Search

Your search keyword '"Tsai, Chih-Chien"' showing total 26 results
Start Over Author "Tsai, Chih-Chien" Publication Year Range Last 3 years
26 results on '"Tsai, Chih-Chien"'

1. A worldwide study of white matter microstructural alterations in people living with Parkinson’s disease

Author

Owens-Walton, Conor, Nir, Talia M., Al-Bachari, Sarah, Ambrogi, Sonia, Anderson, Tim J., Aventurato, Ítalo Karmann, Cendes, Fernando, Chen, Yao-Liang, Ciullo, Valentina, Cook, Phil, Dalrymple-Alford, John C., Dirkx, Michiel F., Druzgal, Jason, Emsley, Hedley C. A., Guimarães, Rachel, Haroon, Hamied A., Helmich, Rick C., Hu, Michele T., Johansson, Martin E., Kim, Ho Bin, Klein, Johannes C., Laansma, Max, Lawrence, Katherine E., Lochner, Christine, Mackay, Clare, McMillan, Corey T., Melzer, Tracy R., Nabulsi, Leila, Newman, Ben, Opriessnig, Peter, Parkes, Laura M., Pellicano, Clelia, Piras, Fabrizio, Piras, Federica, Pirpamer, Lukas, Pitcher, Toni L., Poston, Kathleen L., Roos, Annerine, Silva, Lucas Scárdua, Schmidt, Reinhold, Schwingenschuh, Petra, Shahid-Besanti, Marian, Spalletta, Gianfranco, Stein, Dan J., Thomopoulos, Sophia I., Tosun, Duygu, Tsai, Chih-Chien, van den Heuvel, Odile A., van Heese, Eva, Vecchio, Daniela, Villalón-Reina, Julio E., Vriend, Chris, Wang, Jiun-Jie, Wu, Yih-Ru, Yasuda, Clarissa Lin, Thompson, Paul M., Jahanshad, Neda, and van der Werf, Ysbrand

Published
2024
Full Text
View/download PDF

2. Video and Synthetic MRI Pre-training of 3D Vision Architectures for Neuroimage Analysis

Author

Dhinagar, Nikhil J., Singh, Amit, Ozarkar, Saket, Buwa, Ketaki, Thomopoulos, Sophia I., Owens-Walton, Conor, Laltoo, Emily, Chen, Yao-Liang, Cook, Philip, McMillan, Corey, Tsai, Chih-Chien, Wang, J-J, Wu, Yih-Ru, and Thompson, Paul M.

Subjects
Electrical Engineering and Systems Science - Image and Video Processing, Computer Science - Artificial Intelligence, Computer Science - Computer Vision and Pattern Recognition
Abstract

Transfer learning represents a recent paradigm shift in the way we build artificial intelligence (AI) systems. In contrast to training task-specific models, transfer learning involves pre-training deep learning models on a large corpus of data and minimally fine-tuning them for adaptation to specific tasks. Even so, for 3D medical imaging tasks, we do not know if it is best to pre-train models on natural images, medical images, or even synthetically generated MRI scans or video data. To evaluate these alternatives, here we benchmarked vision transformers (ViTs) and convolutional neural networks (CNNs), initialized with varied upstream pre-training approaches. These methods were then adapted to three unique downstream neuroimaging tasks with a range of difficulty: Alzheimer's disease (AD) and Parkinson's disease (PD) classification, "brain age" prediction. Experimental tests led to the following key observations: 1. Pre-training improved performance across all tasks including a boost of 7.4% for AD classification and 4.6% for PD classification for the ViT and 19.1% for PD classification and reduction in brain age prediction error by 1.26 years for CNNs, 2. Pre-training on large-scale video or synthetic MRI data boosted performance of ViTs, 3. CNNs were robust in limited-data settings, and in-domain pretraining enhanced their performances, 4. Pre-training improved generalization to out-of-distribution datasets and sites. Overall, we benchmarked different vision architectures, revealing the value of pre-training them with emerging datasets for model initialization. The resulting pre-trained models can be adapted to a range of downstream neuroimaging tasks, even when training data for the target task is limited.

Published
2023

3. Curriculum Based Multi-Task Learning for Parkinson's Disease Detection

Author

Dhinagar, Nikhil J., Owens-Walton, Conor, Laltoo, Emily, Boyle, Christina P., Chen, Yao-Liang, Cook, Philip, McMillan, Corey, Tsai, Chih-Chien, Wang, J-J, Wu, Yih-Ru, van der Werf, Ysbrand, and Thompson, Paul M.

