Robust data-driven segmentation of pulsatile cerebral vessels using functional magnetic resonance imaging.
Adam M WrightTianyin XuJacob IngramJohn KooYi ZhaoYunjie TongQiuting WenPublished in: bioRxiv : the preprint server for biology (2024)
Functional magnetic resonance imaging (fMRI) captures rich physiological and neuronal information that can offer insights into neurofluid dynamics, vascular health, and waste clearance function. The availability of cerebral vessel segmentation could facilitate fluid dynamics research in fMRI. However, without magnetic resonance angiography scans, cerebral vessel segmentation is challenging and time-consuming. This study leverages cardiac-induced pulsatile fMRI signal to develop a data-driven, automatic segmentation of large cerebral arteries and the superior sagittal sinus (SSS). The method was validated in a local dataset by comparing it to ground truth cerebral artery and SSS segmentations. Using the Human Connectome Project (HCP) aging dataset, the method's reproducibility was tested on 422 participants aged 36 to 100 years, each with four repeated fMRI scans. The method demonstrated high reproducibility, with an intraclass correlation coefficient > 0.7 in both cerebral artery and SSS segmentation volumes. This study demonstrates that the large cerebral arteries and SSS can be reproducibly and automatically segmented in fMRI datasets, facilitating the investigation of fluid dynamics in these regions.
Keyphrases
- resting state
- subarachnoid hemorrhage
- deep learning
- functional connectivity
- magnetic resonance imaging
- computed tomography
- convolutional neural network
- magnetic resonance
- cerebral ischemia
- contrast enhanced
- healthcare
- brain injury
- public health
- left ventricular
- optical coherence tomography
- machine learning
- health information
- mental health
- cerebral blood flow
- diffusion weighted imaging
- quality improvement
- single cell
- risk assessment
- rna seq
- oxidative stress
- blood brain barrier
- sewage sludge