A Systematic Review of Trustworthy and Explainable AI Frameworks for Motor Imagery-Based Brain-Computer Interfaces in Robotic Control Systems

This study focuses on the reliability, robustness, and explainability of Motor Imagery-based Brain-Computer Interfaces (MI-BCIs). MI-BCIs can control external devices through imagined movements and have promising applications in neurorehabilitation, assistive robotics, and smart environments. Their practical application is hampered by signal reliability, variable evaluation, and the lack of defined assessment frameworks. A total of 43 peer-reviewed research studies in ScienceDirect, Scopus, and IEEE Xplore were analyzed and divided into six areas: robotic systems, healthcare and neurorehabilitation, smart environments and IoT, trustworthy and secure MI-BCIs, explainable AI, and learning frameworks. This review highlights similar motivations, important hurdles, and field-advancing strategies. The findings show significant repeatability, cross-subject generalization, and reliability gaps. To our knowledge, this is the first comprehensive investigation on the trust, robustness, and explainability of MI-BCI frameworks. Interdisciplinary collaboration, ethical norms, and subject-invariant techniques are needed to expedite MI-BCI development from the lab to the field.