Subjects
Electrical Engineering and Systems Science - Image and Video Processing, Computer Science - Artificial Intelligence, Computer Science - Computer Vision and Pattern Recognition, Computer Science - Machine Learning, Quantitative Biology - Quantitative Methods
Abstract

There is great interest in developing radiological classifiers for diagnosis, staging, and predictive modeling in progressive diseases such as Parkinson's disease (PD), a neurodegenerative disease that is difficult to detect in its early stages. Here we leverage severity-based meta-data on the stages of disease to define a curriculum for training a deep convolutional neural network (CNN). Typically, deep learning networks are trained by randomly selecting samples in each mini-batch. By contrast, curriculum learning is a training strategy that aims to boost classifier performance by starting with examples that are easier to classify. Here we define a curriculum to progressively increase the difficulty of the training data corresponding to the Hoehn and Yahr (H&Y) staging system for PD (total N=1,012; 653 PD patients, 359 controls; age range: 20.0-84.9 years). Even with our multi-task setting using pre-trained CNNs and transfer learning, PD classification based on T1-weighted (T1-w) MRI was challenging (ROC AUC: 0.59-0.65), but curriculum training boosted performance (by 3.9%) compared to our baseline model. Future work with multimodal imaging may further boost performance., Comment: Accepted for publication at the 20th IEEE International Symposium on Biomedical Imaging, ISBI 2023

Published
2023

12. Video and synthetic MRI pre-training of 3D vision architectures for neuroimage analysis

Author

Dhinagar, Nikhil J., primary, Singh, Amit, additional, Ozarkar, Saket S., additional, Buwa, Ketaki U., additional, Thomopoulos, Sophia, additional, Owens-Walton, Conor, additional, Laltoo, Emily, additional, Chen, Yao-Liang, additional, Cook, Philip A., additional, McMillan, Corey, additional, Tsai, Chih-Chien, additional, Wang, J-J, additional, Wu, Yih-Ru, additional, and Thompson, Paul M., additional

Published
2024
Full Text
View/download PDF

13. A Worldwide Study of White Matter Microstructural Alterations in People Living with Parkinson’s Disease

Author

Owens-Walton, Conor, primary, Nir, Talia M., additional, Al-Bachari, Sarah, additional, Ambrogi, Sonia, additional, Anderson, Tim J., additional, Aventurato, Ítalo Karmann, additional, Cendes, Fernando, additional, Chen, Yao-Liang, additional, Ciullo, Valentina, additional, Cook, Phil, additional, Dalrymple-Alford, John C., additional, Dirkx, Michiel F., additional, Druzgal, Jason, additional, Emsley, Hedley C. A., additional, Guimarães, Rachel, additional, Haroon, Hamied A., additional, Helmich, Rick C., additional, Hu, Michele T., additional, Johansson, Martin E., additional, Kim, Ho Bin, additional, Klein, Johannes C., additional, Laansma, Max, additional, Lawrence, Katherine E., additional, Lochner, Christine, additional, Mackay, Clare, additional, McMillan, Corey, additional, Melzer, Tracy R., additional, Nabulsi, Leila, additional, Newman, Ben, additional, Opriessnig, Peter, additional, Parkes, Laura M., additional, Pellicano, Clelia, additional, Piras, Fabrizio, additional, Piras, Federica, additional, Pirpamer, Lukas, additional, Pitcher, Toni L., additional, Poston, Kathleen L., additional, Roos, Annerine, additional, Silva, Lucas Scárdua, additional, Schmidt, Reinhold, additional, Schwingenschuh, Petra, additional, Shahid, Marian, additional, Spalletta, Gianfranco, additional, Stein, Dan J., additional, Thomopoulos, Sophia I., additional, Tosun, Duygu, additional, Tsai, Chih-Chien, additional, van den Heuvel, Odile A., additional, van Heese, Eva, additional, Vecchio, Daniela, additional, Villalón-Reina, Julio E., additional, Vriend, Chris, additional, Wang, Jiun-Jie, additional, Wu, Yih-Ru, additional, Yasuda, Clarissa Lin, additional, Thompson, Paul M., additional, Jahanshad, Neda, additional, and van der Werf, Ysbrand, additional

Published
2024
Full Text
View/download PDF

16. Comparison of Anatomical and Diffusion MRI for detecting Parkinson’s Disease using Deep Convolutional Neural Network

Author

Chattopadhyay, Tamoghna, primary, Singh, Amit, additional, Laltoo, Emily, additional, Boyle, Christina P., additional, Owens-Walton, Conor, additional, Chen, Yao-Liang, additional, Cook, Philip, additional, McMillan, Corey, additional, Tsai, Chih-Chien, additional, Wang, J-J, additional, Wu, Yih-Ru, additional, van der Werf, Ysbrand, additional, and Thompson, Paul M., additional

