Scholarlink

Purpose <p>Alzheimer's disease (AD) is a progressive neurodegenerative disorder characterized by cognitive decline and memory loss, imposing significant emotional, financial, and societal burdens on patients, caregivers, and healthcare systems. This research investigates the application of Convolutional Neural Networks (CNNs) to automate the early detection of AD using Magnetic Resonance Imaging (MRI) scans, aiming to identify the most effective CNN architectures for distinguishing between Healthy Controls (CN), Mild Cognitive Impairment (MCI), and AD patients.</p> Methods <p>The study evaluated several CNN architectures, including DenseNet121, InceptionV3, VGG16, LeNet, and ResNet50, to determine their efficacy in classifying MRI scans into CN, MCI, and AD categories. Each model's performance was assessed based on metrics such as accuracy, precision, recall, Area Under the Curve (AUC), and F1-score. The models were trained and tested on benchmark datasets from UCI and KEEL repositories, ensuring a comprehensive evaluation across diverse data samples.</p> Results <p>Among the evaluated architectures, DenseNet121 demonstrated superior performance. Specifically, it achieved a 93.9% accuracy rate in distinguishing between AD and CN subjects using the "Tanh" activation function. When employing the "SELU" activation function, the model attained a 94.9% accuracy in differentiating between CN and MCI subjects and a 93.8% accuracy in classifying AD and MCI subjects. These results highlight DenseNet121's robustness and reliability in early AD detection tasks.</p> Conclusions <p>The findings suggest that employing advanced CNN architectures, particularly DenseNet121, can significantly enhance the accuracy and efficiency of AD diagnosis through automated MRI analysis. This approach holds promise for reducing the workload on healthcare professionals and minimizing human error, facilitating earlier and more reliable detection of Alzheimer's disease.</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

Leveraging deep learning for early diagnosis of Alzheimer’s disease via brain MRI imaging

  • 期刊论文

  • 发表日期：

    2025年10月13日

查看全文
  • Sanidhya Diwakar,
  • Deepak Gupta,
  • Jyoti Maurya,
  • Barenya Bikash Hazarika

摘要

Purpose

Alzheimer's disease (AD) is a progressive neurodegenerative disorder characterized by cognitive decline and memory loss, imposing significant emotional, financial, and societal burdens on patients, caregivers, and healthcare systems. This research investigates the application of Convolutional Neural Networks (CNNs) to automate the early detection of AD using Magnetic Resonance Imaging (MRI) scans, aiming to identify the most effective CNN architectures for distinguishing between Healthy Controls (CN), Mild Cognitive Impairment (MCI), and AD patients.

Methods

The study evaluated several CNN architectures, including DenseNet121, InceptionV3, VGG16, LeNet, and ResNet50, to determine their efficacy in classifying MRI scans into CN, MCI, and AD categories. Each model's performance was assessed based on metrics such as accuracy, precision, recall, Area Under the Curve (AUC), and F1-score. The models were trained and tested on benchmark datasets from UCI and KEEL repositories, ensuring a comprehensive evaluation across diverse data samples.

Results

Among the evaluated architectures, DenseNet121 demonstrated superior performance. Specifically, it achieved a 93.9% accuracy rate in distinguishing between AD and CN subjects using the "Tanh" activation function. When employing the "SELU" activation function, the model attained a 94.9% accuracy in differentiating between CN and MCI subjects and a 93.8% accuracy in classifying AD and MCI subjects. These results highlight DenseNet121's robustness and reliability in early AD detection tasks.

Conclusions

The findings suggest that employing advanced CNN architectures, particularly DenseNet121, can significantly enhance the accuracy and efficiency of AD diagnosis through automated MRI analysis. This approach holds promise for reducing the workload on healthcare professionals and minimizing human error, facilitating earlier and more reliable detection of Alzheimer's disease.