Tensorial independent component analysis of multi-echo multiband magnetic resonance imaging improves data quality and identification of altered networks in depression
Year: 2025
Type: dissertation
Study design: Observational Study
Abstract: Multi-echo echo-planar imaging (ME-EPI) acquires images at multiple echo times (TEs), enabling the differentiation of blood oxygen level-dependent (BOLD) and non-BOLD fluctuations through TE-dependent analysis of transverse relaxation time and initial intensity. Decomposing ME-EPI in tensor space is a promising approach to characterize the distribution of changes across TEs (TE patterns) directly and aid the classification of components by providing information from an additional domain. In this study, we decomposed multiband ME-EPI data using the tensorial extension of independent component analysis (tensor-ICA) and summarized the characteristics of different types of independent components (ICs) according to their TE patterns. The findings support the feasibility of using tensor-ICA to decompose ME-EPI data and its usefulness for simplifying the distinction of components. Based on differences in the TE patterns (decreased, short-TE-peak, and long-TE-peak patterns), ICs could be assigned to one of three groups. By removing the identified noise ICs, quality control metrics were significantly improved, and stronger and more concise BOLD responses were observed. Further, the present study demonstrates that extracting physiologically meaningful independent temporal components based on BOLD tensor ICs from ME-EPI data offers a valuable approach to delineating perceptual alterations in mental disorders such as major depressive disorder (MDD). Thirty-nine individuals diagnosed with MDD and 36 healthy controls, matched for age and gender ratio, finished the scans under naturalistic stimuli (German film "Lola rennt") and were included to characterize the BOLD signal of perceptual processing. Two components associated with auditory and visual features exhibited significant group differences. These deficits primarily localize to higher-order processing units such as the prefrontal cortex rather than primary sensory cortices, thus confirming its pivotal role in the pathophysiology of MDD. Future studies should address how therapeutic interventions targeting the prefrontal cortex affect the brain network changes identified in this study, as well as their contribution to the overall treatment response. This study suggests that tensor-ICA can provide useful information for classifying ME-EPI ICs, yield efficient denoising of fMRI data, and promote deep analysis like temporal response to naturalistic stimuli.
Source: RWTH Publications (RWTH Aachen)
Author: Feng, Tengfei
Institution: RWTH Aachen University
Language: English
Subfield: Radiology, Nuclear Medicine and Imaging
Field: Medicine
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W7119519237
Cites:
Cited by:
Related to: