

BACKGROUND

The widespread application of data-driven factorization-based methods, such as independent component analysis (ICA), to functional magnetic resonance imaging data facilitates the study of neural function and how it is disrupted by psychiatric disorders such as schizophrenia. While the increasing number of these methods motivates a comparison of their relative performance, such a comparison is difficult to perform on real fMRI data, since the ground truth is, relatively, unknown and the alignment of factors across different methods is impractical and imprecise.

NEW METHOD

We present a novel method, global difference maps (GDMs), to compare the results of different fMRI analysis techniques on real fMRI data, quantify their relative performances, and highlight the differences between the decompositions visually.

COMPARISON WITH EXISTING METHODS

We apply this method to compare the performances of two different factorization-based methods, ICA and its multiset extension independent vector analysis (IVA), for the analysis of fMRI data from 109 patients with schizophrenia and 138 healthy controls during the performance of three tasks.

RESULTS

Through this application of GDMs, we find that IVA can determine regions that are more discriminatory between patients and controls than ICA, though IVA is less effective at emphasizing regions found in only a subset of the tasks.

CONCLUSIONS

These results demonstrate that GDMs are an effective way to compare the performances of different factorization-based methods as well as regression-based analyses.

Deconstructing multivariate decoding for the study of brain function

Multivariate decoding methods were developed originally as tools to enable accurate predictions in real-world applications. The realization that these methods can also be employed to study brain function has led to their widespread adoption in the neurosciences. However, prior to the rise of multivariate decoding, the study of brain function was firmly embedded in a statistical philosophy grounded on univariate methods of data analysis. In this way, multivariate decoding for brain interpretation grew out of two established frameworks: multivariate decoding for predictions in real-world applications, and classical univariate analysis based on the study and interpretation of brain activation. We argue that this led to two confusions, one reflecting a mixture of multivariate decoding for prediction or interpretation, and the other a mixture of the conceptual and statistical philosophies underlying multivariate decoding and classical univariate analysis. Here we attempt to systematically disambiguate multivariate decoding for the study of brain function from the frameworks it grew out of. After elaborating these confusions and their consequences, we describe six, often unappreciated, differences between classical univariate analysis and multivariate decoding. We then focus on how the common interpretation of what is signal and noise changes in multivariate decoding. Finally, we use four examples to illustrate where these confusions may impact the interpretation of neuroimaging data. We conclude with a discussion of potential strategies to help resolve these confusions in interpreting multivariate decoding results, including the potential departure from multivariate decoding methods for the study of brain function.

Quantifying the Interaction and Contribution of Multiple Datasets in Fusion: Application to the Detection of Schizophrenia

The extraction of information from multiple sets of data is a problem inherent to many disciplines. This is possible by either analyzing the data sets jointly as in data fusion or separately and then combining as in data integration. However, selecting the optimal method to combine and analyze multiset data is an ever-present challenge. The primary reason for this is the difficulty in determining the optimal contribution of each data set to an analysis as well as the amount of potentially exploitable complementary information among data sets. In this paper, we propose a novel classification rate-based technique to unambiguously quantify the contribution of each data set to a fusion result as well as facilitate direct comparisons of fusion methods on real data and apply a new method, independent vector analysis (IVA), to multiset fusion. This classification rate-based technique is used on functional magnetic resonance imaging data collected from 121 patients with schizophrenia and 150 healthy controls during the performance of three tasks. Through this application, we find that though optimal performance is achieved by exploiting all tasks, each task does not contribute equally to the result and this framework enables effective quantification of the value added by each task. Our results also demonstrate that data fusion methods are more powerful than data integration methods, with the former achieving a classification rate of 73.5 % and the latter achieving one of 70.9 %, a difference which we show is significant when all three tasks are analyzed together. Finally, we show that IVA, due to its flexibility, has equivalent or superior performance compared with the popular data fusion method, joint independent component analysis.

