Robust data-driven segmentation of pulsatile cerebral vessels using functional magnetic resonance imaging.

Area of Science:

• Neuroimaging
• Biomedical Engineering
• Fluid Dynamics

Background:

• Functional magnetic resonance imaging (fMRI) offers insights into neurovascular health but lacks detailed cerebral vessel segmentation.
• Accurate segmentation of cerebral arteries and the superior sagittal sinus (SSS) is crucial for studying neurofluid dynamics.
• Current segmentation methods are often time-consuming and require specialized magnetic resonance angiography (MRA) scans.

Purpose of the Study:

• To develop and validate a data-driven, automatic method for segmenting large cerebral arteries and the SSS directly from fMRI data.
• To assess the reproducibility of this automated segmentation method across a large, aging cohort.

Main Methods:

• Leveraged cardiac-induced pulsatile fMRI signals for automated segmentation.
• Developed a data-driven approach to identify large cerebral arteries and the SSS.
• Validated the method against ground truth segmentations and tested reproducibility on the Human Connectome Project (HCP) aging dataset (422 participants).

Main Results:

• The automated method successfully segmented large cerebral arteries and the SSS in fMRI data.
• High reproducibility was demonstrated, with intraclass correlation coefficients > 0.7 for both artery and SSS segmentation volumes.
• The method proved reliable across a wide age range (36-100 years) and repeated scans.

Conclusions:

• Automated segmentation of large cerebral arteries and the SSS is feasible and reproducible using fMRI.
• This technique facilitates non-invasive investigation of neurofluid dynamics and vascular health.
• The method reduces reliance on MRA, making advanced fMRI analysis more accessible.