Big Data Approaches for the Analysis of Large-Scale fMRI Data Using Apache Spark and GPU Processing: A Demonstration on Resting-State fMRI Data from the Human Connectome Project

AbstractFunctional connectivity (FC) networks are typically inferred from resting-state fMRI data using the Pearson correlation between BOLD time series from pairs of brain regions. However, alternative methods of estimating functional connectivity have not been systematically tested for their sensitivity or robustness to head motion artifact. Here, we evaluate the sensitivity of six different functional connectivity measures to motion artifact using resting-state data from the Human Connectome Project. We report that FC estimated using full correlation has a relatively high residual distance-dependent relationship with motion compared to partial correlation, coherence and information theory-based measures, even after implementing rigorous methods for motion artifact mitigation. This disadvantage of full correlation, however, may be offset by higher test-retest reliability and system identifiability. FC estimated by partial correlation offers the best of both worlds, with low sensitivity to motion artifact and intermediate system identifiability, with the caveat of low test-retest reliability. We highlight spatial differences in the sub-networks affected by motion with different FC metrics. Further, we report that intra-network edges in the default mode and retrosplenial temporal sub-networks are highly correlated with motion in all FC methods. Our findings indicate that the method of estimating functional connectivity is an important consideration in resting-state fMRI studies and must be chosen carefully based on the parameters of the study.

Download Full-text

Large-Scale Intrinsic Functional Brain Organization Emerges from Three Canonical Spatiotemporal Patterns

10.1101/2021.06.20.448984 ◽

2021 ◽

Author(s):

Taylor S Bolt ◽

Jason Nomi ◽

Danilo Bzdok ◽

Catie Chang ◽

B.T. Thomas Yeo ◽

...

Keyword(s):

Resting State ◽

Large Scale ◽

Low Frequency ◽

Resting State Fmri ◽

Spatiotemporal Patterns ◽

Brain Organization ◽

Functional Brain ◽

Temporal Properties ◽

Human Connectome Project ◽

Functional Brain Organization

The characterization of intrinsic functional brain organization has been approached from a multitude of analytic techniques and methods. We are still at a loss of a unifying conceptual framework for capturing common insights across this patchwork of empirical findings. By analyzing resting-state fMRI data from the Human Connectome Project using a large number of popular analytic techniques, we find that all results can be seamlessly reconciled by three fundamental low-frequency spatiotemporal patterns that we have identified via a novel time-varying complex pattern analysis. Overall, these three spatiotemporal patterns account for a wide variety of previously observed phenomena in the resting-state fMRI literature including the task-positive/task-negative anticorrelation, the global signal, the primary functional connectivity gradient and the network community structure of the functional connectome. The shared spatial and temporal properties of these three canonical patterns suggest that they arise from a single hemodynamic mechanism.

Download Full-text

Alternative Labeling Tool: a minimal algorithm for denoising single-subject resting-state fMRI data with ICA-MELODIC

10.21203/rs.3.rs-1061924/v1 ◽

2021 ◽

Author(s):

Peter Zhukovsky ◽

Gillian Coughlan ◽

Erin W Dickie ◽

Colin Hawco ◽

Aristotle N Voineskos

Keyword(s):

Resting State ◽

Large Scale ◽

Resting State Fmri ◽

Fmri Data ◽

Single Subject ◽

Imaging Studies ◽

Computationally Intensive ◽

Statistical Library ◽

User Friendly ◽

High Degree

Abstract Subject-level independent component analysis (ICA) is a well-established and widely used approach in denoising of resting-state functional magnetic resonance imaging (fMRI) data. However, approaches such as ICA-FIX and ICA-AROMA require advanced setups and/or are computationally intensive. Here, we aim to introduce a user-friendly, computationally lightweight toolbox for labeling independent signal and noise components, termed Alternative Labeling Tool (ALT). ALT uses two features that require manual tuning: proportion of an independent component’s spatial map located inside gray matter and positive skew of the power spectrum. ALT is tightly integrated with the commonly used FMRIB’s statistical library (FSL). Using the Open Access Series of Imaging Studies (OASIS-3) ageing dataset (n=30), we found that ALT shows a high degree of inter-rater agreement with manual labeling (over 86% of true positives for both signal and noise components on average). Crucially, denoising using ALT-generated labels significantly reduced mean framewise displacement (p<0.001). In conclusion, ALT can be extended to small and large-scale datasets when the use of more complex tools such as ICA-FIX is not possible. ALT will thus allow for more widespread adoption of ICA-based denoising of resting-state fMRI data.

