Temporally and spatially constrained ICA of fMRI data analysis

Three layered sparse dictionary learning algorithm for enhancing the subject wise segregation of brain networks

Independent component analysis (ICA) and dictionary learning (DL) are the most successful blind source separation (BSS) methods for functional magnetic resonance imaging (fMRI) data analysis. However, ICA to higher and DL to lower extent may suffer from performance degradation by the presence of anomalous observations in the recovered time courses (TCs) and high overlaps among spatial maps (SMs). This paper addressed both problems using a novel three-layered sparse DL (TLSDL) algorithm that incorporated prior information in the dictionary update process and recovered full-rank outlier-free TCs from highly corrupted measurements. The associated sequential DL model involved factorizing each subject's data into a multi-subject (MS) dictionary and MS sparse code while imposing a low-rank and a sparse matrix decomposition restriction on the dictionary matrix. It is derived by solving three layers of feature extraction and component estimation. The first and second layers captured brain regions with low and moderate spatial overlaps, respectively. The third layer that segregated regions with significant spatial overlaps solved a sequence of vector decomposition problems using the proximal alternating linearized minimization (PALM) method and solved a decomposition restriction using the alternating directions method (ALM). It learned outlier-free dynamics that integrate spatiotemporal diversities across brains and external information. It differs from existing DL methods owing to its unique optimization model, which incorporates prior knowledge, subject-wise/multi-subject representation matrices, and outlier handling. The TLSDL algorithm was compared with existing dictionary learning algorithms using experimental and synthetic fMRI datasets to verify its performance. Overall, the mean correlation value was found to be 26 % higher for the TLSDL than for the state-of-the-art subject-wise sequential DL (swsDL) technique.

Unraveling the Neural Landscape of Mental Disorders using Double Functional Independent Primitives (dFIPs)

Mental illnesses extract a high personal and societal cost, and thus explorations of the links between mental illness and functional connectivity in the brain are critical. Investigating major mental illnesses, believed to arise from disruptions in sophisticated neural connections, allows us to comprehend how these neural network disruptions may be linked to altered cognition, emotional regulation, and social interactions. Although neuroimaging has opened new avenues to explore neural alterations linked to mental illnesses, the field still requires precise and sensitive methodologies to inspect these neural substrates of various psychological disorders. In this study, we employ a hierarchical methodology to derive double functionally independent primitives (dFIPs) from resting state functional magnetic resonance neuroimaging data (rs-fMRI). These dFIPs encapsulate canonical overlapping patterns of functional network connectivity (FNC) within the brain. Our investigation focuses on the examination of how combinations of these dFIPs relate to different mental disorder diagnoses. The central aim is to unravel the complex patterns of FNC that correspond to the diverse manifestations of mental illnesses. To achieve this objective, we used a large brain imaging dataset from multiple sites, comprising 5805 total individuals diagnosed with schizophrenia (SCZ), autism spectrum disorder (ASD), bipolar disorder (BPD), major depressive disorder (MDD), and controls. The key revelations of our study unveil distinct patterns associated with each mental disorder through the combination of dFIPs. Notably, certain individual dFIPs exhibit disorder-specific characteristics, while others demonstrate commonalities across disorders. This approach offers a novel, data-driven synthesis of intricate neuroimaging data, thereby illuminating the functional changes intertwined with various mental illnesses. Our results show distinct signatures associated with psychiatric disorders, revealing unique connectivity patterns such as heightened cerebellar connectivity in SCZ and sensory domain hyperconnectivity in ASD, both contrasted with reduced cerebellar-subcortical connectivity. Utilizing the dFIP concept, we pinpoint specific functional connections that differentiate healthy controls from individuals with mental illness, underscoring its utility in identifying neurobiological markers. In summary, our findings delineate how dFIPs serve as unique fingerprints for different mental disorders.

A method for estimating dynamic functional network connectivity gradients (dFNG) from ICA captures smooth inter-network modulation

