A novel neuroimaging based early detection framework for alzheimer disease using deep learning.

Area of Science:

• Neuroscience
• Artificial Intelligence
• Medical Imaging

Background:

• Alzheimer's disease (AD) is a progressive neurodegenerative disorder with increasing global prevalence.
• Early diagnosis of AD is crucial for effective intervention but remains a significant challenge.
• Current diagnostic methods often lack the sensitivity for timely detection, impacting patient management.

Purpose of the Study:

• To develop and evaluate a novel deep learning framework, NEDA-DL, for the early detection of Alzheimer's disease.
• To leverage neuroimaging data (MRI and PET) for enhanced diagnostic accuracy.
• To improve upon existing computer-aided diagnostic (CAD) systems for AD detection.

Main Methods:

• A hybrid deep learning architecture combining ResNet-50 and AlexNet was employed.
• CUDA-based parallel processing and depthwise separable convolutions were utilized for computational efficiency.
• The NEDA-DL framework processed MRI and PET neuroimaging data for classification.

Main Results:

• The NEDA-DL framework demonstrated state-of-the-art classification performance.
• The Softmax classifier achieved a high accuracy of 99.87%.
• Comparative analyses confirmed the superiority of NEDA-DL over existing methods.

Conclusions:

• NEDA-DL offers a highly accurate and efficient approach for early Alzheimer's disease detection using neuroimaging.
• The framework integrates structural and functional imaging insights to enhance diagnostic precision.
• NEDA-DL shows significant potential to support clinical decision-making in Alzheimer's disease diagnosis.