Spellcasters began as an entertainment VR game in which teams of five wizards duel by drawing gestures with a magic wand to cast spells. The team translated that gesture-drawing mechanic into a serious game for stroke rehabilitation, specifically targeting hemiparesis spells now require therapeutic gestures to cast, and the curriculum is customizable per patient. Two companion artifacts extend the system: (1) a clinician-facing sandbox for authoring custom spell gestures matched to a patient's post-stroke motor profile, and (2) a companion clinician dashboard web application with co-designed data visualizations driven by the game's telemetry, so therapists can track progress, edit in-game goals between sessions, and interact with patients.
Post-stroke rehabilitation requires high-volume, often tedious gesture practice for which intrinsic motivation is famously hard to sustain the same dynamic the lab targets across its serious-games portfolio. At the same time, a therapy game built around one fixed set of "right" gestures can't accommodate the wide variability of post-stroke motor profiles. Spellcasters explores what happens when an engaging entertainment mechanic is re-authored for therapy, with clinicians retaining authoring power over the gesture set itself.
Participatory design sessions and user studies were conducted with 14 physical and occupational therapists to develop the customizable-gesture sandbox and clinician dashboard. Earlier related stroke-rehabilitation work was conducted at UCSD (Cosman Lab) using bodystorming, Wizard-of-Oz, and object-theater methods. Lead Researcher: Jared Duval | Co-authors: Rutul Thakkar, Delong Du, Kassandra Chin, Sherry Luo, Aviv Elor, Magy Seif El-Nasr, Michael John | Publication: Duval, J., et al. Designing Spellcasters from Clinician Perspectives: A Customizable Gesture-Based Immersive VR Game for Stroke Rehabilitation. ACM Transactions on Accessible Computing (TACCESS) 15(3), Article 26, September 2022. https://doi.org/10.1145/3530820 | Demo video: https://youtu.be/zXLUNgFvpgg | Companion app Figma prototype: https://www.figma.com/file/0XGZRuvnULvt0BdOcQKvdl/Spellcasters-HighFi-Prototype