Sparsity Is Better with Stability: Combining Accuracy and Stability for Model Selection in Brain Decoding

Structured sparse methods have received significant attention in neuroimaging. These methods allow the incorporation of domain knowledge through additional spatial and temporal constraints in the predictive model and carry the promise of being more interpretable than non-structured sparse methods, such as LASSO or Elastic Net methods. However, although sparsity has often been advocated as leading to more interpretable models it can also lead to unstable models under subsampling or slight changes of the experimental conditions. In the present work we investigate the impact of using stability/reproducibility as an additional model selection criterion¹ on several different sparse (and structured sparse) methods that have been recently applied for fMRI brain decoding. We compare three different model selection criteria: (i) classification accuracy alone; (ii) classification accuracy and overlap between the solutions; (iii) classification accuracy and correlation between the solutions. The methods we consider include LASSO, Elastic Net, Total Variation, sparse Total Variation, Laplacian and Graph Laplacian Elastic Net (GraphNET). Our results show that explicitly accounting for stability/reproducibility during the model optimization can mitigate some of the instability inherent in sparse methods. In particular, using accuracy and overlap between the solutions as a joint optimization criterion can lead to solutions that are more similar in terms of accuracy, sparsity levels and coefficient maps even when different sparsity methods are considered.

An Automated, Adaptive Framework for Optimizing Preprocessing Pipelines in Task-Based Functional MRI

BOLD fMRI is sensitive to blood-oxygenation changes correlated with brain function; however, it is limited by relatively weak signal and significant noise confounds. Many preprocessing algorithms have been developed to control noise and improve signal detection in fMRI. Although the chosen set of preprocessing and analysis steps (the “pipeline”) significantly affects signal detection, pipelines are rarely quantitatively validated in the neuroimaging literature, due to complex preprocessing interactions. This paper outlines and validates an adaptive resampling framework for evaluating and optimizing preprocessing choices by optimizing data-driven metrics of task prediction and spatial reproducibility. Compared to standard “fixed” preprocessing pipelines, this optimization approach significantly improves independent validation measures of within-subject test-retest, and between-subject activation overlap, and behavioural prediction accuracy. We demonstrate that preprocessing choices function as implicit model regularizers, and that improvements due to pipeline optimization generalize across a range of simple to complex experimental tasks and analysis models. Results are shown for brief scanning sessions (<3 minutes each), demonstrating that with pipeline optimization, it is possible to obtain reliable results and brain-behaviour correlations in relatively small datasets.

A Novel Brain Networks Enhancement Model (BNEM) for BOLD fMRI Data Analysis With Highly Spatial Reproducibility

Independent component analysis aiming at detecting the functional connectivity among discrete cortical brain regions has been extensively used to explore the functional magnetic resonance imaging data. Although the independent components (ICs) were with relatively high quality, the noise embedding in ICs has a great impact on the true active/inactive region inference and the reproducibility, in postprocessing stage, e.g., the extraction of statistical parametrical maps (SPMs). In this paper, a novel brain network enhancement model (BNEM) is proposed, which mainly consists of two key techniques: 1) 3-D wavelet noise filter (3DWNF) for the meaningful ICs, which greatly suppresses noise and enforces the real activation inference of SPMs; and 2) a spatial reproducibility enhancement algorithm (SREA), aiming to improve the reproducibility of SPMs. The simulated experiment demonstrated that the postfiltering signals by 3DWNF were with higher correlation and less normalized mean square error to the ground truths than the prefiltering ones; SREA could further enhance the quality of most postfiltering ones, preserving the consistency with 3DWNF. The real data experiments also revealed that 1) 3DWNF could lead to more accurate preservation of the true positive voxels by correctly identifying the high proportionally misclassified voxels of the nonenhanced SPMs; 2) SREA could further improve the classification accuracy of the active/inactive voxels of SPMs corresponding to the 3DWNF denoised ICs; and 3) both 3DWNF and SREA contribute to the reproducibility enhancement of the reproduced SPMs by BNEM. Thus, BNEM is expected to have wide applicability in the neuroscience and clinical domain.

How many separable sources? Model selection in independent components analysis

Unlike mixtures consisting solely of non-Gaussian sources, mixtures including two or more Gaussian components cannot be separated using standard independent components analysis methods that are based on higher order statistics and independent observations. The mixed Independent Components Analysis/Principal Components Analysis (mixed ICA/PCA) model described here accommodates one or more Gaussian components in the independent components analysis model and uses principal components analysis to characterize contributions from this inseparable Gaussian subspace. Information theory can then be used to select from among potential model categories with differing numbers of Gaussian components. Based on simulation studies, the assumptions and approximations underlying the Akaike Information Criterion do not hold in this setting, even with a very large number of observations. Cross-validation is a suitable, though computationally intensive alternative for model selection. Application of the algorithm is illustrated using Fisher's iris data set and Howells' craniometric data set. Mixed ICA/PCA is of potential interest in any field of scientific investigation where the authenticity of blindly separated non-Gaussian sources might otherwise be questionable. Failure of the Akaike Information Criterion in model selection also has relevance in traditional independent components analysis where all sources are assumed non-Gaussian.

