Manual segmentation of the human claustrum in high resolution MRI

Nicholas R. Prunier, BA

Brigham and Women’s Hospital
Manual segmentation of the human claustrum in high resolution MRI

Scientific Abstract

Background: The claustrum is a deep forebrain structure characterized by a thin, sheet-like layer of neurons highly connected to different cortical and subcortical areas. Although its structure and function are poorly understood, recent studies suggest that the widespread connectivity of the claustrum may reflect a role in multimodal perceptual binding, or consciousness. Accurate and reliable segmentation of the claustrum is important for morphological, fMRI, and diffusion MRI tractography analyses. However, morphology of the claustrum in standard T1 weighted MRI datasets has been difficult to analyze because of its small size, its indistinct border with adjacent brain structures, and the resolution of standard structural MRI datasets. The goal of the present project is to generate a reliable method for segmenting the claustrum in high resolution MRI datasets.

Methods: High resolution (0.7mm isotropic) structural MRI data will be obtained from 10 randomly selected subjects in the HCP Young Adults dataset. Manual segmentation of the claustrum will be performed using the Neurosegmentation Module of 3D Slicer v4.10 using histogram-based segmentation methods, as utilized in the Center for Morphometric Analysis approach to brain segmentation. Each claustrum will be segmented twice by two different analysts. The Dice coefficient will be used to analyze the degree of agreement within and between observers.

Results: We expect to generate a reliable method of segmenting the claustrum of the human brain, as reflected by moderate-to-high Dice coefficient values for inter and intrarater comparisons.

Conclusions: Valid and repeatable segmentation of the human claustrum is crucial to developing a more substantial literature on this elusive structure. The development of a reliable and accurate method of segmenting the claustrum in neuroimaging will facilitate tractography and functional connectivity studies that can provide further insight into the function and connectivity of this structure.

Live Zoom Session – April 21st

research Areas

Authors

Nicholas R. Prunier, BA, Nikos Makris, MD, PhD, Martha E. Shenton, PhD, Richard Jarrett Rushmore III, PhD

Principal Investigator

Richard Jarrett Rushmore III, PhD