Scholarlink

Assessing the alternations of the fornix region in structural magnetic resonance (MR) images of the brain is essential for early detection of Alzheimer’s disease (AD). To distinguish between early and late conditions of mild cognitive impairment (MCI), this study attempted to segment 3D volumetric fornix structures and measure geometric changes. Images of T1-weighted sMR brain images obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. The FreeSurfer toolbox is used to segment these images after pre-processing. Geometrical features, including volume, convex volume, surface area, major and minor axis lengths, and equivalent diameter, are obtained from the segmented fornix region. The statistical significance of the extracted features is assessed using the Wilcoxon rank-sum test, student t-test, and Shapiro–Wilk test. These statistically significant features are fed into a random forest (RF) model, which binary classifies MCI conditions using a ten fold cross-validation technique. The results showed that every feature taken into consideration was statistically significant (p < 0.05). The RF classifier obtained an F1-score of 80% and an accuracy of 70%. Therefore, the alterations in the fornix structure carried on by late MCI can be studied using geometrical features.

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

Machine Learning-Based Classification of Early and Late Mild Cognitive Impairment Conditions Using Structural MR Images of Fornix and Geometric Features

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Prince Lassemento Fully,
  • Ahsan Ali,
  • Jac Fredo Agastinose Ronickom,
  • Ramakrishnan Swaminathan

摘要

Assessing the alternations of the fornix region in structural magnetic resonance (MR) images of the brain is essential for early detection of Alzheimer’s disease (AD). To distinguish between early and late conditions of mild cognitive impairment (MCI), this study attempted to segment 3D volumetric fornix structures and measure geometric changes. Images of T1-weighted sMR brain images obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. The FreeSurfer toolbox is used to segment these images after pre-processing. Geometrical features, including volume, convex volume, surface area, major and minor axis lengths, and equivalent diameter, are obtained from the segmented fornix region. The statistical significance of the extracted features is assessed using the Wilcoxon rank-sum test, student t-test, and Shapiro–Wilk test. These statistically significant features are fed into a random forest (RF) model, which binary classifies MCI conditions using a ten fold cross-validation technique. The results showed that every feature taken into consideration was statistically significant (p < 0.05). The RF classifier obtained an F1-score of 80% and an accuracy of 70%. Therefore, the alterations in the fornix structure carried on by late MCI can be studied using geometrical features.