Zeitgebers-Based User Time Perception Analysis and Data-Driven Modeling via Transformer in VR

TL;DR

The paper investigates how VR zeitgebers—light color, music tempo, and task factors—shape subjective time perception and its relation to user experience. It introduces Relative Subjective Time Change (RSTC) and the Time Perception Modeling Network (TPM-Net), a CNN-Transformer fusion that analyzes multimodal physiological data paired with zeitgebers to predict time perception categories. Key findings show that task factors and red light/slow tempo cues drive time underestimation, and that RSTC correlates with improvements in presence and engagement. TPM-Net achieves high predictive performance (Accuracy 86.11%, Macro-F1 83.10%, UAR 79.45%) and outperforms several baselines, highlighting potential for VR design, therapy, and training applications.

Abstract

Virtual Reality (VR) creates a highly realistic and controllable simulation environment that can manipulate users' sense of space and time. While the sensation of "losing track of time" is often associated with enjoyable experiences, the link between time perception and user experience in VR and its underlying mechanisms remains largely unexplored. This study investigates how different zeitgebers-light color, music tempo, and task factor-influence time perception. We introduced the Relative Subjective Time Change (RSTC) method to explore the relationship between time perception and user experience. Additionally, we applied a data-driven approach called the Time Perception Modeling Network (TPM-Net), which integrates Convolutional Neural Network (CNN) and Transformer architectures to model time perception based on multimodal physiological and zeitgebers data. With 56 participants in a between-subject experiment, our results show that task factors significantly influence time perception, with red light and slow-tempo music further contributing to time underestimation. The RSTC method reveals that underestimating time in VR is strongly associated with improved user experience, presence, and engagement. Furthermore, TPM-Net shows potential for modeling time perception in VR, enabling inference of relative changes in users' time perception and corresponding changes in user experience. This study provides insights into the relationship between time perception and user experience in VR, with applications in VR-based therapy and specialized training.

Figures (6)

• Figure 1: Experiment Process.
• Figure 2: Example of the sequence of one trail. Each trail is divided into four sections. Participants start timed tasks after the countdown has ended.
• Figure 3: Overview of the data-driven TPM-Net architecture. The model consists of five stages. In the multimodal input stage, the BW Seq includes $\alpha$, low$\beta$, high$\beta$, $\theta$, and $\gamma$, while the HS Seq includes stress, awareness, drowsiness, and meditation. The Zei Seq represents the corresponding experimental variables. $d_{eeg}$, $d_{bw}$, $d_{hs}$, $d_{hr}$, $d_{zei}$ denotes the data channels corresponding to each sequence. Details of the temporal encoding module can be found in Figure \ref{['fig:convolution']}.
• Figure 4: Temporal encoding module.
• Figure 5: Subjective time perception results with means and data test. (a) results of subjective time estimation with/without task. (b) results of subjective time estimation with task in color group. (c)results of subjective time estimation with task in music group. (d) results of subjective time estimation without task in color group. (d)results of subjective time estimation without task in music group. (a'), (b'), (c'), and (d') are the results of corresponding relative time estimates to baseline as described above.