Comparing within-subject classification and regularization methods in fMRI for large and small sample sizes

In recent years, a variety of multivariate classifier models have been applied to fMRI, with different modeling assumptions. When classifying high-dimensional fMRI data, we must also regularize to improve model stability, and the interactions between classifier and regularization techniques are still being investigated. Classifiers are usually compared on large, multisubject fMRI datasets. However, it is unclear how classifier/regularizer models perform for within-subject analyses, as a function of signal strength and sample size. We compare four standard classifiers: Linear and Quadratic Discriminants, Logistic Regression and Support Vector Machines. Classification was performed on data in the linear kernel (covariance) feature space, and classifiers are tuned with four commonly-used regularizers: Principal Component and Independent Component Analysis, and penalization of kernel features using L₁ and L₂ norms. We evaluated prediction accuracy (P) and spatial reproducibility (R) of all classifier/regularizer combinations on single-subject analyses, over a range of three different block task contrasts and sample sizes for a BOLD fMRI experiment. We show that the classifier model has a small impact on signal detection, compared to the choice of regularizer. PCA maximizes reproducibility and global SNR, whereas Lp -norms tend to maximize prediction. ICA produces low reproducibility, and prediction accuracy is classifier-dependent. However, trade-offs in (P,R) depend partly on the optimization criterion, and PCA-based models are able to explore the widest range of (P,R) values. These trends are consistent across task contrasts and data sizes (training samples range from 6 to 96 scans). In addition, the trends in classifier performance are consistent for ROI-based classifier analyses.

Pattern classification of fMRI data: applications for analysis of spatially distributed cortical networks

The field of fMRI data analysis is rapidly growing in sophistication, particularly in the domain of multivariate pattern classification. However, the interaction between the properties of the analytical model and the parameters of the BOLD signal (e.g. signal magnitude, temporal variance and functional connectivity) is still an open problem. We addressed this problem by evaluating a set of pattern classification algorithms on simulated and experimental block-design fMRI data. The set of classifiers consisted of linear and quadratic discriminants, linear support vector machine, and linear and nonlinear Gaussian naive Bayes classifiers. For linear discriminant, we used two methods of regularization: principal component analysis, and ridge regularization. The classifiers were used (1) to classify the volumes according to the behavioral task that was performed by the subject, and (2) to construct spatial maps that indicated the relative contribution of each voxel to classification. Our evaluation metrics were: (1) accuracy of out-of-sample classification and (2) reproducibility of spatial maps. In simulated data sets, we performed an additional evaluation of spatial maps with ROC analysis. We varied the magnitude, temporal variance and connectivity of simulated fMRI signal and identified the optimal classifier for each simulated environment. Overall, the best performers were linear and quadratic discriminants (operating on principal components of the data matrix) and, in some rare situations, a nonlinear Gaussian naïve Bayes classifier. The results from the simulated data were supported by within-subject analysis of experimental fMRI data, collected in a study of aging. This is the first study that systematically characterizes interactions between analysis model and signal parameters (such as magnitude, variance and correlation) on the performance of pattern classifiers for fMRI.

Increased insula coactivation with salience networks in insomnia

Insomnia is among the most prevalent and costly of all sleep-related disorders. To characterize the neural mechanisms underlying subjective dysfunction in insomnia, we examined brain activity in 17 female insomniacs and 17 female healthy controls using simultaneous functional magnetic resonance imaging (fMRI) and electroencephalography (EEG) while they were resting and while they were trying to fall asleep. In examining the dynamic regional activity within intrinsic brain networks, we found that, compared with controls, insomniacs had greater involvement of the anterior insula with salience networks, as well as insula BOLD correlation with EEG gamma frequency power during rest in insomniacs. This increased involvement of the anterior insula was associated with negative affect in insomniacs. Aberrant activation of the insula, which integrates temporal and bodily states, in arousal networks may underlie the misperception of sleep quality and subjective distress in insomnia.

