Immersive VoiceSpace: Development and Pilot Testing of a Virtual Reality System for Contextualized Vocal Training

Behavioral voice therapy is effective for many voice disorders, yet gains achieved in the clinic often fail to transfer reliably to everyday communication. Motor learning frameworks suggest that transfer improves when practice preserves the sensory and contextual demands of the target environment. Immersive virtual reality (IVR) may help address this gap by embedding voice tasks within ecologically relevant visual-spatial contexts. The purpose of the present study was to evaluate the feasibility of a prototype IVR platform, Immersive VoiceSpace (IVS), and to examine vocal behavior across graded task conditions as an early step toward contextualized voice training.

Seventeen adults participated: 10 vocally healthy speakers (typical) and 7 individuals with dysphonia (atypical). Participants completed a functional speech task under four conditions: Baseline, Normal, Effortful, and Calling. The three IVS conditions were performed in a virtual restaurant in which a waiter avatar's behavior was modulated by speaker output through graded listener distance, relative vocal intensity thresholds, and timeout parameters. Sound pressure level (SPL) and mean speaking fundamental frequency (mean f₀) were analyzed and statistically examined using linear mixed-effects models. Participants also completed a feasibility questionnaire addressing usability and interaction, immersion and realism, engagement and perceived benefit, and comfort and safety.

Task condition (IVS Levels) significantly affected both SPL, F(3,48) = 33.94, P < 0.001, and mean f₀, F(3,45) = 17.63, P < 0.001, with graded increases across levels. Group effects were also significant for SPL, F(1,15) = 15.76, P = 0.001, and mean f₀, F(1,15) = 13.29, P = 0.002, with atypical speakers producing lower overall values. The IVS Level × Group interaction was not significant for SPL, but was significant for mean f₀, F(3,45) = 3.94, P = 0.014. Feasibility ratings ranged from moderate to excellent, with highest scores for comfort and safety. No adverse events or cybersickness were reported.

IVS was feasible, well-tolerated, and the findings provide preliminary evidence that vocal output differed across graded IVS Levels. Further multisession study is needed to determine its clinical utility.