SAR Journal of Psychiatry and Neuroscience
Volume-7 | Issue-04
Original Research Article
Implementing a Specialty-Specific Artificial Intelligence Curriculum for Neurology Residents: A Pilot Educational Initiative
Shravani Khisti, Simren Jayaraman, Anshika Kapoor, Sarah Ebenezer, Ruju Talati, Abbey Jin, Yash Khanna, Sean Gratton
Published : July 20, 2026
Abstract
Introduction: Artificial intelligence (AI) has been increasingly revolutionizing the field of neurology, yet formal training in AI tools, appraisal, and limitations remains limited in residency education. Neurology training should reflect growing advancements in technology, as this gap risks inappropriate reliance or underutilization of AI tools. Objectives: To evaluate how a structured, three-session artificial intelligence curriculum for neurology residents improves AI literacy, clinical confidence, and critical appraisal to use AI tools in neurologic practice. Methods: We conducted a prospective quality improvement study involving 9 neurology residents at University Health through an AI curriculum consisting of 3 sessions that incorporated case-based learning, ethics discussions, and exposure to real-world AI tools (Appendix A). Using the Kirkpatrick model, we aimed to measure outcomes at 3 levels: reaction, learning, and behavior. To do this, we administered a pre- and post-survey (Appendix B) using 5-point Likert-scale items and open-ended questions. These were analyzed with the Wilcoxon signed-rank test for paired ordinal responses and qualitative analysis. The main themes assessed included confidence and knowledge, adoption and attitude, and risk awareness. Qualitative measures included perceived effectiveness of the lecture series and self-reported, intended behavior change. Results: Residents demonstrated improvements in confidence and knowledge, as well as in attitudes and adoption. The highest level of improvement was observed in confidence and knowledge, with a 0.96-point change (p value = 0.004). In terms of resident attitudes, 89% agreed that the lecture series improved their understanding of AI in neurology, the content was relevant to their practice, and that they are more prepared to engage with AI. Intended adoption and attitudes showed a favorable but non-significant trend, with a 0.17-point improvement (p=0.078). Finally, risk awareness had a non-significant, minimal improvement from 0.06 points. Conclusion/Discussion: Implementation of a structured, discussion-based AI curriculum significantly improved neurology residents’ confidence, knowledge, and skills in AI tools. This project demonstrates that targeted educational interventions can address critical gaps in AI literacy.