A Multimodal Approach to Alzheimer’s Disease Classification Using Clinical and MRI Data

Abstract

Alzheimer’s disease (AD) is one of the significant health issues globally, as its neurodegenerative progression and challenges in early diagnosis create serious difficulties for clinical management. In recent years, artificial intelligence, particularly deep learning methods, has demonstrated strong performance in automated disease classification using medical imaging and clinical data. However, many existing models operate as black-box systems, limiting their transparency and clinical applicability.
In this study, an explainable multimodal AI approach is proposed for Alzheimer’s disease classification by combining deep learning and machine learning techniques. A convolutional neural network based on ResNet is used to extract spatial features from MRI images, while a machine learning model based on XGBoost is applied to structured clinical and biomarker data. To improve interpretability, explainability methods including SHAP and Grad-CAM are employed to identify key contributing features and relevant brain regions influencing model predictions. The models are evaluated using standard classification metrics, including accuracy, precision, recall, and F1-score. The results indicate that the proposed approach achieves strong classification performance while providing clinically meaningful explanations. These findings highlight the importance of combining predictive accuracy with interpretability, supporting the development of reliable AI-based diagnostic systems for Alzheimer’s disease.