Learning Anatomy in Virtual Environments: Cybersickness and Cognitive Load

Abstract Body : Introduction & Objective
Anatomy education has historically relied on cadavers and plastic models, but technological advances have created affordable and functional alternative teaching tools. For example, virtual reality (VR) employs a head-mounted display (HMD) that immerses the user in a digitally created stereoscopic environment for viewing three-dimensional (3D) models, while autostereoscopic devices provide an HMD-free stereoscopic view, and 3D printing uses various techniques to create replicas of specimens with appropriate tactile affordances. However, the comparative education efficacy and physiological side effects, such as levels of cybersickness and cognitive load, which may influence learning are not well studied. This research compares the effectiveness of VR, autostereoscopic screens (AS), and 3D-printed physical models (PH) in learning anatomical landmarks with a specific focus on how cybersickness and cognitive load may play a role in learning in these different environments.

Materials & Methods
61 undergraduate students with no prior anatomy education were randomized and counterbalanced into one of three groups. All groups had to learn three bones (pelvic bone, zygoma, and calcaneus) in three modalities (VR, AS, and PH). Participants had four minutes to learn 10 bony landmarks on one of the three bones. While learning, they simultaneously completed a dual-task paradigm, which measured their reaction time as a proxy for cognitive load exerted. Participants then completed an untimed, recognition-based test on a 3D-printed bone, identical to the one they had just learned, with randomized landmarks and a word bank. Afterwards, participants completed the Simulator Sickness Questionnaire (SSQ), which measures cybersickness, and the Paas scale and Simulation Task Load Index, both of which measure aspects of cognitive load. Stereo Fly Test and Operation Span Task, which measure stereoacuity and working memory capacity, respectively, were also completed and used as covariates.

Results
A one-way analysis of covariance (ANCOVA) assessing for test performance differences across learning modality, and accounting for stereoacuity and working memory capacity, found no significant effect of modality (p > 0.05). A repeated-measures multivariate ANCOVA (MANCOVA) assessing for cybersickness differences across learning modality, and accounting for stereoacuity and working memory capacity, found a significant effect of modality on the overall weighted SSQ score (p < 0.05; VR > AS > PH) and the SSQ oculomotor subscore (p < 0.05; VR > AS > PH). Another repeated-measures MANCOVA assessing for cognitive load differences across learning modality, and accounting for stereoacuity and working memory capacity, found no significant effect of modality (p > 0.05).

Conclusion
Although learning modality has a significant effect on participants’ self-reported cybersickness levels, it does not affect skeletal anatomy learning or cognitive load outcomes.

Significance/Implication
Equivalent learning across modalities is encouraging for anatomy education programs looking to employ virtual teaching and/or learning methods in the classroom as an alternative to physical specimens. However, more research is needed to establish these effects in more complex material beyond the skeletal anatomy learned here. Further, investigating determinants (e.g., sound, fidelity, etc.) and opportunities to mitigate the effects of cybersickness remain a priority.