Dataset Information

A scalable method to improve gray matter segmentation at ultra high field MRI.

ABSTRACT: High-resolution (functional) magnetic resonance imaging (MRI) at ultra high magnetic fields (7 Tesla and above) enables researchers to study how anatomical and functional properties change within the cortical ribbon, along surfaces and across cortical depths. These studies require an accurate delineation of the gray matter ribbon, which often suffers from inclusion of blood vessels, dura mater and other non-brain tissue. Residual segmentation errors are commonly corrected by browsing the data slice-by-slice and manually changing labels. This task becomes increasingly laborious and prone to error at higher resolutions since both work and error scale with the number of voxels. Here we show that many mislabeled, non-brain voxels can be corrected more efficiently and semi-automatically by representing three-dimensional anatomical images using two-dimensional histograms. We propose both a uni-modal (based on first spatial derivative) and multi-modal (based on compositional data analysis) approach to this representation and quantify the benefits in 7 Tesla MRI data of nine volunteers. We present an openly accessible Python implementation of these approaches and demonstrate that editing cortical segmentations using two-dimensional histogram representations as an additional post-processing step aids existing algorithms and yields improved gray matter borders. By making our data and corresponding expert (ground truth) segmentations openly available, we facilitate future efforts to develop and test segmentation algorithms on this challenging type of data.

SUBMITTER: Gulban OF

PROVIDER: S-EPMC5991408 | biostudies-literature | 2018

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

A scalable method to improve gray matter segmentation at ultra high field MRI.

Gulban Omer Faruk OF Schneider Marian M Marquardt Ingo I Haast Roy A M RAM De Martino Federico F

PloS one 20180606 6

High-resolution (functional) magnetic resonance imaging (MRI) at ultra high magnetic fields (7 Tesla and above) enables researchers to study how anatomical and functional properties change within the cortical ribbon, along surfaces and across cortical depths. These studies require an accurate delineation of the gray matter ribbon, which often suffers from inclusion of blood vessels, dura mater and other non-brain tissue. Residual segmentation errors are commonly corrected by browsing the data sl ...[more]

PMID: 29874295

Similar Datasets

Project description:ObjectiveMagnetic resonance imaging at 7T offers improved image spatial and contrast resolution for visualization of small brain nuclei targeted in neuromodulation. However, greater image geometric distortion and a lack of compatible instrumentation preclude implementation. In this report, the authors detail the development of a stereotactic image localizer and accompanying imaging sequences designed to mitigate geometric distortion, enabling accurate image registration and surgical planning of basal ganglia nuclei.MethodsMagnetization-prepared rapid acquisition with gradient echo (MPRAGE), fast gray matter acquisition T1 inversion recovery (FGATIR), T2-weighted, and T2*-weighted sequences were optimized for 7T in 9 human subjects to visualize basal ganglia nuclei, minimize image distortion, and maximize target contrast-to-noise and signal-to-noise ratios. Extracranial spatial distortions were mapped to develop a skull-contoured image localizer embedded with spherical silicone fiducials for improved MR image registration and target guidance. Surgical plan accuracy testing was initially performed in a custom-developed MRI phantom (n = 5 phantom studies) and finally in a human trial.ResultsMPRAGE and T2*-weighted sequences had the best measures among global measures of image quality (3.8/4, p < 0.0001; and 3.7/4, p = 0.0002, respectively). Among basal ganglia nuclei, FGATIR outperformed MPRAGE for globus pallidus externus (GPe) visualization (2.67/4 vs 1.78/4, p = 0.008), and FGATIR, T2-weighted imaging, and T2*-weighted imaging outperformed MPRAGE for substantia nigra visualization (1.44/4 vs 2.56/4, p = 0.04; vs 2.56/4, p = 0.04; vs 2.67/4, p = 0.003). Extracranial distortion was lower in the head's midregion compared with the base and apex ( 1.17-1.33 mm; MPRAGE and FGATIR, p < 0.0001; T2-weighted imaging, p > 0.05; and T2*-weighted imaging, p = 0.013). Fiducial placement on the localizer in low distortion areas improved image registration (fiducial registration error, 0.79-1.19 mm; p < 0.0001) and targeting accuracy (target registration error, 0.60-1.09 mm; p = 0.04). Custom surgical software and the refined image localizer enabled successful surgical planning in a human trial (fiducial registration error = 1.0 mm).ConclusionsA skull-contoured image localizer that accounts for image distortion is necessary to enable high-accuracy 7T imaging-guided targeting for surgical neuromodulation. These results may enable improved clinical efficacy for the treatment of neurological disease.

Dataset Information

A scalable method to improve gray matter segmentation at ultra high field MRI.

Publications

A scalable method to improve gray matter segmentation at ultra high field MRI.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets