Multimodal Image Classification Using Genetic Programming for Alzheimer’s Disease Diagnosis

Alzheimer’s disease (AD) is a progressive neurological disorder and a major contributor to dementia cases across the world. Timely and accurate diagnosis is crucial for effective clinical management and therapeutic intervention. This paper presents a genetic programming (GP) method with a multi-tree representation designed to effectively integrate multimodal neuroimaging data while preserving spatial information for AD classification. Unlike existing GP approaches that focus on single-modality data, our GP approach directly uses the images from multiple imaging sources as inputs into the evolutionary process. A new GP representation is designed to handle multimodal data effectively, enabling feature extraction and classification. Experiments on the commonly used public database of Alzheimer’s disease neuroimaging initiative (ADNI) show that the proposed method performs effectively in diagnosing AD. These findings suggest that multi-tree GP has the potential to serve as a powerful and interpretable tool for neuroimaging-based AD diagnosis, offering a promising approach to improve AD detection and clinical decision-making.