AI-Wired Projects

​

Our AI-Wired Projects aim to build the next generation of AI-informed, expert-driven clinical decision support tools for aphasia rehabilitation. Although the Western Aphasia Battery–Revised (WAB-R) is widely used to classify aphasia, clinicians consistently report that individuals with the same aphasia type show highly diverse language profiles. As a result, treatment selection often varies significantly across clinicians, limiting consistency, transparency, and scalability in clinical practice.

​

This research program addresses that challenge by examining how experts make treatment decisions and how AI systems can learn from these decisions to support evidence-based, individualized care. In this project, licensed speech-language pathologists and aphasia researchers review anonymized WAB-R profiles and select treatments using a structured checklist. Their decisions are compared with both rule-based AI recommendations and supervised machine-learning models trained on expert responses.

​

We analyze inter-rater agreement, human–AI alignment, and machine-learning feasibility to determine how expert clinical reasoning can be transformed into structured, explainable, and scalable frameworks for rehabilitation planning. This work lays the foundation for developing AI-supported treatment recommendation systems that enhance decision-making for clinicians, support early-career practitioners, and improve the reliability of aphasia care.