Scholarlink

Early detection of Alzheimer’s disease (AD) is challenging due to the complex of MRI data and the time demands of full diffusion-weighted imaging (DWI). This research presents an ethical AI-based deep learning framework that uses sparse DWI for rapid, tract-based AD classification. By leveraging diffusion tensor imaging (DTI) parameters, the method analyzes white matter (WM) tracts to distinguish Mild Cognitive Impairment (MCI) from Cognitively Normal (CN) individuals, focusing on longitudinal WM changes linked to AD. The framework reduces DWI acquisition time while maintaining performance comparable to full DWI. Results show that sparse DWI improves efficiency without compromising accuracy, offering a timely, cost-effective, and ethical approach for early AD diagnosis. AUC = 0.95, matching performance of dense acquisitions, and 91% classification accuracy were attained using just 5 diffusion directions.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Tract-Specific Biomarker Discovery for Early Alzheimer’s Disease Using Sparse Diffusion MRI and AI Framework

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Abhishek Tiwari,
  • Saurabh J. Shigwan

摘要

Early detection of Alzheimer’s disease (AD) is challenging due to the complex of MRI data and the time demands of full diffusion-weighted imaging (DWI). This research presents an ethical AI-based deep learning framework that uses sparse DWI for rapid, tract-based AD classification. By leveraging diffusion tensor imaging (DTI) parameters, the method analyzes white matter (WM) tracts to distinguish Mild Cognitive Impairment (MCI) from Cognitively Normal (CN) individuals, focusing on longitudinal WM changes linked to AD. The framework reduces DWI acquisition time while maintaining performance comparable to full DWI. Results show that sparse DWI improves efficiency without compromising accuracy, offering a timely, cost-effective, and ethical approach for early AD diagnosis. AUC = 0.95, matching performance of dense acquisitions, and 91% classification accuracy were attained using just 5 diffusion directions.