Separating 4D multi-task fMRI data of multiple subjects by independent component analysis with projection

Current analytical techniques applied to functional magnetic resonance imaging (fMRI) data require a priori knowledge or specific assumptions about the time courses of processes contributing to the measured signals. Here we describe a new method for analyzing fMRI data based on the independent component analysis (ICA) algorithm of Bell and Sejnowski ([1995]: Neural Comput 7:1129-1159). We decomposed eight fMRI data sets from 4 normal subjects performing Stroop color-naming, the Brown and Peterson work/number task, and control tasks into spatially independent components. Each component consisted of voxel values at fixed three-dimensional locations (a component "map"), and a unique associated time course of activation. Given data from 144 time points collected during a 6-min trial, ICA extracted an equal number of spatially independent components. In all eight trials, ICA derived one and only one component with a time course closely matching the time course of 40-sec alternations between experimental and control tasks. The regions of maximum activity in these consistently task-related components generally overlapped active regions detected by standard correlational analysis, but included frontal regions not detected by correlation. Time courses of other ICA components were transiently task-related, quasiperiodic, or slowly varying. By utilizing higher-order statistics to enforce successively stricter criteria for spatial independence between component maps, both the ICA algorithm and a related fourth-order decomposition technique (Comon [1994]: Signal Processing 36:11-20) were superior to principal component analysis (PCA) in determining the spatial and temporal extent of task-related activation. For each subject, the time courses and active regions of the task-related ICA components were consistent across trials and were robust to the addition of simulated noise. Simulated movement artifact and simulated task-related activations added to actual fMRI data were clearly separated by the algorithm. ICA can be used to distinguish between nontask-related signal components, movements, and other artifacts, as well as consistently or transiently task-related fMRI activations, based on only weak assumptions about their spatial distributions and without a priori assumptions about their time courses. ICA appears to be a highly promising method for the analysis of fMRI data from normal and clinical populations, especially for uncovering unpredictable transient patterns of brain activity associated with performance of psychomotor tasks.

SCTICA: Sub-packet constrained temporal ICA method for fMRI data analysis

Independent Component Analysis (ICA) is a popular method that uses statistical principles to separate the mixture into statistically independent non-Gaussian sources. It has been well used in functional Magnetic Resonance Imaging (fMRI) data. However, real fMRI data can rarely be accurately modeled as mixtures of independent components, the convergence of ICA may be impaired. This paper is based on the idea of preconditioned ICA for real data (Picard), which involves a preprocessing L-BFGS strategy based on orthogonal matrix sets. In this study, we designed an experiment to validate the idea that Picard can improve ICA algorithms such as Infomax, Extended-Infomax, and FastICA, respectively named Picard 1, Picard 2, and Picard 3, for fMRI data analysis. Three Picard versions were performed on the simulated and noisy fMRI mixtures to verify the ability to separate independent sources. Experimental results showed that Picard 3 outperformed Picard 1 and Picard 2 on both noiseless and noisy simulated fMRI data, which implied the priority of Picard 3 in fMRI data analysis.

Blind source separation (BSS) is commonly used in functional magnetic resonance imaging (fMRI) data analysis. Recently, BSS models based on restricted Boltzmann machine (RBM), one of the building blocks of deep learning models, have been shown to improve brain network identification compared to conventional single matrix factorization models such as independent component analysis (ICA). These models, however, trained RBM on fMRI volumes, and are hence challenged by model complexity and limited training set. In this article, we propose to apply RBM to fMRI time courses instead of volumes for BSS. The proposed method not only interprets fMRI time courses explicitly to take advantages of deep learning models in latent feature learning but also substantially reduces model complexity and increases the scale of training set to improve training efficiency. Our experimental results based on Human Connectome Project (HCP) datasets demonstrated the superiority of the proposed method over ICA and the one that applied RBM to fMRI volumes in identifying task-related components, resulted in more accurate and specific representations of task-related activations. Moreover, our method separated out components representing intermixed effects between task events, which could reflect inherent interactions among functionally connected brain regions. Our study demonstrates the value of RBM in mining complex structures embedded in large-scale fMRI data and its potential as a building block for deeper models in fMRI data analysis.

