Autonomous shuttle buses offer significant potential for improving public transportation, enhancing traffic safety, and reducing environmental impact. However, their successful implementation depends not only on technological development but also crucially on user acceptance. While previous studies have primarily investigated acceptance based on surveys of individuals without direct experience, especially in critical traffic situations, our study addresses this gap through a simulated driving test incorporating such scenarios. Extending the UTAUT2 framework, we integrate objective NeuroIS indicators to propose a psychophysiological mediation model. We captured both autonomic arousal (electrocardiogram, galvanic skin response) and cognitive stress appraisals (Perceived Stress Scale, NASA-TLX, single item scales) to examine their impact on the behavioral intention to use autonomous shuttles. Data from N = 104 participants were evaluated using Partial Least Squares Structural Equation Modeling (PLS-SEM). The results show that Social Influence and Perceived Usefulness are the primary drivers of usage intention, whereas Trust and Perceived Risk showed no significant effect in the simulated environment. Furthermore, the analysis supports a significant negative impact of perceived Cognitive Reaction on acceptance. Crucially, physiological indicators demonstrated a divergent mediation pattern: while heightened cardiac activation marginally increased cognitive stress, electrodermal activity actively reduced it. This points towards a “eustress” or heightened task-focus phenomenon, mitigating perceived psychological burden during critical rides. These findings highlight the importance of social validation and emotional responses in the acceptance of novel mobility technologies. The study demonstrates that subjective cognitive load, shaped by specific physiological arousals, acts as a pivotal barrier to acceptance, providing a robust foundation for future real-world exposure studies.