Using Machine Learning for Alzheimer’s disease diagnosis part1
- Evidence-empowered Transfer Learning for Alzheimer’s Disease(arXiv)
Author : : Kai Tzu-iunn Ong, Hana Kim, Minjin Kim, Jinseong Jang, Beomseok Sohn, Yoon Seong Choi, Dosik Hwang, Seong Jae Hwang, Jinyoung Yeo
Abstract : Transfer learning has been widely utilized to mitigate the data scarcity problem in the field of Alzheimer’s disease (AD). Conventional transfer learning relies on re-using models trained on AD-irrelevant tasks such as natural image classification. However, it often leads to negative transfer due to the discrepancy between the non-medical source and target medical domains. To address this, we present evidence-empowered transfer learning for AD diagnosis. Unlike conventional approaches, we leverage an AD-relevant auxiliary task, namely morphological change prediction, without requiring additional MRI data. In this auxiliary task, the diagnosis model learns the evidential and transferable knowledge from morphological features in MRI scans. Experimental results demonstrate that our framework is not only effective in improving detection performance regardless of model capacity, but also more data-efficient and faithful
2. Characterize the non-Gaussian diffusion property of cerebrospinal fluid using Diffusion Kurtosis Imaging and explore its diagnostic efficacy for Alzheimer’s disease(arXiv)
Author : Yingnan Xue, Min Wen, Qiong Ye
Abstract : Differentiating Alzheimer’s disease (AD) patients from healthy controls (HCs) remains a challenge. The changes of protein level in cerebrospinal fluid (CSF) of AD patients have been reported in the literature. Macromolecules will hinder the movement of water in CSF and lead to non-Gaussian diffusion. Diffusion kurtosis imaging (DKI) is a commonly used technique for quantifying non-Gaussian diffusivity. In this study, we used DKI to evaluate the non-Gaussian diffusion of CSF in AD patients and HC. Between-group difference was explored. In addition, we have built a prediction model using cross-validation Support Vector Machines (SVM), and achieved excellent performance. The validated area under the receiver operating characteristic curve(AUC) is in the range of 0.96–1.00, and the correct prediction is in the range of 87.1% — 90.0%
