Multi-subject Task-related fMRI Data Analysis via Generalized Canonical Correlation Analysis^.

Area of Science:

• Neuroimaging
• Statistical analysis
• Brain activity mapping

Background:

• Functional magnetic resonance imaging (fMRI) is a key technique for brain study, measuring Blood Oxygen Level Dependent (BOLD) signals.
• Unsupervised multivariate statistical methods are increasingly used for fMRI data analysis to extract information without prior experimental knowledge.
• Generalized canonical correlation analysis (gCCA) estimates common linear subspaces within multiple random linear subspaces.

Purpose of the Study:

• To propose a new fMRI data generating model accounting for common task-related and resting-state components.
• To estimate the common spatial task-related component using a two-stage gCCA approach.
• To validate the proposed model and method with real-world fMRI data.

Main Methods:

• Development of a novel fMRI data generating model incorporating common task-related and resting-state components.
• Application of a two-stage generalized canonical correlation analysis (gCCA) for estimating common spatial task-related components.
• Validation using real-world, multi-subject task-related fMRI datasets.

Main Results:

• The proposed model effectively accounts for commonalities in fMRI data.
• The two-stage gCCA successfully estimates the common spatial task-related component.
• Experimental findings corroborate the theoretical results, demonstrating the method's efficacy.

Conclusions:

• The developed approach is highly suitable for multi-subject, task-related fMRI data processing.
• The methods amplify and recover commonalities across subjects in fMRI experiments.
• This work advances the analysis of complex brain activity patterns in group studies.