Addictions and Personality Traits: Impulsivity and Related Constructs

Behavioral tendencies that might be captured through self-report measures may provide insight into personality features that are associated with substance addictions. Recently, impulsivity and related constructs, such as sensation-seeking, have been examined to help better understand their relationships with addictions. Here, we review recent findings that show links over developmental epochs between addictive behaviors and impulsivity, sensation-seeking, and other constructs that are theoretically linked. These findings have significant implications for generating improved treatments and interventions aimed at preventing the development of addictive disorders.

Age differences in the intrinsic functional connectivity of default network subsystems

Recent work suggests that the default mode network (DMN) includes two core regions, the ventromedial prefrontal cortex and posterior cingulate cortex (PCC), and several unique subsystems that are functionally distinct. These include a medial temporal lobe (MTL) subsystem, active during remembering and future projection, and a dorsomedial prefrontal cortex (dmPFC) subsystem, active during self-reference. The PCC has been further subdivided into ventral (vPCC) and dorsal (dPCC) regions that are more strongly connected with the DMN and cognitive control networks, respectively. The goal of this study was to examine age differences in resting state functional connectivity within these subsystems. After applying a rigorous procedure to reduce the effects of head motion, we used a multivariate technique to identify both common and unique patterns of functional connectivity in the MTL vs. the dmPFC, and in vPCC vs. dPCC. All four areas had robust functional connectivity with other DMN regions, and each also showed distinct connectivity patterns in both age groups. Young and older adults had equivalent functional connectivity in the MTL subsystem. Older adults showed weaker connectivity in the vPCC and dmPFC subsystems, particularly with other DMN areas, but stronger connectivity than younger adults in the dPCC subsystem, which included areas involved in cognitive control. Our data provide evidence for distinct subsystems involving DMN nodes, which are maintained with age. Nevertheless, there are age differences in the strength of functional connectivity within these subsystems, supporting prior evidence that DMN connectivity is particularly vulnerable to age, whereas connectivity involving cognitive control regions is relatively maintained. These results suggest an age difference in the integrated activity among brain networks that can have implications for cognition in older adults.

On applicability of PCA, voxel-wise variance normalization and dimensionality assumptions for sliding temporal window sICA in resting-state fMRI

Subject-level resting-state fMRI (RS-fMRI) spatial independent component analysis (sICA) may provide new ways to analyze the data when performed in the sliding time window. However, whether principal component analysis (PCA) and voxel-wise variance normalization (VN) are applicable pre-processing procedures in the sliding-window context, as they are for regular sICA, has not been addressed so far. Also model order selection requires further studies concerning sliding-window sICA. In this paper we have addressed these concerns. First, we compared PCA-retained subspaces concerning overlapping parts of consecutive temporal windows to answer whether in-window PCA and VN can confound comparisons between sICA analyses in consecutive windows. Second, we compared the PCA subspaces between windowed and full data to assess expected comparability between windowed and full-data sICA results. Third, temporal evolution of dimensionality estimates in RS-fMRI data sets was monitored to identify potential challenges in model order selection in a sliding-window sICA context. Our results illustrate that in-window VN can be safely used, in-window PCA is applicable with most window widths and that comparisons between windowed and full data should not be performed from a subspace similarity point of view. In addition, our studies on dimensionality estimates demonstrated that there are sustained, periodic and very case-specific changes in signal-to-noise ratio within RS-fMRI data sets. Consequently, dimensionality estimation is needed for well-founded model order determination in the sliding-window case. The observed periodic changes correspond to a frequency band of ≤0.1 Hz, which is commonly associated with brain activity in RS-fMRI and become on average most pronounced at window widths of 80 and 60 time points (144 and 108 s, respectively). Wider windows provided only slightly better comparability between consecutive windows, and 60 time point or shorter windows also provided the best comparability with full-data results. Further studies are needed to determine the cause for dimensionality variations.

Dimensionality of brain networks linked to life-long individual differences in self-control

The ability to delay gratification in childhood has been linked to positive outcomes in adolescence and adulthood. Here we examine a subsample of participants from a seminal longitudinal study of self-control throughout a subject’s lifespan. Self control, first studied in children at age 4, is now reexamined 40 years later, on a task that required control over the contents of working memory. We examine whether patterns of brain activation on this task can reliably distinguish participants with consistently low and high self-control abilities (low vs. high delayers). We find that low delayers recruit significantly higher-dimensional neural networks when performing the task compared to high delayers. High delayers are also more homogeneous as a group in their neural patterns compared to low delayers. From these brain patterns we can predict with 71% accuracy, whether a participant is a high or low delayer. The present results suggest that dimensionality of neural networks is a biological predictor of self-control abilities.

PHYCAA+: an optimized, adaptive procedure for measuring and controlling physiological noise in BOLD fMRI

The presence of physiological noise in functional MRI can greatly limit the sensitivity and accuracy of BOLD signal measurements, and produce significant false positives. There are two main types of physiological confounds: (1) high-variance signal in non-neuronal tissues of the brain including vascular tracts, sinuses and ventricles, and (2) physiological noise components which extend into gray matter tissue. These physiological effects may also be partially coupled with stimuli (and thus the BOLD response). To address these issues, we have developed PHYCAA+, a significantly improved version of the PHYCAA algorithm (Churchill et al., 2011) that (1) down-weights the variance of voxels in probable non-neuronal tissue, and (2) identifies the multivariate physiological noise subspace in gray matter that is linked to non-neuronal tissue. This model estimates physiological noise directly from EPI data, without requiring external measures of heartbeat and respiration, or manual selection of physiological components. The PHYCAA+ model significantly improves the prediction accuracy and reproducibility of single-subject analyses, compared to PHYCAA and a number of commonly-used physiological correction algorithms. Individual subject denoising with PHYCAA+ is independently validated by showing that it consistently increased between-subject activation overlap, and minimized false-positive signal in non gray-matter loci. The results are demonstrated for both block and fast single-event task designs, applied to standard univariate and adaptive multivariate analysis models.

Functional subdivision of group-ICA results of fMRI data collected during cinema viewing

Independent component analysis (ICA) can unravel functional brain networks from functional magnetic resonance imaging (fMRI) data. The number of the estimated components affects both the spatial pattern of the identified networks and their time-course estimates. Here group-ICA was applied at four dimensionalities (10, 20, 40, and 58 components) to fMRI data collected from 15 subjects who viewed a 15-min silent film ("At land" by Maya Deren). We focused on the dorsal attention network, the default-mode network, and the sensorimotor network. The lowest dimensionalities demonstrated most prominent activity within the dorsal attention network, combined with the visual areas, and in the default-mode network; the sensorimotor network only appeared with ICA comprising at least 20 components. The results suggest that even very low-dimensional ICA can unravel the most prominent functionally-connected brain networks. However, increasing the number of components gives a more detailed picture and functionally feasible subdivision of the major networks. These results improve our understanding of the hierarchical subdivision of brain networks during viewing of a movie that provides continuous stimulation embedded in an attention-directing narrative.

Optimizing preprocessing and analysis pipelines for single-subject fMRI: 2. Interactions with ICA, PCA, task contrast and inter-subject heterogeneity

A variety of preprocessing techniques are available to correct subject-dependant artifacts in fMRI, caused by head motion and physiological noise. Although it has been established that the chosen preprocessing steps (or "pipeline") may significantly affect fMRI results, it is not well understood how preprocessing choices interact with other parts of the fMRI experimental design. In this study, we examine how two experimental factors interact with preprocessing: between-subject heterogeneity, and strength of task contrast. Two levels of cognitive contrast were examined in an fMRI adaptation of the Trail-Making Test, with data from young, healthy adults. The importance of standard preprocessing with motion correction, physiological noise correction, motion parameter regression and temporal detrending were examined for the two task contrasts. We also tested subspace estimation using Principal Component Analysis (PCA), and Independent Component Analysis (ICA). Results were obtained for Penalized Discriminant Analysis, and model performance quantified with reproducibility (R) and prediction metrics (P). Simulation methods were also used to test for potential biases from individual-subject optimization. Our results demonstrate that (1) individual pipeline optimization is not significantly more biased than fixed preprocessing. In addition, (2) when applying a fixed pipeline across all subjects, the task contrast significantly affects pipeline performance; in particular, the effects of PCA and ICA models vary with contrast, and are not by themselves optimal preprocessing steps. Also, (3) selecting the optimal pipeline for each subject improves within-subject (P,R) and between-subject overlap, with the weaker cognitive contrast being more sensitive to pipeline optimization. These results demonstrate that sensitivity of fMRI results is influenced not only by preprocessing choices, but also by interactions with other experimental design factors. This paper outlines a quantitative procedure to denoise data that would otherwise be discarded due to artifact; this is particularly relevant for weak signal contrasts in single-subject, small-sample and clinical datasets.