In application to functional magnetic resonance imaging (fMRI) data analysis, a number of data fusion algorithms have shown success in extracting interpretable brain networks that can distinguish two groups such two populations-patients with mental disorder and the healthy controls. However, there are situations where more than two groups exist such as the fusion of multi-task fMRI data. Therefore, in this work we propose the use of IVA to effectively extract information that is able to distinguish across multiple groups when applied to data fusion. The performance of IVA is investigated using a simulated fMRI-like data. The simulation results illustrate that IVA with multivariate Laplacian distribution and second-order statistics (IVA-L-SOS) yields better performance compared to joint independent component analysis and IVA with multivariate Gaussian distribution in terms of both estimation accuracy and robustness. When applied to real multi-task fMRI data, IVA-L-SOS successfully extract task-related brain networks that are able to distinguish three tasks.

Validating the performance of one-time decomposition for fMRI analysis using ICA with automatic target generation process

A two-step super-Gaussian independent component analysis approach for fMRI data

A new adaptive signal-preserving technique for noise suppression in functional magnetic resonance imaging (fMRI) data is proposed based on spectrum subtraction. The proposed technique estimates a model for the power spectrum of random noise from the acquired data. This model is used to estimate a noise-suppressed power spectrum for any given pixel time course by simple subtraction of power spectra. The new technique is tested using computer simulations and real data for event-related fMRI experiments. The results show the potential of the new technique in suppressing noise while preserving the other deterministic components. Moreover, further analysis using principal component analysis (PCA) and independent component analysis (ICA) shows a significant improvement in both convergence and clarity of results when the new technique is used. This suggests the value of the new technique as a useful preprocessing step for this type of signal.

A data-driven unsupervised learning such as an independent component analysis was gainfully applied to bloodoxygenation- level-dependent (BOLD) functional magnetic resonance imaging (fMRI) data compared to a model-based general linear model (GLM). This is due to an ability of this unsupervised learning method to extract a meaningful neuronal activity from BOLD signal that is a mixture of confounding non-neuronal artifacts such as head motions and physiological artifacts as well as neuronal signals. In this study, we support this claim by identifying neuronal underpinnings of cigarette craving and cigarette resistance. The fMRI data were acquired from heavy cigarette smokers (n = 14) while they alternatively watched images with and without cigarette smoking. During acquisition of two fMRI runs, they were asked to crave when they watched cigarette smoking images or to resist the urge to smoke. Data driven approaches of group independent component analysis (GICA) method based on temporal concatenation (TC) and TCGICA with an extension of iterative dual-regression (TC-GICA-iDR) were applied to the data. From the results, cigarette craving and cigarette resistance related neuronal activations were identified in the visual area and superior frontal areas, respectively with a greater statistical significance from the TC-GICA-iDR method than the TC-GICA method. On the other hand, the neuronal activity levels in many of these regions were not statistically different from the GLM method between the cigarette craving and cigarette resistance due to potentially aberrant BOLD signals.

The last two decades have seen a surge in the interest in research based on Functional Magnetic Resonance Imaging (fMRI) data. Decoding of cognitive states based on fMRI activation profiles has become a very active topic in this area. fMRI data is very high dimensional and noisy. However, there is a dearth of datasets to work on. Sharing of learning by analyzing datasets drawn across multiple subjects in an experiment can help in increasing the amount of data we have for analysis. Decoding of cognitive states using classifiers trained across multiple subjects is a challenging task because of differences in anatomy and cognition. Selecting features to analyze from the dataset is a key step in the analysis of fMRI data. In this paper we apply PCA, ICA and five non-linear dimensionality reduction techniques to the fMRI data. The aim of this work is to analyze which technique can provide the best feature selection to capture the commonality across multiple subjects. The reduced datasets are then used to train classifiers to solve a multiple-subject decoding problem.KeywordsFeature SelectionIndependent Component AnalysisCognitive StateCanonical Correlation AnalysisFeature Selection MethodThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Analysis of time-evolving data is crucial to understand the functioning of dynamic systems such as the brain. For instance, analysis of functional magnetic resonance imaging (fMRI) data collected during a task may reveal spatial regions of interest, and how they evolve during the task. However, capturing underlying spatial patterns as well as their change in time is challenging. The traditional approach in fMRI data analysis is to assume that underlying spatial regions of interest are static. In this article, using fractional amplitude of low-frequency fluctuations (fALFF) as an effective way to summarize the variability in fMRI data collected during a task, we arrange time-evolving fMRI data as a subjects by voxels by time windows tensor, and analyze the tensor using a tensor factorization-based approach called a PARAFAC2 model to reveal spatial dynamics. The PARAFAC2 model jointly analyzes data from multiple time windows revealing subject-mode patterns, evolving spatial regions (also referred to as networks) and temporal patterns. We compare the PARAFAC2 model with matrix factorization-based approaches relying on independent components, namely, joint independent component analysis (ICA) and independent vector analysis (IVA), commonly used in neuroimaging data analysis. We assess the performance of the methods in terms of capturing evolving networks through extensive numerical experiments demonstrating their modeling assumptions. In particular, we show that (i) PARAFAC2 provides a compact representation in all modes, i.e., subjects, time, and voxels, revealing temporal patterns as well as evolving spatial networks, (ii) joint ICA is as effective as PARAFAC2 in terms of revealing evolving networks but does not reveal temporal patterns, (iii) IVA's performance depends on sample size, data distribution and covariance structure of underlying networks. When these assumptions are satisfied, IVA is as accurate as the other methods, (iv) when subject-mode patterns differ from one time window to another, IVA is the most accurate. Furthermore, we analyze real fMRI data collected during a sensory motor task, and demonstrate that a component indicating statistically significant group difference between patients with schizophrenia and healthy controls is captured, which includes primary and secondary motor regions, cerebellum, and temporal lobe, revealing a meaningful spatial map and its temporal change.

