Federated Learning for Intrusion Detection in Internet of Medical Things (IoMT): A PRISMA-Based Systematic Literature Review
Abstract
Abstract The rapid expansion of Internet of Medical Things (IoMT) devices has introduced new opportunities for real-time healthcare monitoring but also increased cybersecurity risks. Traditional centralized intrusion detection systems (IDS) struggle to scale across distributed IoMT networks while preserving sensitive patient data. Federated Learning (FL) offers a privacy-preserving alternative, enabling collaborative model training without sharing raw data. This study presents a PRISMA-based systematic literature review of FL applications for IoMT intrusion detection, analyzing 86 peer-reviewed studies published between 2015 and 2025 from Scopus, IEEE Xplore, Web of Science, and PubMed. Findings indicate that FL-based IDS frameworks achieve competitive accuracy (85–93%) and F1-scores (~ 0.87), yet face challenges in handling non-IID heterogeneous data, optimizing energy and communication-efficiency, and ensuring robust privacy guarantees. Based on the synthesis, we propose a conceptual framework integrating privacy-preserving aggregation, adaptive federated optimization, and edge-intelligence mechanisms, tailored to IoMT constraints. This work provides a comprehensive foundation for developing scalable, privacy-aware, and resource-efficient IoMT intrusion detection systems, while identifying key research gaps and future directions for enhancing robustness, personalization, and real-world deployment.
Read the paper
Related studies
- Active Learning during Federated Learning for Object Detection
- Fairness-aware federated learning
- Decentralized Federated Learning
- Building a Personally Identifiable Information Recognizer in a Privacy Preserved Manner Using Automated Annotation and Federated Learning
- Preservation of Patient Level Privacy: Federated Classification and Calibration Models
- Efficient explainable federated learning
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.