SimTB, a simulation toolbox for fMRI data under a model of spatiotemporal separability

We introduce SimTB, a MATLAB toolbox designed to simulate functional magnetic resonance imaging (fMRI) datasets under a model of spatiotemporal separability. The toolbox meets the increasing need of the fMRI community to more comprehensively understand the effects of complex processing strategies by providing a ground truth that estimation methods may be compared against. SimTB captures the fundamental structure of real data, but data generation is fully parameterized and fully controlled by the user, allowing for accurate and precise comparisons. The toolbox offers a wealth of options regarding the number and configuration of spatial sources, implementation of experimental paradigms, inclusion of tissue-specific properties, addition of noise and head movement, and much more. A straightforward data generation method and short computation time (3-10 seconds for each dataset) allow a practitioner to simulate and analyze many datasets to potentially understand a problem from many angles. Beginning MATLAB users can use the SimTB graphical user interface (GUI) to design and execute simulations while experienced users can write batch scripts to automate and customize this process. The toolbox is freely available at http://mialab.mrn.org/software together with sample scripts and tutorials.

Nonlinear denoising and analysis of neuroimages with kernel principal component analysis and pre-image estimation

We investigate the use of kernel principal component analysis (PCA) and the inverse problem known as pre-image estimation in neuroimaging: i) We explore kernel PCA and pre-image estimation as a means for image denoising as part of the image preprocessing pipeline. Evaluation of the denoising procedure is performed within a data-driven split-half evaluation framework. ii) We introduce manifold navigation for exploration of a nonlinear data manifold, and illustrate how pre-image estimation can be used to generate brain maps in the continuum between experimentally defined brain states/classes. We base these illustrations on two fMRI BOLD data sets - one from a simple finger tapping experiment and the other from an experiment on object recognition in the ventral temporal lobe.

Capturing inter-subject variability with group independent component analysis of fMRI data: a simulation study

A key challenge in functional neuroimaging is the meaningful combination of results across subjects. Even in a sample of healthy participants, brain morphology and functional organization exhibit considerable variability, such that no two individuals have the same neural activation at the same location in response to the same stimulus. This inter-subject variability limits inferences at the group-level as average activation patterns may fail to represent the patterns seen in individuals. A promising approach to multi-subject analysis is group independent component analysis (GICA), which identifies group components and reconstructs activations at the individual level. GICA has gained considerable popularity, particularly in studies where temporal response models cannot be specified. However, a comprehensive understanding of the performance of GICA under realistic conditions of inter-subject variability is lacking. In this study we use simulated functional magnetic resonance imaging (fMRI) data to determine the capabilities and limitations of GICA under conditions of spatial, temporal, and amplitude variability. Simulations, generated with the SimTB toolbox, address questions that commonly arise in GICA studies, such as: (1) How well can individual subject activations be estimated and when will spatial variability preclude estimation? (2) Why does component splitting occur and how is it affected by model order? (3) How should we analyze component features to maximize sensitivity to intersubject differences? Overall, our results indicate an excellent capability of GICA to capture between-subject differences and we make a number of recommendations regarding analytic choices for application to functional imaging data.

