Efficient Blind Source Separation Method for fMRI Using Autoencoder and Spatiotemporal Sparsity Constraints
Diversity measures exploited by blind source separation (BSS) methods are usually based on either statistical attributes/geometrical structures or sparse/overcomplete (underdetermined) representations of the signals. This leads to some inefficient BSS methods that are derived from either a mixing ma...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2023-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10128979/ |
| _version_ | 1797814582239035392 |
|---|---|
| author | Muhammad Usman Khalid Bilal A. Khawaja Malik Muhammad Nauman |
| author_facet | Muhammad Usman Khalid Bilal A. Khawaja Malik Muhammad Nauman |
| author_sort | Muhammad Usman Khalid |
| collection | DOAJ |
| description | Diversity measures exploited by blind source separation (BSS) methods are usually based on either statistical attributes/geometrical structures or sparse/overcomplete (underdetermined) representations of the signals. This leads to some inefficient BSS methods that are derived from either a mixing matrix (mm), sparse weight vectors (sw), or sparse code (sc). In contrast, the proposed efficient method, sparse spatiotemporal BSS (ssBSS), avoids computational complications associated with lag sets, deflation strategy, and repeated error matrix computation using the whole dataset. It solves the spatiotemporal data reconstruction model (STEM) with <inline-formula> <tex-math notation="LaTeX">$l_{1}$ </tex-math></inline-formula>-norm penalization and <inline-formula> <tex-math notation="LaTeX">$l_{0}$ </tex-math></inline-formula>-norm constraints using Neumann’s alternating projection lemma and block coordinate descent approach to yield the desired bases. Its specific solution allows incorporating a three-step autoencoder and univariate soft thresholding for a block update of the source/mixing matrices. Due to the utilization of both spatial and temporal information, it can better distinguish between the sources and yield interpretable results. These steps also make ssBSS unique because, to the best of my knowledge, no mixing matrix based BSS method incorporates sparsity of both features and a multilayer network structure. The proposed method is validated using synthetic and various functional magnetic resonance imaging (fMRI) datasets. Results reveal the superior performance of the proposed ssBSS method compared to the existing methods based on mmBSS and swBSS. Specifically, overall, a 14% increase in the mean correlation value and 91% reduction in computation time over the ssICA algorithm was discovered. |
| first_indexed | 2024-03-13T08:09:47Z |
| format | Article |
| id | doaj.art-be195276b34a4dc4b720f08ff5d21c51 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-03-13T08:09:47Z |
| publishDate | 2023-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-be195276b34a4dc4b720f08ff5d21c512023-05-31T23:00:29ZengIEEEIEEE Access2169-35362023-01-0111503645038110.1109/ACCESS.2023.327754310128979Efficient Blind Source Separation Method for fMRI Using Autoencoder and Spatiotemporal Sparsity ConstraintsMuhammad Usman Khalid0Bilal A. Khawaja1https://orcid.org/0000-0003-1537-5502Malik Muhammad Nauman2https://orcid.org/0000-0003-3489-6792College of Computer and Information Sciences, Imam Mohammad Ibn Saud Islamic University, Riyadh, Saudi ArabiaDepartment of Electrical Engineering, Faculty of Engineering, Islamic University of Madinah, Madinah, Saudi ArabiaFaculty of Integrated Technologies, Universiti Brunei Darussalam, Bandar Seri Begawan, BruneiDiversity measures exploited by blind source separation (BSS) methods are usually based on either statistical attributes/geometrical structures or sparse/overcomplete (underdetermined) representations of the signals. This leads to some inefficient BSS methods that are derived from either a mixing matrix (mm), sparse weight vectors (sw), or sparse code (sc). In contrast, the proposed efficient method, sparse spatiotemporal BSS (ssBSS), avoids computational complications associated with lag sets, deflation strategy, and repeated error matrix computation using the whole dataset. It solves the spatiotemporal data reconstruction model (STEM) with <inline-formula> <tex-math notation="LaTeX">$l_{1}$ </tex-math></inline-formula>-norm penalization and <inline-formula> <tex-math notation="LaTeX">$l_{0}$ </tex-math></inline-formula>-norm constraints using Neumann’s alternating projection lemma and block coordinate descent approach to yield the desired bases. Its specific solution allows incorporating a three-step autoencoder and univariate soft thresholding for a block update of the source/mixing matrices. Due to the utilization of both spatial and temporal information, it can better distinguish between the sources and yield interpretable results. These steps also make ssBSS unique because, to the best of my knowledge, no mixing matrix based BSS method incorporates sparsity of both features and a multilayer network structure. The proposed method is validated using synthetic and various functional magnetic resonance imaging (fMRI) datasets. Results reveal the superior performance of the proposed ssBSS method compared to the existing methods based on mmBSS and swBSS. Specifically, overall, a 14% increase in the mean correlation value and 91% reduction in computation time over the ssICA algorithm was discovered.https://ieeexplore.ieee.org/document/10128979/Sparse representationautoencoderBSSfMRIactivation mapsPCA |
| spellingShingle | Muhammad Usman Khalid Bilal A. Khawaja Malik Muhammad Nauman Efficient Blind Source Separation Method for fMRI Using Autoencoder and Spatiotemporal Sparsity Constraints IEEE Access Sparse representation autoencoder BSS fMRI activation maps PCA |
| title | Efficient Blind Source Separation Method for fMRI Using Autoencoder and Spatiotemporal Sparsity Constraints |
| title_full | Efficient Blind Source Separation Method for fMRI Using Autoencoder and Spatiotemporal Sparsity Constraints |
| title_fullStr | Efficient Blind Source Separation Method for fMRI Using Autoencoder and Spatiotemporal Sparsity Constraints |
| title_full_unstemmed | Efficient Blind Source Separation Method for fMRI Using Autoencoder and Spatiotemporal Sparsity Constraints |
| title_short | Efficient Blind Source Separation Method for fMRI Using Autoencoder and Spatiotemporal Sparsity Constraints |
| title_sort | efficient blind source separation method for fmri using autoencoder and spatiotemporal sparsity constraints |
| topic | Sparse representation autoencoder BSS fMRI activation maps PCA |
| url | https://ieeexplore.ieee.org/document/10128979/ |
| work_keys_str_mv | AT muhammadusmankhalid efficientblindsourceseparationmethodforfmriusingautoencoderandspatiotemporalsparsityconstraints AT bilalakhawaja efficientblindsourceseparationmethodforfmriusingautoencoderandspatiotemporalsparsityconstraints AT malikmuhammadnauman efficientblindsourceseparationmethodforfmriusingautoencoderandspatiotemporalsparsityconstraints |