Published
2023
Full Text
View/download PDF

18. Curriculum Based Multi-Task Learning for Parkinson’s Disease Detection

Author

Dhinagar, Nikhil J., primary, Owens-Walton, Conor, additional, Laltoo, Emily, additional, Boyle, Christina P., additional, Chen, Yao-Liang, additional, Cook, Philip, additional, McMillan, Corey, additional, Tsai, Chih-Chien, additional, Wang, J-J, additional, Wu, Yih-Ru, additional, van der Werf, Ysbrand, additional, and Thompson, Paul M., additional

Published
2023
Full Text
View/download PDF

19. Comparison of Anatomical and Diffusion MRI for detecting Parkinson’s Disease using Deep Convolutional Neural Network

Author

Chattopadhyay, Tamoghna, Singh, Amit, Laltoo, Emily, Boyle, Christina P., Owens-Walton, Conor, Chen, Yao-Liang, Cook, Philip, McMillan, Corey, Tsai, Chih-Chien, Wang, J-J, Wu, Yih-Ru, van der Werf, Ysbrand, and Thompson, Paul M.

Subjects
Article
Abstract

Parkinson’s disease (PD) is a progressive neurodegenerative disease that affects over 10 million people worldwide. Brain atrophy and microstructural abnormalities tend to be more subtle in PD than in other age-related conditions such as Alzheimer’s disease, so there is interest in how well machine learning methods can detect PD in radiological scans. Deep learning models based on convolutional neural networks (CNNs) can automatically distil diagnostically useful features from raw MRI scans, but most CNN-based deep learning models have only been tested on T1-weighted brain MRI. Here we examine the added value of diffusion-weighted MRI (dMRI) - a variant of MRI, sensitive to microstructural tissue properties - as an additional input in CNN-based models for PD classification. Our evaluations used data from 3 separate cohorts - from Chang Gung University, the University of Pennsylvania, and the PPMI dataset. We trained CNNs on various combinations of these cohorts to find the best predictive model. Although tests on more diverse data are warranted, deep-learned models from dMRI show promise for PD classification.Clinical RelevanceThis study supports the use of diffusion-weighted images as an alternative to anatomical images for AI-based detection of Parkinson’s disease.

Published
2023
Full Text
View/download PDF

23. T1 and T2∗ relaxation time in the parcellated myocardium of healthy Taiwanese participants: A single center study

Author

Tsai, Chih-Chien, primary, Ng, Shu-Hang, additional, Chen, Yao-Liang, additional, Juan, Yu-Hsiang, additional, Wang, Chao-Hung, additional, Lin, Gigin, additional, Chien, Chih-Wei, additional, Lin, Yu-Ching, additional, Lin, Yu-Chun, additional, Huang, Yu-Chieh, additional, Huang, Pei-Ching, additional, and Wang, Jiun-Jie, additional

Published
2021
Full Text
View/download PDF

24. Cerebellar Volume and Disease Staging in Parkinson's Disease: An ENIGMA-PD Study.

Author

Kerestes R, Laansma MA, Owens-Walton C, Perry A, van Heese EM, Al-Bachari S, Anderson TJ, Assogna F, Aventurato ÍK, van Balkom TD, Berendse HW, van den Berg KRE, Betts R, Brioschi R, Carr J, Cendes F, Clark LR, Dalrymple-Alford JC, Dirkx MF, Druzgal J, Durrant H, Emsley HCA, Garraux G, Haroon HA, Helmich RC, van den Heuvel OA, João RB, Johansson ME, Khachatryan SG, Lochner C, McMillan CT, Melzer TR, Mosley PE, Newman B, Opriessnig P, Parkes LM, Pellicano C, Piras F, Pitcher TL, Poston KL, Rango M, Roos A, Rummel C, Schmidt R, Schwingenschuh P, Silva LS, Smith V, Squarcina L, Stein DJ, Tavadyan Z, Tsai CC, Vecchio D, Vriend C, Wang JJ, Wiest R, Yasuda CL, Young CB, Jahanshad N, Thompson PM, van der Werf YD, and Harding IH

Subjects
Humans, Cross-Sectional Studies, Magnetic Resonance Imaging, Cerebellum, Brain, Parkinson Disease complications
Abstract

