This paper proposes an intrusion detection system based on federated learning to address security vulnerabilities in Internet of Things (IoMT) healthcare systems. While IoMT enables early disease diagnosis and personalized treatment through real-time health data collection, the sensitivity of the data makes it vulnerable to security threats. In this paper, we address these challenges by implementing artificial neural network-based intrusion detection, federated learning-based privacy protection, and enhanced model interpretability using explainable artificial intelligence (XAI). Using network and medical datasets that simulate various attack types, we compare the effectiveness of the proposed framework with that of a centralized approach. We demonstrate that the federated learning approach performs similarly to the centralized approach while simultaneously preserving privacy and providing model explainability.