Download Full-text

Computing personalized brain functional networks from fMRI using self-supervised deep learning

10.1101/2021.09.25.461829 ◽

2021 ◽

Author(s):

Hongming Li ◽

Srinivasan Dhivya ◽

Zaixu Cui ◽

Chuanjun Zhuo ◽

Raquel E. Gur ◽

...

Keyword(s):

Deep Learning ◽

Resting State ◽

Brain Function ◽

Resting State Fmri ◽

Fmri Data ◽

Functional Networks ◽

Brain Functional Networks ◽

Human Connectome Project ◽

Functional Homogeneity ◽

Decoder Architecture

ABSTRACTA novel self-supervised deep learning (DL) method is developed for computing bias-free, personalized brain functional networks (FNs) that provide unique opportunities to better understand brain function, behavior, and disease. Specifically, convolutional neural networks with an encoder-decoder architecture are employed to compute personalized FNs from resting-state fMRI data without utilizing any external supervision by optimizing functional homogeneity of personalized FNs in a self-supervised setting. We demonstrate that a DL model trained on fMRI scans from the Human Connectome Project can identify canonical FNs and generalizes well across four different datasets. We further demonstrate that the identified personalized FNs are informative for predicting individual differences in behavior, brain development, and schizophrenia status. Taken together, self-supervised DL allows for rapid, generalizable computation of personalized FNs.

Download Full-text

Human Connectome Project resting state fMRI data organized into 60 fine-scaled retinotopically organized regions in cortical areas V1, V2 and V3

10.1101/340174 ◽

2018 ◽

Cited By ~ 2

Author(s):

Debra Ann Dawson ◽

Zixuan Yin ◽

Jack Lam ◽

Amir Shmuel

Keyword(s):

Visual Cortex ◽

Resting State ◽

Resting State Fmri ◽

Regions Of Interest ◽

Fmri Data ◽

Cortical Areas ◽

Human Connectome Project ◽

Public Data ◽

Visual Areas ◽

Data Release

AbstractThe data comprises 60 regions of interest (ROIs) from V1, V2, and V3 of the human visual cortex. Preprocessed data from the Human Connectome Project (HCP) 900 subjects public data release were utilized: 220 subjects were randomly selected, each with 4 scans of resting state fMRI data. Given that these subjects did not have retinotopy scans performed, the visual areas were defined using an anatomical template from Benson et al. (2014). Visual areas from each hemisphere were further divided along dorsal-ventral lines into quadrants, resulting in 4 quadrants per subject. Within each quadrant, fine scaled ROIs were defined by subdividing each visual area into 5 regions according to eccentricity. These data may be useful for studying retinotopically organized functional connectivity in the visual cortex using the HCP 3 Tesla dataset.

Download Full-text

Leveraging big data for classification of children who stutter from fluent peers

10.1101/2020.10.28.359711 ◽

2020 ◽

Author(s):

Saige Rutherford ◽

Mike Angstadt ◽

Chandra Sripada ◽

Soo-Eun Chang

Keyword(s):

Big Data ◽

Large Scale ◽

Brain Networks ◽

Resting State Fmri ◽

Large Datasets ◽

Human Connectome Project ◽

Out Of Sample ◽

Neural Bases ◽

Clinical Populations

AbstractLarge datasets, consisting of hundreds or thousands of subjects, are becoming the new data standard within the neuroimaging community. While big data creates numerous benefits, such as detecting smaller effects, many of these big datasets have focused on non-clinical populations. The heterogeneity of clinical populations makes creating datasets of equal size and quality more challenging. There is a need for methods to connect these robust large datasets with the carefully curated clinical datasets collected over the past decades. In this study, we use resting-state fMRI data from the Adolescent Brain Cognitive Development study and the Human Connectome Project to discover generalizable brain features for use in an out-of-sample predictive model to classify young (3-10yrs) children who stutter from fluent peers. We achieve up to 72% classification accuracy using 10-fold cross validation. This study suggests that big data has the potential to yield generalizable biomarkers that are clinically meaningful. Specifically, this is the first study to demonstrate that big data-derived brain features can help differentiate children who stutter from their fluent peers and provide novel information on brain networks relevant to stuttering pathophysiology. The results provide a significant expansion to previous understanding of the neural bases of stuttering. In addition to auditory, somatomotor, and subcortical networks, the big data-based models highlight the importance of considering large scale brain networks supporting error sensitivity, attention, cognitive control, and emotion regulation/self-inspection in the neural bases of stuttering.

