Authors: Nazlı Tokatlı¹, Muhammed Tayyip Koçak², Gürkan Göztepeli², Onur Kılıç², Yakuphan Bilmez², Halis Altun², Fatma Nur Kömür³
¹Department of Computer Engineering, Istanbul Health and Technology University, Istanbul, Türkiye
²Department of Software Engineering, Istanbul Health and Technology University, Istanbul, Türkiye
³Department of Audiology and Speech Disorders, Institute of Health Sciences, Marmara University, Istanbul, Türkiye / Department of Otolaryngology, University of Michigan, Ann Arbor, Michigan, USA
Background: In audiology education, a persistent gap remains between theoretical instruction and hands-on clinical experience, particularly in regions where access to a wide range of patient profiles and modern training facilities is limited. To address this challenge, we proposed AICSA, an AI-powered virtual reality (VR) audiometry training software designed to enhance preclinical competence and educational engagement among audiology students, trainees, and professionals who need practice.
Methods:Â AICAS was developed using Unity for the interactive 3D environment, FastAPI for backend communication, and three fine-tuned LLaMA3 language models to enable natural language interaction. The system includes a virtual clinic setting with functional audiometry, AI-driven virtual patients, and a variety of clinical scenarios. Based on various clinically validated auditory pathologies and contextual variables, patient behaviors and cases were created using AI modeling. Behavioral avatars were designed to simulate patient responses during a pure tone audiometry procedure.
Results: In preliminary data, the integration of fine-tuned LLaMA3 models resulted in 92% consistency in generating dialogues, 86% flexibility in handling various student interactions, and a 7% increase in perceived realism and teaching value, creating a more engaging and functional learning environment for users. 82% of audiology student participants indicated that the platform was user-friendly and easy to navigate to achieve desired functions.
Conclusion: AICAS provides training in pure tone audiometry by simulating real-life patient scenarios and clinical environments in VR. It combines theory and practice effectively, increasing users’ confidence and readiness for clinical challenges. AICAS sets a precedent for integrating AI and VR technologies into healthcare education in audiological practice, laying the foundation for broader adoption in global audiology curricula.
