Research on an AI Model for Early Alzheimer's Detection

Abstract

Alzheimer's Disease (AD) is a neurodegenerative ailment with significant impact. This study aims to enhance early AD detection accuracy using a novel AI model. Leveraging multimodal image fusion (MRI and PET) and a self-attention mechanism, the model captures complex brain region relationships. Experiments on ADNI and OASIS datasets, with data preprocessing and augmentation, yielded a 92.5% accuracy and 0.95 AUC on the test set, outperforming traditional methods. Grad-CAM heatmaps enhanced model interpretability. However, challenges like data quality dependence and computational complexity remain. Future work will focus on data augmentation, model compression, and cross-domain validation to improve clinical application potential and further AD research.