Download Full-text

Reconfiguration of Directed Functional Connectivity Among Neurocognitive Networks with Aging: Considering the Role of Thalamo-Cortical Interactions

Cerebral Cortex ◽

10.1093/cercor/bhaa334 ◽

2020 ◽

Author(s):

Moumita Das ◽

Vanshika Singh ◽

Lucina Q Uddin ◽

Arpan Banerjee ◽

Dipanjan Roy

Keyword(s):

Functional Connectivity ◽

Resting State ◽

Large Scale ◽

Network Connectivity ◽

Resting State Fmri ◽

Brain Network ◽

Fmri Data ◽

Independent Dataset ◽

Left And Right

Abstract A complete picture of how subcortical nodes, such as the thalamus, exert directional influence on large-scale brain network interactions across age remains elusive. Using directed functional connectivity and weighted net causal outflow on resting-state fMRI data, we provide evidence of a comprehensive reorganization within and between neurocognitive networks (default mode: DMN, salience: SN, and central executive: CEN) associated with age and thalamocortical interactions. We hypothesize that thalamus subserves both modality-specific and integrative hub role in organizing causal weighted outflow among large-scale neurocognitive networks. To this end, we observe that within-network directed functional connectivity is driven by thalamus and progressively weakens with age. Secondly, we find that age-associated increase in between CEN- and DMN-directed functional connectivity is driven by both the SN and the thalamus. Furthermore, left and right thalami act as a causal integrative hub exhibiting substantial interactions with neurocognitive networks with aging and play a crucial role in reconfiguring network outflow. Notably, these results were largely replicated on an independent dataset of matched young and old individuals. Our findings strengthen the hypothesis that the thalamus is a key causal hub balancing both within- and between-network connectivity associated with age and maintenance of cognitive functioning with aging.

Download Full-text

Representation Learning of Resting State fMRI with Variational Autoencoder

10.1101/2020.06.16.155937 ◽

2020 ◽

Author(s):

Jung-Hoon Kim ◽

Yizhen Zhang ◽

Kuan Han ◽

Minkyu Choi ◽

Zhongming Liu

Keyword(s):

Resting State ◽

Latent Variables ◽

Large Data ◽

Representation Learning ◽

Resting State Fmri ◽

Fmri Data ◽

Human Connectome Project ◽

Variational Autoencoder ◽

Short Period ◽

Task Positive Network

AbstractResting state functional magnetic resonance imaging (rs-fMRI) data exhibits complex but structured patterns. However, the underlying origins are unclear and entangled in rs-fMRI data. Here we establish a variational auto-encoder, as a generative model trainable with unsupervised learning, to disentangle the unknown sources of rs-fMRI activity. After being trained with large data from the Human Connectome Project, the model has learned to represent and generate patterns of cortical activity and connectivity using latent variables. Of the latent representation, its distribution reveals overlapping functional networks, and its geometry is unique to each individual. Our results support the functional opposition between the default mode network and the task-positive network, while such opposition is asymmetric and non-stationary. Correlations between latent variables, rather than cortical connectivity, can be used as a more reliable feature to accurately identify subjects from a large group, even if only a short period of data is available per subject.

Download Full-text

A multi-measure approach for assessing the performance of fMRI preprocessing strategies in resting-state functional connectivity

10.1101/837609 ◽

2019 ◽

Cited By ~ 2

Author(s):

Michalis Kassinopoulos ◽

Georgios D. Mitsis

Keyword(s):

Resting State ◽

Signal To Noise Ratio ◽

Resting State Fmri ◽

Motion Artifacts ◽

Fmri Data ◽

Quality Of Data ◽

Resting State Functional Connectivity ◽

Physiological Processes ◽

Human Connectome Project ◽

Low Pass

AbstractIt is well established that head motion and physiological processes (e.g. cardiac and breathing activity) should be taken into consideration when analyzing and interpreting results in fMRI studies. However, even though recent studies aimed to evaluate the performance of different preprocessing pipelines there is still no consensus on the optimal strategy. This is partly due to the fact that the quality control (QC) metrics used to evaluate differences in performance across pipelines have often yielded contradictory results. Furthermore, preprocessing techniques based on physiological recordings or data decomposition techniques (e.g. aCompCor) have not been comprehensively examined. Here, to address the aforementioned issues, we propose a framework that summarizes the scores from eight previously proposed and novel QC metrics to a reduced set of two QC metrics that reflect the signal-to-noise ratio and the reduction in motion artifacts and biases in the preprocessed fMRI data. Using this framework, we evaluate the performance of three commonly used practices on the quality of data: 1) Removal of nuisance regressors from fMRI data, 2) discarding motion-contaminated volumes (i.e., scrubbing) before regression, and 3) low-pass filtering the data and the nuisance regressors before their removal. Using resting-state fMRI data from the Human Connectome Project, we show that the best data quality, is achieved when the global signal (GS) and about 17% of principal components from white matter (WM) are removed from the data. Finally, we observe a small further improvement with low-pass filtering at 0.20 Hz, but not with scrubbing.

Download Full-text