Scholarlink

<p>Alzheimer's disease (AD) is the most common form of dementia and represents a severe neurodegenerative condition. Identifying individuals at high risk of developing AD from Mild Cognitive Impairment (MCI) is critical for early detection and effective management. This study aims to develop a robust experimental design to predict the clinical status of patients at subsequent visits via baseline neuroimaging data. A large dataset comprising lifestyle, cognitive, and medical assessments (MRI, DTI, PET, and CSF tests) from more than 12,741 individuals has been analysed. Data has been preprocessed, followed by feature selection using LASSO technique. Multiple Machine Learning (ML) methods, such as Gaussian Naïve Bayes, Decision Tree, Support Vector Machine, Logistic Regression, Multilayer Perceptron, Random Forest, and Gradient Boosting, have been applied for predicting AD Progression. The models have been tested on hold-out test datasets, ranging from single to multimodal datasets. Results indicate that cognitive tests have played a major role in improving the performance while combining with other modalities<i>.</i> Disease progression prediction models had been studied over a period from month 6 (m06) to month 48 (m48). Significance of the models has been tested with the Friedman test. The present study contributes to improved early diagnosis of Alzheimer's patients, leading to better health outcomes.</p>

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

Enhancing early Alzheimer's diagnosis through machine learning on multimodal neuroimaging and cognitive data

  • 期刊论文

  • 发表日期：

    2025年6月30日

查看全文
  • Parneet Kaur,
  • Rehan Ali,
  • Gurpreet Kaur,
  • Jasvinder Singh Bhatti,
  • Satwinder Singh

摘要

Alzheimer's disease (AD) is the most common form of dementia and represents a severe neurodegenerative condition. Identifying individuals at high risk of developing AD from Mild Cognitive Impairment (MCI) is critical for early detection and effective management. This study aims to develop a robust experimental design to predict the clinical status of patients at subsequent visits via baseline neuroimaging data. A large dataset comprising lifestyle, cognitive, and medical assessments (MRI, DTI, PET, and CSF tests) from more than 12,741 individuals has been analysed. Data has been preprocessed, followed by feature selection using LASSO technique. Multiple Machine Learning (ML) methods, such as Gaussian Naïve Bayes, Decision Tree, Support Vector Machine, Logistic Regression, Multilayer Perceptron, Random Forest, and Gradient Boosting, have been applied for predicting AD Progression. The models have been tested on hold-out test datasets, ranging from single to multimodal datasets. Results indicate that cognitive tests have played a major role in improving the performance while combining with other modalities. Disease progression prediction models had been studied over a period from month 6 (m06) to month 48 (m48). Significance of the models has been tested with the Friedman test. The present study contributes to improved early diagnosis of Alzheimer's patients, leading to better health outcomes.