Immersive virtual reality (VR) enables controlled tests of how emotional context distorts subjective time, yet findings remain mixed. Fifty-seven adults explored three five-minute VR environments (tranquil garden, neutral room, threatening sewer) while eye tracking recorded pupil diameter. Baseline stress was assessed before VR exposure. After each environment, participants estimated elapsed time and rated mood, calmness, and arousal. Repeated-measures ANCOVA (baseline stress as covariate) and linear mixed-effects models converged on a valence gradient: garden intervals were overestimated, sewer intervals underestimated, and neutral estimates were intermediate. Pupil diameter showed the complementary pattern (sewer > garden > neutral) and covaried with the magnitude of temporal distortion. Baseline stress did not predict time estimates and did not moderate valence effects. Results indicate that affectively distinct VR contexts reliably shift perceived time and elicit convergent physiological and self-report responses. These findings inform emotion-aware design of immersive experiences and digital-wellbeing safeguards that support safe, sustained immersion.

翻译：沉浸式虚拟现实（VR）为研究情绪情境如何扭曲主观时间提供了可控的测试环境，但现有研究结果仍存在分歧。57名成年人在探索三个五分钟的VR环境（宁静花园、中性房间、威胁性下水道）的同时，通过眼动追踪记录其瞳孔直径。在VR暴露前评估了基线压力水平。每个环境体验结束后，参与者需估计经过的时间，并对情绪、平静度和唤醒度进行评分。重复测量协方差分析（以基线压力为协变量）与线性混合效应模型的结果均指向一个情绪效价梯度：花园环境的时间间隔被高估，下水道环境的时间间隔被低估，中性环境的估计值介于两者之间。瞳孔直径呈现出互补的模式（下水道 > 花园 > 中性），并与时间扭曲的幅度存在共变关系。基线压力未能预测时间估计，也未调节情绪效价效应。结果表明，具有不同情感效价的VR环境能够可靠地改变感知时间，并引发一致的生理与自我报告反应。这些发现为沉浸式体验的情感感知设计以及支持安全、持续沉浸的数字健康保障措施提供了参考。