To VR or not to VR: Assessing Cybersickness in Navigational Tasks at Different Levels of Immersion

Luis Eudave ^* Martín Martínez

• University of Navarra, Pamplona, Spain

The advent of affordable, high-quality virtual reality (VR) devices has revolutionized experimental research in cognitive psychology and neuroscience, offering more immersive and naturalistic environments for studying skills like spatial navigation. However, the increased incidence of cybersickness in VR may compromise its advantages, necessitating appropriate tools to assess this phenomenon and understand its impact on experimental outcomes. Despite the growing use of VR in research, there is a lack of consensus on the most effective methods for measuring cybersickness across different experimental modalities and over time. Here, we compared two cybersickness assessment tools: the widely-used Simulator Sickness Questionnaire (SSQ) and the more recent Cybersickness in VR Questionnaire (CSQ-VR). Using a maze navigation task, we examined how cybersickness is modulated by modality (Desktop vs. VR) and habituation (decrease in cybersickness between morning and afternoon sessions) in a gender-balanced, young Spanish sample (n=26) with a within-subjects design. We also investigated potential predictors of cybersickness related to the task and individual differences. Our results demonstrate high internal consistency for both tools, performing particularly well in VR, and SSQ showing higher reliability in Desktop conditions. Robust mixed factorial analyses revealed small to moderate effects of modality (VR >Desktop) and habituation in both SSQ and CSQ-VR scores.Robust regression analyses indicated that SSQ scores were predicted by modality and habituation, while CSQ-VR scores were mainly predicted by modality and VR experience. These findings highlight that: (1) both SSQ and CSQ-VR are reliable tools for assessing cybersickness during navigation tasks, especially in VR; (2) VR-induced cybersickness decreases with task repetition without apparent impact on performance; and (3) other performance and individual differences do not predict cybersickness. Our study provides valuable insights for optimizing VR task design in experimental settings, contributing to the broader field of VR-based research methodology in cognitive science.