Group-ICA Model Order Highlights Patterns of Functional Brain Connectivity

Resting-state networks (RSNs) can be reliably and reproducibly detected using independent component analysis (ICA) at both individual subject and group levels. Altering ICA dimensionality (model order) estimation can have a significant impact on the spatial characteristics of the RSNs as well as their parcellation into sub-networks. Recent evidence from several neuroimaging studies suggests that the human brain has a modular hierarchical organization which resembles the hierarchy depicted by different ICA model orders. We hypothesized that functional connectivity between-group differences measured with ICA might be affected by model order selection. We investigated differences in functional connectivity using so-called dual regression as a function of ICA model order in a group of unmedicated seasonal affective disorder (SAD) patients compared to normal healthy controls. The results showed that the detected disease-related differences in functional connectivity alter as a function of ICA model order. The volume of between-group differences altered significantly as a function of ICA model order reaching maximum at model order 70 (which seems to be an optimal point that conveys the largest between-group difference) then stabilized afterwards. Our results show that fine-grained RSNs enable better detection of detailed disease-related functional connectivity changes. However, high model orders show an increased risk of false positives that needs to be overcome. Our findings suggest that multilevel ICA exploration of functional connectivity enables optimization of sensitivity to brain disorders.

Optimizing preprocessing and analysis pipelines for single-subject fMRI. I. Standard temporal motion and physiological noise correction methods

Subject-specific artifacts caused by head motion and physiological noise are major confounds in BOLD fMRI analyses. However, there is little consensus on the optimal choice of data preprocessing steps to minimize these effects. To evaluate the effects of various preprocessing strategies, we present a framework which comprises a combination of (1) nonparametric testing including reproducibility and prediction metrics of the data-driven NPAIRS framework (Strother et al. [2002]: NeuroImage 15:747-771), and (2) intersubject comparison of SPM effects, using DISTATIS (a three-way version of metric multidimensional scaling (Abdi et al. [2009]: NeuroImage 45:89-95). It is shown that the quality of brain activation maps may be significantly limited by sub-optimal choices of data preprocessing steps (or "pipeline") in a clinical task-design, an fMRI adaptation of the widely used Trail-Making Test. The relative importance of motion correction, physiological noise correction, motion parameter regression, and temporal detrending were examined for fMRI data acquired in young, healthy adults. Analysis performance and the quality of activation maps were evaluated based on Penalized Discriminant Analysis (PDA). The relative importance of different preprocessing steps was assessed by (1) a nonparametric Friedman rank test for fixed sets of preprocessing steps, applied to all subjects; and (2) evaluating pipelines chosen specifically for each subject. Results demonstrate that preprocessing choices have significant, but subject-dependant effects, and that individually-optimized pipelines may significantly improve the reproducibility of fMRI results over fixed pipelines. This was demonstrated by the detection of a significant interaction with motion parameter regression and physiological noise correction, even though the range of subject head motion was small across the group (≪ 1 voxel). Optimizing pipelines on an individual-subject basis also revealed brain activation patterns either weak or absent under fixed pipelines, which has implications for the overall interpretation of fMRI data, and the relative importance of preprocessing methods.

Data-driven optimization and evaluation of 2D EPI and 3D PRESTO for BOLD fMRI at 7 Tesla: I. Focal coverage

Blood oxygenation level dependent (BOLD) functional magnetic resonance imaging (fMRI) is commonly performed using 2D single-shot echo-planar imaging (EPI). However, single-shot EPI at 7 Tesla (T) often suffers from significant geometric distortions (due to low bandwidth (BW) in the phase-encode (PE) direction) and amplified physiological noise. Recent studies have suggested that 3D multi-shot sequences such as PRESTO may offer comparable BOLD contrast-to-noise ratio with increased volume coverage and decreased geometric distortions. Thus, a four-way group-level comparison was performed between 2D and 3D acquisition sequences at two in-plane resolutions. The quality of fMRI data was evaluated via metrics of prediction and reproducibility using NPAIRS (Non-parametric Prediction, Activation, Influence and Reproducibility re-Sampling). Group activation maps were optimized for each acquisition strategy by selecting the number of principal components that jointly maximized prediction and reproducibility, and showed good agreement in sensitivity and specificity for positive BOLD changes. High-resolution EPI exhibited the highest z-scores of the four acquisition sequences; however, it suffered from the lowest BW in the PE direction (resulting in the worst geometric distortions) and limited spatial coverage, and also caused some subject discomfort through peripheral nerve stimulation (PNS). In comparison, PRESTO also had high z-scores (higher than EPI for a matched in-plane resolution), the highest BW in the PE direction (producing images with superior geometric fidelity), the potential for whole-brain coverage, and no reported PNS. This study provides evidence to support the use of 3D multi-shot acquisition sequences in lieu of single-shot EPI for ultra high field BOLD fMRI at 7T.