AUTHOR=Yang Hongqin , Mao Jiangbing , Ye Qinyong , Bucholc Magda , Liu Shuo , Gao Wenzhao , Pan Jie , Xin Jiawei , Ding Xuemei 

TITLE=Distance-based novelty detection model for identifying individuals at risk of developing Alzheimer's disease

JOURNAL=Frontiers in Aging Neuroscience

VOLUME=Volume 16 - 2024

YEAR=2024

URL=https://www.frontiersin.org/journals/aging-neuroscience/articles/10.3389/fnagi.2024.1285905

DOI=10.3389/fnagi.2024.1285905

ISSN=1663-4365

ABSTRACT=Novelty detection (ND), also known as anomaly detection, is a machine learning technique used to identify patterns that are typical of the majority class and can discriminate deviations as novelties. In the context of Alzheimer's disease (AD), ND could be employed to detect abnormal or atypical behavior that may indicate early signs of cognitive decline or the presence of the disease. So far, few research studies used ND to discriminate the risk of developing AD and mild cognitive impairment (MCI) from healthy controls (HC). In this project, we utilized four easily interpretable ND methods based on k-nearest neighbors, Mixture of Gaussian (MoG), k-means clustering, and support vector data description algorithms. They were employed to create a decision boundary (novelty detector) trained solely on HC data. Two distinct cohorts with highly heterogeneous data, derived from the Australian Imaging Biomarkers and Lifestyle Study of Ageing (AIBL) and the Fujian Medical University Union Hospital (FMUUH) China, were employed. Our experimental results indicated that the best overall performance of detecting AD individuals in AIBL and FMUUH datasets was obtained by using the MoG-based ND algorithm, with AUC of 0.8757 and 0.9443, sensitivity of 96.79% and 89.09%, specificity of 89.63% and 90.92%, respectively. As such, a user-friendly web-based graphical interface tailored for non-technical stakeholders was developed with the proposed ND-based framework. This platform offers an interactive environment to aid in making diagnoses of MCI and AD, enabling streamlined decision-making processes. More importantly, the proposed DtB strategy could visually and quantitatively reflect the severity of developing AD of individuals.