Dictionary Learning-Based fMRI Data Analysis for Capturing Common and Individual Neural Activation Maps

Jin, Rui; Dontaraju, Krishna K.; Kim, Seung-Jun; Akhonda, M. A. B. S.

doi:10.1109/jstsp.2020.2992430

Cited by 11 publications

(8 citation statements)

References 64 publications

(71 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Nevertheless, ICA's assumption of independence was again challenged in [20] where ICA, compared to sparse DL, faced difficulty in retrieving neural dynamics when moderate to significant overlaps among functional networks were present. A similar trend was encountered in [21] where ICA could not reveal the activation maps that DL discovered. The sparse assumption has also been supported by biological evidence of sparse coding in the brain [22], and highlighted in an earlier ICA study [23].…”

Section: Introductionsupporting

confidence: 68%

“…These algorithms might suffer from convergence issues because they are based on alternating minimization (AM) approach that updates dictionary and sparse code separately and performance issues because they ignore fMRI data's prior information such as temporal smoothness and spatial dependence among neighbouring voxels. Moreover, recently, an MS-DL algorithm was applied to spatial features to reveal SMs that were common among healthy controls and schizophrenic subjects and those that were specific to each group [21], and more recently, MS-fMRI data were decomposed using the Tucker-2 model into shared, and individual TCs/SMs [43].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Sparse Group Bases for Multisubject fMRI Data

Khalid

2022

IEEE Access

View full text Add to dashboard Cite

Considering that functional magnetic resonance imaging (fMRI) signals from multiple subjects (MS) can be represented together as a sum of common and a sum of distinct rank-1 matrices, a new MS dictionary learning (DL) algorithm named sparse group (common + distinct) bases (sgBACES) is proposed. Unlike existing MS-DL algorithms that ignore fMRI data's prior information, it is formulated as a penalized plus constrained rank-1 matrix approximation, where l 1 norm-based adaptive sparse penalty, l 0 norm-based dictionary regularization, and lag-1 based autocorrelation maximization have been introduced in the minimization problem. Besides, spatial dependence among voxels has been exploited for fine-tuning the sparsity parameters. To my best knowledge, the sgBACES algorithm is the first to effectively take both temporal and spatial prior information into account for an MS-fMRI-DL framework. It also has the advantage of not requiring a separate sparse coding stage. Studies based on synthetic and experimental fMRI datasets are used to compare the performance of sgBACES with the state-of-the-art algorithms in terms of correlation strength and computation time. It emerged that the proposed sgBACES algorithm enhanced the signal-to-noise ratio (SNR) of the recovered time courses (TCs) and the precision of the recovered spatial maps (SMs). A 9.2% increase in correlation value over the ShSSDL algorithm is observed for motor-task based fMRI data.

show abstract

Section: Introductionsupporting

confidence: 68%

Section: Introductionmentioning

confidence: 99%

Sparse Group Bases for Multisubject fMRI Data

Khalid

2022

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Along with brain imaging data, behavioral scores have been used recently to study the aberrant behavior in SZ [28], [36]- [37]. A total of 51 behavioral variables (BVs) were available for 253 subjects.…”

Section: Behavioral Scoresmentioning

confidence: 99%

Data-driven spatio-temporal dynamic brain connectivity analysis using fALFF: Application to sensorimotor task data

Hossain

Bhinge

Long

et al. 2022

2022 56th Annual Conference on Information Sciences and Systems (CISS)

Self Cite

View full text Add to dashboard Cite

Dynamic functional connectivity (dFC) analysis enables us to capture the time-varying interactions between brain regions and can lead to powerful biomarkers. Most dFC studies are focused on the study of temporal dynamics and require significant postprocessing to summarize the results of the dynamics analysis. In this paper, we introduce an effective framework that makes use of independent vector analysis (IVA) with fractional amplitude of low frequency fluctuation (fALFF) features extracted from task functional magnetic resonance imaging (fMRI) data. Our approach, which is based on IVA with fALLF features as input, (IVA-fALLF) produces an effective summary of the dynamics also greatly facilitating the study of both spatial and temporal dynamics in a more concise manner. IVA-fALLF captures the spatial and temporal dynamics of sensorimotor task data and identifies a component with significant difference in dynamic behavior between healthy controls (HC) and patients with schizophrenia (SZ). Finally, our post analysis using behavioral scores finds significant correlation between brain imaging data and the associated behavioral scores, increasing confidence on our results. Our results are consistent with the previous data-driven dFC analysis as we find similar brain networks showing abnormal behavior in patients with SZ. Moreover, our analysis identifies component behavior in task and rest windows separately and provides additional confirmation of results through correlation with behavioral scores.

show abstract

“…The Fisher discrimination criterion is to cluster the samples in the same modality and keep the samples in different modalities as far away from each other as possible, which helps to extract features corresponding to the specific modality [22]- [24]. Assume that the multimodal data X = (x 1 , x 2 , • • • , x N ) ∈ R p×N contains M modalities with N m samples belonging to the m-th modality N m and ∑ M m=1 N m = N , where pdimensional vector x n is the n-th sample of X.…”

Section: B Fisher Costmentioning

confidence: 99%

Explainable Multimodal Deep Dictionary Learning to Capture Developmental Differences From Three fMRI Paradigms

Yang

Qiao

Zhou

et al. 2023

IEEE Trans. Biomed. Eng.

View full text Add to dashboard Cite

Objective: Multimodal-based methods show great potential for neuroscience studies by integrating complementary information. There has been less multimodal work focussed on brain developmental changes. Methods: We propose an explainable multimodal deep dictionary learning method to uncover both the commonality and specificity of different modalities, which learns the shared dictionary and the modality-specific sparse representations based on the multimodal data and their encodings of a sparse deep autoencoder. Results: By regarding three fMRI paradigms collected during two tasks and resting state as modalities, we apply the proposed method on multimodal data to identify the brain developmental differences. We found that both children and young adults prefer to switch among states during two tasks while staying within a particular state during rest, but the difference is that children possess more diffuse functional connectivity patterns while young adults have more focused functional connectivity patterns. Conclusion and Significance: To uncover the commonality and specificity of three fMRI paradigms to developmental differences, multimodal data and their encodings are used to train the shared dictionary and the modality-specific sparse representations. Identifying brain network differences helps to understand how the neural circuits and brain networks form and develop with age.

show abstract

Dictionary Learning-Based fMRI Data Analysis for Capturing Common and Individual Neural Activation Maps

Abstract: County (UMBC) ScholarWorks@UMBC digital repository on the Maryland Shared Open Access (MD-SOAR) platform.

Cited by 11 publications

References 64 publications

Sparse Group Bases for Multisubject fMRI Data

Sparse Group Bases for Multisubject fMRI Data

Data-driven spatio-temporal dynamic brain connectivity analysis using fALFF: Application to sensorimotor task data

Explainable Multimodal Deep Dictionary Learning to Capture Developmental Differences From Three fMRI Paradigms

Contact Info

Product

Resources

About