An explainable AI framework for integrated authentication, intrusion detection, and adaptive data transmission in IoMT networks
Year: 2026
Type: article
Abstract: Real-time monitoring, remote diagnostics, and automated medical treatments are all made possible by the Internet of Medical Things (IoMT), this has revolutionised healthcare. However, due of their improved connectivity, IoMT devices are more exposed to a range of cybersecurity attacks, such as malware insertion, spoofing, man-in-the-middle attacks, and distributed denial-of-service assaults. Traditional black-box AI models provide robust security mechanisms but lack transparency, making it difficult to identify and mitigate cyber threats effectively. This work introduces an explainable AI (XAI)-driven cybersecurity framework for IoMT applications to enhance trust, interpretability, and resilience against cyber threats. The proposed framework integrates deep learning networks with local model networks for authentication, employs a enhanced deep Q-Network for secure data transmission, and leverages XAI-based techniques for real-time attack detection and resistance. The model incorporates SHAP and LIME to provide interpretable security decisions, ensuring that IoMT stakeholders can understand and act on AI-generated security insights. In addition, this paper discusses a detailed attack resistance study using XAI, which shows the degree to which the proposed framework withstands various cyber threats.The paper evaluates security effectiveness metrics such as false positive rate, attack detection rate, response time, and the explainability score to judge the performance of XAI in IoMT security. Compared to conventional AI models, XAI-based detection accuracy has been proved to be significantly better experimentally, the reaction time was shortened, and the system was more transparent.
Source: Scientific Reports
Authors: Nithyavani G, Naga Raja G
Institution: Vellore Institute of Technology University
Language: English
Topic: Internet of Things and AI
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $2,690
Funder: VIT University
OpenAlex ID: W7236094516