Recent research has shown that it is possible to classify cognitive states of human subjects based on fMRI (functional magnetic resonance imaging) data. One of the obstacles in classifying fMRI data is the problem of high dimensionality. A single fMRI snapshot consists of thousands of voxels and since a single experiment contains many fMRI snapshots, the dimensionality of an fMRI data instance easily surpasses the order of tens of thousands. So, feature selection methods become a must from both classification and running time performance points of view. To this end several feature selection methods are studied, either general or specific to fMRI data. So far, one of the best such methods, which is specific to fMRI data, is called the method [9]. In this work we combine genetic algorithms with the active method in order to improve the performance of feature selection. Specifically, we first reduce the feature dimension using the active method and search for informative features in that reduced space using genetic algorithms. We achieve better or similar levels of classification performance using a much smaller number of voxels than the active method offers.

Independent component analysis (ICA) is a promising method that is increasingly used to analyze brain imaging data such as functional magnetic resonance imaging (fMRI), structural MRI, and electroencephalography and has also proved useful for group comparison, e.g., differentiating healthy controls from patients. An advantage of ICA is its ability to identify components that are mixed in an unknown manner. However, ICA is not necessarily robust and optimal in identifying between-group effects, especially in highly noisy situations. Here, we propose a modified ICA framework for multi-group data analysis that incorporates prior information regarding group membership as a constraint into the mixing coefficients. Our approach, called coefficient-constrained ICA (CC-ICA), prioritizes identification of components that show a significant group difference. The performance of CC-ICA via synthetic and hybrid data simulations is evaluated under different hypothesis testing assumptions and signal to noise ratios (SNRs). Group analysis is also conducted on real multitask fMRI data. Results show that CC-ICA improves the estimation accuracy of the independent components greatly, especially those that have different patterns for different groups (e.g., patients vs. controls); In addition, it enhances the data extraction sensitivity to group differences by ranking components with P value or J-divergence more consistently with the ground truth. The proposed algorithm performs quite well for both group-difference detection and multitask fMRI data fusion, which may prove especially important for the identification of relevant disease biomarkers.

Investigating the enhancement of template-free activation detection of event-related fMRI data using wavelet shrinkage and figures of merit

Inferring resting-state connectivity patterns from functional magnetic resonance imaging (fMRI) data is a challenging task for any analytical technique. In this paper, we review a probabilistic independent component analysis (PICA) approach, optimized for the analysis of fMRI data, and discuss the role which this exploratory technique can take in scientific investigations into the structure of these effects. We apply PICA to fMRI data acquired at rest, in order to characterize the spatio-temporal structure of such data, and demonstrate that this is an effective and robust tool for the identification of low-frequency resting-state patterns from data acquired at various different spatial and temporal resolutions. We show that these networks exhibit high spatial consistency across subjects and closely resemble discrete cortical functional networks such as visual cortical areas or sensory-motor cortex.