Dynamic functional network connectivity (dFNC) analysis is a widely used approach for studying brain function and offering insight into how brain networks evolve over time. Typically, dFNC studies utilized fixed spatial maps and evaluate transient changes in coupling among time courses estimated from independent component analysis (ICA). This manuscript presents a complementary approach that relaxes this assumption by spatially reordering the components dynamically at each timepoint to optimize for a smooth gradient in the FNC (i.e., a smooth gradient among ICA connectivity values). Several methods are presented to summarize dynamic FNC gradients (dFNGs) over time, starting with static FNC gradients (sFNGs), then exploring the reordering properties as well as the dynamics of the gradients themselves. We then apply this approach to a dataset of schizophrenia (SZ) patients and healthy controls (HC). Functional dysconnectivity between different brain regions has been reported in schizophrenia, yet the neural mechanisms behind it remain elusive. Using resting state fMRI and ICA on a dataset consisting of 151 schizophrenia patients and 160 age and gender-matched healthy controls, we extracted 53 intrinsic connectivity networks (ICNs) for each subject using a fully automated spatially constrained ICA approach. We develop several summaries of our functional network connectivity gradient analysis, both in a static sense, computed as the Pearson correlation coefficient between full time series, and a dynamic sense, computed using a sliding window approach followed by reordering based on the computed gradient, and evaluate group differences. Static connectivity analysis revealed significantly stronger connectivity between subcortical (SC), auditory (AUD) and visual (VIS) networks in patients, as well as hypoconnectivity in sensorimotor (SM) network relative to controls. sFNG analysis highlighted distinctive clustering patterns in patients and HCs along cognitive control (CC)/ default mode network (DMN), as well as SC/ AUD/ SM/ cerebellar (CB), and VIS gradients. Furthermore, we observed significant differences in the sFNGs between groups in SC and CB domains. dFNG analysis suggested that SZ patients spend significantly more time in a SC/ CB state based on the first gradient, while HCs favor the SM/DMN state. For the second gradient, however, patients exhibited significantly higher activity in CB domains, contrasting with HCs' DMN engagement. The gradient synchrony analysis conveyed more shifts between SM/ SC networks and transmodal CC/ DMN networks in patients. In addition, the dFNG coupling revealed distinct connectivity patterns between SC, SM and CB domains in SZ patients compared to HCs. To recap, our results advance our understanding of brain network modulation by examining smooth connectivity trajectories. This provides a more complete spatiotemporal summary of the data, contributing to the growing body of current literature regarding the functional dysconnectivity in schizophrenia patients. By employing dFNG, we highlight a new perspective to capture large scale fluctuations across the brain while maintaining the convenience of brain networks and low dimensional summary measures.

Generative Models of Brain Dynamics

This review article gives a high-level overview of the approaches across different scales of organization and levels of abstraction. The studies covered in this paper include fundamental models in computational neuroscience, nonlinear dynamics, data-driven methods, as well as emergent practices. While not all of these models span the intersection of neuroscience, AI, and system dynamics, all of them do or can work in tandem as generative models, which, as we argue, provide superior properties for the analysis of neuroscientific data. We discuss the limitations and unique dynamical traits of brain data and the complementary need for hypothesis- and data-driven modeling. By way of conclusion, we present several hybrid generative models from recent literature in scientific machine learning, which can be efficiently deployed to yield interpretable models of neural dynamics.

Enhanced design matrix for task-related fMRI data analysis

In this paper, we introduce a novel methodology for the analysis of task-related fMRI data. In particular, we propose an alternative way for constructing the design matrix, based on the newly suggested Information-Assisted Dictionary Learning (IADL) method. This technique offers an enhanced potential, within the conventional GLM framework, (a) to efficiently cope with uncertainties in the modeling of the hemodynamic response function, (b) to accommodate unmodeled brain-induced sources, beyond the task-related ones, as well as potential interfering scanner-induced artifacts, uncorrected head-motion residuals and other unmodeled physiological signals, and (c) to integrate external knowledge regarding the natural sparsity of the brain activity that is associated with both the experimental design and brain atlases. The capabilities of the proposed methodology are evaluated via a realistic synthetic fMRI-like dataset, and demonstrated using a test case of a challenging fMRI study, which verifies that the proposed approach produces substantially more consistent results compared to the standard design matrix method. A toolbox extension for SPM is also provided, to facilitate the use and reproducibility of the proposed methodology.

Brain Connectivity Studies on Structure-Function Relationships: A Short Survey with an Emphasis on Machine Learning

This short survey reviews the recent literature on the relationship between the brain structure and its functional dynamics. Imaging techniques such as diffusion tensor imaging (DTI) make it possible to reconstruct axonal fiber tracks and describe the structural connectivity (SC) between brain regions. By measuring fluctuations in neuronal activity, functional magnetic resonance imaging (fMRI) provides insights into the dynamics within this structural network. One key for a better understanding of brain mechanisms is to investigate how these fast dynamics emerge on a relatively stable structural backbone. So far, computational simulations and methods from graph theory have been mainly used for modeling this relationship. Machine learning techniques have already been established in neuroimaging for identifying functionally independent brain networks and classifying pathological brain states. This survey focuses on methods from machine learning, which contribute to our understanding of functional interactions between brain regions and their relation to the underlying anatomical substrate.

