Linking functional and structural brain organisation with behaviour in healthy adults

Area of Science:

• Neuroimaging
• Cognitive Neuroscience
• Data Science

Background:

• Multimodal data integration enhances sensitivity to brain-behavior relationships.
• Investigating cross-modal organizational patterns can reveal new insights.
• Linked Independent Component Analysis (LICA) is a method for integrating diverse data types.

Purpose of the Study:

• To determine if organizational patterns persist across structural and functional brain imaging modalities.
• To assess if multimodal integration increases sensitivity to brain-behavior associations.
• To explore the relationship between inter-regional functional and structural organization.

Main Methods:

• Utilized multimodal magnetic resonance imaging (T1w, resting-state fMRI, DWI) and behavioral data from the Human Connectome Project (n=676).
• Extracted unimodal features including grey matter density, tractography connectivity, and connectopic maps.
• Applied LICA to integrate features and examined associations between resulting components and demographic/behavioral variables (n=308).

Main Results:

• 15 out of 100 independent components derived from LICA showed significant associations with demographic/behavioral measures.
• Two components linked to intoxication were driven by diffusion-weighted imaging (DWI) data.
• One component, driven by striatal functional MRI (fMRI) data, related to working memory.
• Few components exhibited shared variance between structural and functional data, with none showing significant behavioral associations.

Conclusions:

• fMRI connectopic mapping shows promise for future working memory research.
• The current study questions the utility of integrating connectopic maps and tractography data due to a lack of behaviorally relevant shared variance.
• Multimodal integration using LICA can identify specific brain-behavior relationships, particularly when driven by a single modality.