Multimodal Neural-Signal-Driven Instant Price Sensitivity Prediction in Virtual Reality Duty-Free Store Scenarios

This work proposes a multimodal neuro-signal fusion model for predicting consumer price sensitivity within immersive virtual reality (VR) duty-free shopping settings. Electroencephalography (EEG), eye-tracking, and galvanic skin response (GSR) signals were concurrently collected and synchronized at the millisecond level to elicit subconscious responses within 500 ms of every price cue. A cross-modal attention-based transformer model was trained and tested on a data set from 120 international travelers, producing 8,640 price-exposure events. The multimodal model resulted in an F1-score of 0.85 and ROC-AUC of 0.91, a 14% improvement over the best single-modal baseline performance. Ablation studies verifiec complementary roles across modalities, for which EEG frontal-theta power and gaze dwell on price tags were established as principal predictors. In an online A/B test imbedded in the VR store, adaptive discount policies activated by the model elevated conversion rates by 9.7 percentage points and lifted average basket sizes by 98.9 while holding profit margins. These findings prove not only the technical possibility of multimodal neural-driven price models but also their potential for creating real economic impacts in cross-border stores.