Frontiers in Neuroscience (Oct 2020)
Automatic Skull Stripping of Rat and Mouse Brain MRI Data Using U-Net
- Li-Ming Hsu,
- Li-Ming Hsu,
- Li-Ming Hsu,
- Li-Ming Hsu,
- Shuai Wang,
- Shuai Wang,
- Paridhi Ranadive,
- Woomi Ban,
- Woomi Ban,
- Tzu-Hao Harry Chao,
- Tzu-Hao Harry Chao,
- Tzu-Hao Harry Chao,
- Sheng Song,
- Sheng Song,
- Sheng Song,
- Domenic Hayden Cerri,
- Domenic Hayden Cerri,
- Domenic Hayden Cerri,
- Lindsay R. Walton,
- Lindsay R. Walton,
- Lindsay R. Walton,
- Margaret A. Broadwater,
- Margaret A. Broadwater,
- Margaret A. Broadwater,
- Sung-Ho Lee,
- Sung-Ho Lee,
- Sung-Ho Lee,
- Dinggang Shen,
- Dinggang Shen,
- Dinggang Shen,
- Yen-Yu Ian Shih,
- Yen-Yu Ian Shih,
- Yen-Yu Ian Shih
Affiliations
- Li-Ming Hsu
- Center for Animal Magnetic Resonance Imaging, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Li-Ming Hsu
- Biomedical Research Imaging Center, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Li-Ming Hsu
- Department of Neurology, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Li-Ming Hsu
- Department of Radiology, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Shuai Wang
- Biomedical Research Imaging Center, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Shuai Wang
- Department of Radiology, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Paridhi Ranadive
- Center for Animal Magnetic Resonance Imaging, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Woomi Ban
- Center for Animal Magnetic Resonance Imaging, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Woomi Ban
- Biomedical Research Imaging Center, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Tzu-Hao Harry Chao
- Center for Animal Magnetic Resonance Imaging, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Tzu-Hao Harry Chao
- Biomedical Research Imaging Center, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Tzu-Hao Harry Chao
- Department of Neurology, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Sheng Song
- Center for Animal Magnetic Resonance Imaging, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Sheng Song
- Biomedical Research Imaging Center, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Sheng Song
- Department of Neurology, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Domenic Hayden Cerri
- Center for Animal Magnetic Resonance Imaging, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Domenic Hayden Cerri
- Biomedical Research Imaging Center, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Domenic Hayden Cerri
- Department of Neurology, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Lindsay R. Walton
- Center for Animal Magnetic Resonance Imaging, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Lindsay R. Walton
- Biomedical Research Imaging Center, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Lindsay R. Walton
- Department of Neurology, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Margaret A. Broadwater
- Center for Animal Magnetic Resonance Imaging, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Margaret A. Broadwater
- Biomedical Research Imaging Center, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Margaret A. Broadwater
- Department of Neurology, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Sung-Ho Lee
- Center for Animal Magnetic Resonance Imaging, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Sung-Ho Lee
- Biomedical Research Imaging Center, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Sung-Ho Lee
- Department of Neurology, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Dinggang Shen
- Biomedical Research Imaging Center, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Dinggang Shen
- Department of Radiology, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Dinggang Shen
- Department of Brain and Cognitive Engineering, Korea University, Seoul, South Korea
- Yen-Yu Ian Shih
- Center for Animal Magnetic Resonance Imaging, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Yen-Yu Ian Shih
- Biomedical Research Imaging Center, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- Yen-Yu Ian Shih
- Department of Neurology, The University of North Carolina at Chapel Hill, Chapel Hill, NC, United States
- DOI
- https://doi.org/10.3389/fnins.2020.568614
- Journal volume & issue
-
Vol. 14
Abstract
Accurate removal of magnetic resonance imaging (MRI) signal outside the brain, a.k.a., skull stripping, is a key step in the brain image pre-processing pipelines. In rodents, this is mostly achieved by manually editing a brain mask, which is time-consuming and operator dependent. Automating this step is particularly challenging in rodents as compared to humans, because of differences in brain/scalp tissue geometry, image resolution with respect to brain-scalp distance, and tissue contrast around the skull. In this study, we proposed a deep-learning-based framework, U-Net, to automatically identify the rodent brain boundaries in MR images. The U-Net method is robust against inter-subject variability and eliminates operator dependence. To benchmark the efficiency of this method, we trained and validated our model using both in-house collected and publicly available datasets. In comparison to current state-of-the-art methods, our approach achieved superior averaged Dice similarity coefficient to ground truth T2-weighted rapid acquisition with relaxation enhancement and T2∗-weighted echo planar imaging data in both rats and mice (all p < 0.05), demonstrating robust performance of our approach across various MRI protocols.
Keywords
