Robust subject-wise dictionary learning for fMRI

• Published in: IEEE access 2024-01, Vol.12, p.1-1
• Main Authors: Khalid, Muhammad Usman, Albahlal, Bader M.
• Format: Article
• Language: English
• Subjects: Algorithms; Blind source separation; Computational modeling; Correlation; Data models; Decomposition; Dictionaries; dictionary learning; Divergence; fMRI; Functional magnetic resonance imaging; group-level analysis; Machine learning; Magnetic resonance imaging; Matrix decomposition; Optimization; Outliers (statistics); Random noise; Representation learning; representation matrices; Robust subject-wise analysis; Robustness; Signal integrity; signal intensity maximization; Smoothness; Sparse matrices; sparse representation; Spatiotemporal phenomena; α-divergence
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2024.3373437

This paper presents a robust subject-wise sequential dictionary learning (swsDL) algorithm named rswsDL for functional magnetic resonance imaging (fMRI) data where the negative impact of dimensionality reduction in the form of information loss and sensitivity to anomalous observations due to the assumption of Gaussian prior has been resolved by rotating the reduced dimensions to optimal direction and replacing the quadratic loss in the data fidelity term with α-divergence based loss function to counter the outliers, respectively. While dimensionality optimization guaranteed robustness to model order by maximizing signal intensity and smoothness, the robust loss function guaranteed decomposition stability against deviations from the Gaussian noise model. The proposed algorithm was derived by deploying spatial and temporal bases from the computationally fast sparse spatiotemporal blind source separation (ssBSS) method and solving a sequence of rank-1 matrix decomposition problems, where the l 1 / l 0 -norm penalty/constraint promoted sparsity, and the estimation of sparse representation matrices was accomplished using a block coordinate descent approach. This strategy allowed the utilization of multi-subject fMRI data to enhance the subject-wise source separation in a robust manner. It, therefore, can be considered a promising alternative to the state-of-the-art robust consistent adaptive sequential dictionary learning (rACSD) algorithm. The rswsDL and existing robust dictionary learning based source separation algorithms were applied to synthetic and experimental fMRI datasets to validate its performance. The rswsDL algorithm manifested a 16.7% increase in the mean correlation value over the rACSD algorithm.