Background: Increasing evidence points to a pathophysiological role for the cerebellum in Parkinson's disease (PD). However, regional cerebellar changes associated with motor and non-motor functioning remain to be elucidated., Objective: To quantify cross-sectional regional cerebellar lobule volumes using three dimensional T1-weighted anatomical brain magnetic resonance imaging from the global ENIGMA-PD working group., Methods: Cerebellar parcellation was performed using a deep learning-based approach from 2487 people with PD and 1212 age and sex-matched controls across 22 sites. Linear mixed effects models compared total and regional cerebellar volume in people with PD at each Hoehn and Yahr (HY) disease stage, to an age- and sex- matched control group. Associations with motor symptom severity and Montreal Cognitive Assessment scores were investigated., Results: Overall, people with PD had a regionally smaller posterior lobe (d max  = -0.15). HY stage-specific analyses revealed a larger anterior lobule V bilaterally (d max  = 0.28) in people with PD in HY stage 1 compared to controls. In contrast, smaller bilateral lobule VII volume in the posterior lobe was observed in HY stages 3, 4, and 5 (d max  = -0.76), which was incrementally lower with higher disease stage. Within PD, cognitively impaired individuals had lower total cerebellar volume compared to cognitively normal individuals (d = -0.17)., Conclusions: We provide evidence of a dissociation between anterior "motor" lobe and posterior "non-motor" lobe cerebellar regions in PD. Whereas less severe stages of the disease are associated with larger motor lobe regions, more severe stages of the disease are marked by smaller non-motor regions. © 2023 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society., (© 2023 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.)

Published
2023
Full Text
View/download PDF

25. Comparison of Anatomical and Diffusion MRI for detecting Parkinson's Disease using Deep Convolutional Neural Network.

Author

Chattopadhyay T, Singh A, Laltoo E, Boyle CP, Owens-Walton C, Chen YL, Cook P, McMillan C, Tsai CC, Wang JJ, Wu YR, van der Werf Y, and Thompson PM

Subjects
Humans, Magnetic Resonance Imaging methods, Neural Networks, Computer, Diffusion Magnetic Resonance Imaging, Parkinson Disease diagnostic imaging, Neurodegenerative Diseases
Abstract

Parkinson's disease (PD) is a progressive neurodegenerative disease that affects over 10 million people worldwide. Brain atrophy and microstructural abnormalities tend to be more subtle in PD than in other age-related conditions such as Alzheimer's disease, so there is interest in how well machine learning methods can detect PD in radiological scans. Deep learning models based on convolutional neural networks (CNNs) can automatically distil diagnostically useful features from raw MRI scans, but most CNN-based deep learning models have only been tested on T1-weighted brain MRI. Here we examine the added value of diffusion-weighted MRI (dMRI) - a variant of MRI, sensitive to microstructural tissue properties - as an additional input in CNN-based models for PD classification. Our evaluations used data from 3 separate cohorts - from Chang Gung University, the University of Pennsylvania, and the PPMI dataset. We trained CNNs on various combinations of these cohorts to find the best predictive model. Although tests on more diverse data are warranted, deep-learned models from dMRI show promise for PD classification.Clinical Relevance- This study supports the use of diffusion-weighted images as an alternative to anatomical images for AI-based detection of Parkinson's disease.

Published
2023
Full Text
View/download PDF

26. Curriculum Based Multi-Task Learning for Parkinson's Disease Detection.

Author

Dhinagar NJ, Owens-Walton C, Laltoo E, Boyle CP, Chen YL, Cook P, McMillan C, Tsai CC, Wang JJ, Wu YR, van der Werf Y, and Thompson PM

Abstract

There is great interest in developing radiological classifiers for diagnosis, staging, and predictive modeling in progressive diseases such as Parkinson's disease (PD), a neurodegenerative disease that is difficult to detect in its early stages. Here we leverage severity-based meta-data on the stages of disease to define a curriculum for training a deep convolutional neural network (CNN). Typically, deep learning networks are trained by randomly selecting samples in each mini-batch. By contrast, curriculum learning is a training strategy that aims to boost classifier performance by starting with examples that are easier to classify. Here we define a curriculum to progressively increase the difficulty of the training data corresponding to the Hoehn and Yahr (H&Y) staging system for PD (total N=1,012; 653 PD patients, 359 controls; age range: 20.0-84.9 years). Even with our multi-task setting using pre-trained CNNs and transfer learning, PD classification based on T1-weighted (T1-w) MRI was challenging (ROC AUC: 0.59-0.65), but curriculum training boosted performance (by 3.9%) compared to our baseline model. Future work with multimodal imaging may further boost performance.

Published
2023

Catalog

Books, media, physical & digital resources
See catalog results