Exploring Virtual Reality for Breastfeeding Education: a Study on Interaction Methods and Proxy Fidelity

Breastfeeding provides essential health benefits for both infants and mothers, yet difficulties with positioning and latching are among the main causes of early cessation. Traditional resources, such as brochures, videos, and mobile apps, offer limited opportunities for hands-on practice with spatial tasks, hindering skill acquisition and confidence. Virtual Reality (VR) offers a promising alternative by delivering immersive, interactive, and safe environments to practice complex tasks. Consumer-level VR, however, relies on one-size-fits-all interaction devices, such as standard controllers or hand tracking, which may lack natural gesture fidelity. While VR-based breastfeeding training has been explored, the impact of input device fidelity on usability, cognitive load, and presence remains largely unknown. This thesis presents a VR breastfeeding application developed with Unity and the Meta XR SDK for the Meta Quest Pro. The application has two phases: a tutorial to learn proper infant holding and latching, and a gamified latching phase, where the virtual infant signals hunger and users must respond correctly within a limited timeframe. Immediate visual and auditory feedback guides each step. Cartoon-style avatars and environments were chosen to avoid the uncanny valley, based on consultation with a lactancy expert, ensuring comfort, reducing anxiety, and supporting full embodiment. Full-body tracking via the Meta Movement SDK allows natural upper-body motion, and iterative refinements based on user feedback improved navigation and avatar representation, enhancing engagement. Two exploratory within-subject studies examined design factors. The first compared hand tracking, VR controllers, and a high-fidelity medical infant doll. The second compared a high-fidelity medical doll with a low-fidelity rolled towel. Usability, cognitive load, and presence were measured via the System Usability Scale (SUS), NASA Task Load Index (TLX), and the Presence Questionnaire (PQ). Condition order was counterbalanced using Latin Squares, and participants provided qualitative feedback. Results show that interaction modality significantly affects user experience: controllers led to lower usability, higher cognitive load, and reduced presence, whereas hand tracking and proxy use improved engagement and performance. Proxy fidelity did not significantly influence outcomes, suggesting high-fidelity models are not strictly necessary. Qualitative feedback confirmed that controller-based interaction felt less natural, while low-fidelity proxies were sufficient for practice. Small sample sizes (n = 4 and n = 8) limit generalizability but highlight the importance of intuitive, natural interaction over visual realism in VR training. Overall, this work demonstrates that VR can provide a scalable, engaging, and safe platform for maternal education, enabling learners to practice breastfeeding skills with immediate feedback. The results provide practical guidance for VR breastfeeding training, emphasizing user-centered interaction, accessibility, and cost-effectiveness, while supporting public health objectives. Immersive, guided, and iterative VR training has the potential to improve skill acquisition, increase confidence, and enhance maternal and infant health outcomes.