Abstract #0399
Deep Learning Methods for Reversible Cerebral Vasoconstriction Syndrome Classification Based on Resting-state fMRI Images
Tun-Wei Hsu1,2,3, Chia-Hung Wu1,4,5, Hsiu-Mei Wu1,4, Kuan-Lin Lai4,6,7, Shih-Pin Chen4,6,7,8, Shuu-Jiun Wang4,6,7, Jiing-Feng Liring1,4, and Wan-Yuo Guo1,4
1Department of Radiology, Taipei Veterans General Hospital, Taipei, Taiwan, 2Integrated PET/MR Imaging Center, Taipei Veterans General Hospital, Taipei, Taiwan, 3Department of Biomedical Imaging and Radiological Sciences, National Yang-Ming Chiao Tung University, Taipei, Taiwan, 4School of Medicine, College of Medicine, National Yang-Ming Chiao Tung University, Taipei, Taiwan, 5Institute of Clinical Medicine, National Yang-Ming Chiao Tung University, Taipei, Taiwan, 6Department of Neurology, Neurological Institute, Taipei Veterans General Hospital, Taipei, Taiwan, 7Brain Research Center, National Yang-Ming Chiao Tung University, Taipei, Taiwan, 8Division of Translational Research, Taipei Veterans General Hospital, Taipei, Taiwan
Synopsis
Reversible cerebral vasoconstriction syndrome (RCVS) is a reversible segmental and multifocal vasoconstriction of the cerebral arteries and is believed to relate to autonomic network over-activity. We used Long Short-Term Memory networks (LSTMs), a type of deep neural network designed to handle time sequence data, to learn directly from the rs-fMRI time-series for classification of individuals with RCVS and healthy controls based on the regions in autonomic and other functional networks. These results provide methodological implications for rs-fMRI data of RCVS patients involved in the analysis and are a key element in future studies.
This abstract and the presentation materials are available to members only;
a login is required.
Join Here
