Fusion of Multi-Modal Neuroimaging Data and Association With Cognitive Data

Thumbnail Image

Files

2023107689.pdf (515 KB)

Links to Files

https://ieeexplore.ieee.org/document/10193147

Permanent Link

https://doi.org/10.1109/ICASSPW59220.2023.10193147
 http://hdl.handle.net/11603/29457

Collections

UMBC Computer Science and Electrical Engineering Department
UMBC Faculty Collection
UMBC Student Collection

Author/Creator

Author/Creator ORCID

https://orcid.org/0000-0001-9035-4555
 https://orcid.org/0000-0003-0826-453X
 https://orcid.org/0000-0003-0594-2796

Date

2023-08-02

Type of Work

conference papers and proceedings
postprints
Text

Department

Program

Citation of Original Publication

M. D. LoPresto, M. A. B. S. Akhonda, V. D. Calhoun and T. Adali, "Fusion of Multi-Modal Neuroimaging Data and Association With Cognitive Data," 2023 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops (ICASSPW), Rhodes Island, Greece, 2023, pp. 1-5, doi: 10.1109/ICASSPW59220.2023.10193147.

Rights

© 2023 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

Subjects

Abstract

Efforts to develop a deeper understanding of the human brain benefit from the joint analysis (fusion) of multiple sources of data to exploit the complementary information in each modality. A number of Blind Source Separation (BSS) techniques have been developed for data-driven analysis and fusion of neuroimaging data, but only recently have fusion frameworks emerged that consider cognitive data alongside the neuroimaging data. In our approach, we first apply transposed Independent Vector Analysis (tIVA) across three neuroimaging modalities to extract subject covariations that are then associated with cognitive data. The tIVA step allows full interaction of the neuroimaging modalities to discover cross-modality relationships. We cluster the data based on the subject covariations to find the cognitive scores that offer significant discrimination between each pair of clusters. Our approach allows full interaction of multiple neuroimaging modalities and makes direct association with the cognitive data, identifying the Brief Assessment of Cognition in Schizophrenia Composite Score (BACS CS), Hopkins Verbal Learning Test (HVLT) and the Neuropsychological Assessment Battery (NAB) Mazes Test as being strongly associated with the cross-modality connections.