Temporally constrained ICA with threshold and its application to fMRI data

Background

Although independent component analysis (ICA) has been widely applied to functional magnetic resonance imaging (fMRI) data to reveal spatially independent brain networks, the order indetermination of ICA leads to the problem of target component selection. The temporally constrained independent component analysis (TCICA) is capable of automatically extracting the desired spatially independent components by adding the temporal prior information of the task to the mixing matrix for fMRI data analysis. However, the TCICA method can only extract a single component that tends to be a mix of multiple task-related components when there exist several independent components related to one task.

Methods

In this study, we proposed a TCICA with threshold (TCICA-Thres) method that performed TCICA outside the threshold and performed FastICA inside the threshold to automatically extract all the target components related to one task. The proposed approach was tested using simulated fMRI data and was applied to a real fMRI experiment using 13 subjects. Additionally, the performance of TCICA-Thres was compared with that of FastICA and TCICA.

Results

The results from the simulation and the fMRI data demonstrated that TCICA-Thres better extracted the task-related components than TCICA. Moreover, TCICA-Thres outperformed FastICA in robustness to noise, spatial detection power and computational time.

Conclusions

The proposed TCICA-Thres solves the limitations of TCICA and extends the application of TCICA in fMRI data analysis.

Electronic supplementary material

The online version of this article (10.1186/s12880-018-0300-6) contains supplementary material, which is available to authorized users.

A New Constrained Spatiotemporal ICA Method Based on Multi-Objective Optimization for fMRI Data Analysis

Compared with independent component analysis (ICA), constrained ICA (CICA) has unique advantages in functional magnetic resonance image (fMRI) data analysis by introducing some priori information into the estimation process. However, there are still some controversies in the current CICA methods, such as how to choose the threshold parameter to restrain the similarity, and how to reduce the accuracy requirements for a priori information. In this paper, we propose a new constrained spatiotemporal ICA (CSTICA) method based on the framework of multi-objective optimization, where the inequality constraint of the traditional CICA method is transformed into the objective optimization function of the CSTICA, and both temporal and spatial a priori information are included simultaneously. The simulated, hybrid, and real fMRI data experiments are designed to evaluate the performance of the proposed CSTICA method in comparison with the classical ICA and CICA methods. Compared with the traditional CICA methods, the CSTICA has circumvented the problem of threshold parameter selection. Furthermore, the experimental results demonstrate that the source recovery ability of the CSTICA has been improved to a certain extent especially in the cases of a priori information with low accuracies. Meanwhile, the results also indicate that the CSTICA reduces dependency on the accuracy of a priori information.

Recovering task fMRI signals from highly under-sampled data with low-rank and temporal subspace constraints

Recent developments in highly accelerated fMRI data acquisition have employed low-rank and/or sparsity constraints for image reconstruction, as an alternative to conventional, time-independent parallel imaging. When under-sampling factors are high or the signals of interest are low-variance, however, functional data recovery can be poor or incomplete. We introduce a method for improving reconstruction fidelity using external constraints, like an experimental design matrix, to partially orient the estimated fMRI temporal subspace. Combining these external constraints with low-rank constraints introduces a new image reconstruction model that is analogous to using a mixture of subspace-decomposition (PCA/ICA) and regression (GLM) models in fMRI analysis. We show that this approach improves fMRI reconstruction quality in simulations and experimental data, focusing on the model problem of detecting subtle 1-s latency shifts between brain regions in a block-design task-fMRI experiment. Successful latency discrimination is shown at acceleration factors up to R = 16 in a radial-Cartesian acquisition. We show that this approach works with approximate, or not perfectly informative constraints, where the derived benefit is commensurate with the information content contained in the constraints. The proposed method extends low-rank approximation methods for under-sampled fMRI data acquisition by leveraging knowledge of expected task-based variance in the data, enabling improvements in the speed and efficiency of fMRI data acquisition without the loss of subtle features.

A Novel Feature-Map Based ICA Model for Identifying the Individual, Intra/Inter-Group Brain Networks across Multiple fMRI Datasets

Independent component analysis (ICA) has been widely used in functional magnetic resonance imaging (fMRI) data analysis to evaluate functional connectivity of the brain; however, there are still some limitations on ICA simultaneously handling neuroimaging datasets with diverse acquisition parameters, e.g., different repetition time, different scanner, etc. Therefore, it is difficult for the traditional ICA framework to effectively handle ever-increasingly big neuroimaging datasets. In this research, a novel feature-map based ICA framework (FMICA) was proposed to address the aforementioned deficiencies, which aimed at exploring brain functional networks (BFNs) at different scales, e.g., the first level (individual subject level), second level (intragroup level of subjects within a certain dataset) and third level (intergroup level of subjects across different datasets), based only on the feature maps extracted from the fMRI datasets. The FMICA was presented as a hierarchical framework, which effectively made ICA and constrained ICA as a whole to identify the BFNs from the feature maps. The simulated and real experimental results demonstrated that FMICA had the excellent ability to identify the intergroup BFNs and to characterize subject-specific and group-specific difference of BFNs from the independent component feature maps, which sharply reduced the size of fMRI datasets. Compared with traditional ICAs, FMICA as a more generalized framework could efficiently and simultaneously identify the variant BFNs at the subject-specific, intragroup, intragroup-specific and intergroup levels, implying that FMICA was able to handle big neuroimaging datasets in neuroscience research.

