LearnAD: Learning Interpretable Rules for Brain Networks in Alzheimer's Disease Classification

2025-12-30

arXiv.org

AI Summary

This work addresses the challenge of achieving high-accuracy and interpretable classification for Alzheimer’s disease diagnosis by proposing the first neuro-symbolic framework that integrates statistical modeling with logical rule learning. Specifically, a graph neural network (GNN) is employed to extract salient brain connectivity patterns from magnetic resonance imaging data, and these features are subsequently used by FastLAS to induce globally interpretable diagnostic rules. The resulting method achieves classification performance comparable to random forests and GNNs, outperforms decision trees, and matches support vector machines, while delivering fully transparent and clinically comprehensible decision logic. This approach significantly enhances the interpretability of GNN-based models in neuroscience applications, bridging the gap between high performance and clinical trustworthiness.

Technology Category

• Machine Learning: Neuro-Symbolic Learning
• Cognitive Modeling & Cognitive Systems: Symbolic Representations
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

We introduce LearnAD, a neuro-symbolic method for predicting Alzheimer's disease from brain magnetic resonance imaging data, learning fully interpretable rules. LearnAD applies statistical models, Decision Trees, Random Forests, or GNNs to identify relevant brain connections, and then employs FastLAS to learn global rules. Our best instance outperforms Decision Trees, matches Support Vector Machine accuracy, and performs only slightly below Random Forests and GNNs trained on all features, all while remaining fully interpretable. Ablation studies show that our neuro-symbolic approach improves interpretability with comparable performance to pure statistical models. LearnAD demonstrates how symbolic learning can deepen our understanding of GNN behaviour in clinical neuroscience.

Problem

• Alzheimer's disease
• interpretable rules
• brain networks
• MRI classification
• neuro-symbolic learning

Innovation

• neuro-symbolic learning
• interpretable rules
• Alzheimer's disease classification
• FastLAS
• graph neural networks