An improved multi-objective optimization-based CICA method with data-driver temporal reference for group fMRI data analysis

Group independent component analysis (GICA) has been successfully applied to study multi-subject functional magnetic resonance imaging (fMRI) data, and the group independent component (GIC) represents the commonality of all subjects in the group. However, some studies show that the performance of GICA can be improved by incorporating a priori information, which is not always considered when looking for GICs in existing GICA methods. In this paper, we propose an improved multi-objective optimization-based constrained independent component analysis (CICA) method to take advantage of the temporal a priori information extracted from all subjects in the group by incorporating it into the computational process of GICA for group fMRI data analysis. The experimental results of simulated and real data show that the activated regions and the time course detected by the improved CICA method are more accurate in some sense. Moreover, the GIC computed by the improved CICA method has a higher correlation with the corresponding independent component of each subject in the group, which means that the improved CICA method with the temporal a priori information extracted from the group can better reflect the commonality of the subjects. These results demonstrate that the improved CICA method has its own advantages in fMRI data analysis.

SCGICAR: Spatial concatenation based group ICA with reference for fMRI data analysis

BACKGROUND AND OBJECTIVE

With the rapid development of big data, the functional magnetic resonance imaging (fMRI) data analysis of multi-subject is becoming more and more important. As a kind of blind source separation technique, group independent component analysis (GICA) has been widely applied for the multi-subject fMRI data analysis. However, spatial concatenated GICA is rarely used compared with temporal concatenated GICA due to its disadvantages.

METHODS

In this paper, in order to overcome these issues and to consider that the ability of GICA for fMRI data analysis can be improved by adding a priori information, we propose a novel spatial concatenation based GICA with reference (SCGICAR) method to take advantage of the priori information extracted from the group subjects, and then the multi-objective optimization strategy is used to implement this method. Finally, the post-processing means of principal component analysis and anti-reconstruction are used to obtain group spatial component and individual temporal component in the group, respectively.

RESULTS

The experimental results show that the proposed SCGICAR method has a better performance on both single-subject and multi-subject fMRI data analysis compared with classical methods. It not only can detect more accurate spatial and temporal component for each subject of the group, but also can obtain a better group component on both temporal and spatial domains.

CONCLUSIONS

These results demonstrate that the proposed SCGICAR method has its own advantages in comparison with classical methods, and it can better reflect the commonness of subjects in the group.

A novel fMRI group data analysis method based on data-driven reference extracting from group subjects

Group-independent component analysis (GICA) is a well-established blind source separation technique that has been widely applied to study multi-subject functional magnetic resonance imaging (fMRI) data. The group-independent components (GICs) represent the commonness of all of the subjects in the group. Similar to independent component analysis on the single-subject level, the performance of GICA can be improved for multi-subject fMRI data analysis by incorporating a priori information; however, a priori information is not always considered while looking for GICs in existing GICA methods, especially when no obvious or specific knowledge about an unknown group is available. In this paper, we present a novel method to extract the group intrinsic reference from all of the subjects of the group and then incorporate it into the GICA extraction procedure. Comparison experiments between FastICA and GICA with intrinsic reference (GICA-IR) are implemented on the group level with regard to the simulated, hybrid and real fMRI data. The experimental results show that the GICs computed by GICA-IR have a higher correlation with the corresponding independent component of each subject in the group, and the accuracy of activation regions detected by GICA-IR was also improved. These results have demonstrated the advantages of the GICA-IR method, which can better reflect the commonness of the subjects in the group.

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.

Spatial uncertainty modeling of fuzzy information in images for pattern classification

The modeling of the spatial distribution of image properties is important for many pattern recognition problems in science and engineering. Mathematical methods are needed to quantify the variability of this spatial distribution based on which a decision of classification can be made in an optimal sense. However, image properties are often subject to uncertainty due to both incomplete and imprecise information. This paper presents an integrated approach for estimating the spatial uncertainty of vagueness in images using the theory of geostatistics and the calculus of probability measures of fuzzy events. Such a model for the quantification of spatial uncertainty is utilized as a new image feature extraction method, based on which classifiers can be trained to perform the task of pattern recognition. Applications of the proposed algorithm to the classification of various types of image data suggest the usefulness of the proposed uncertainty modeling technique for texture feature extraction.