Learning Seriously with VR: An evaluation on the effect of a low-fidelity pre-training on cognitive workload and task performance in a virtual reality sailing task

Abstract

The rise of virtual reality-based (VR) technology has prompted the gradual adoption of VR-based simulation training in the maritime industry. However, there is limited research on strategies to improve the effectiveness of VR training through cognitive workload reduction, especially with the pre-training strategy that is rooted in Cognitive Load Theory (CLT). This study investigates whether a low-fidelity mock-up pre-training prior to a VR sailing simulation improves task performance and reduces cognitive workload. Two groups of participants (n = 16) completed a VR sailing task. The experimental group received a low structural fidelity mock-up pre-training before performing the VR sailing task, while the control group did not. Task performance was measured by a performance index including time, distance, and error rate. Cognitive workload was subjectively measured with the NASA Task Load Index (TLX) and Klepsch et al.’s (2017) differentiated cognitive load (DCL) questionnaire. Results indicated that pre-training did not significantly improve performance, though performance increased significantly over the three VR sailing trials. NASA TLX revealed no significant reduction in cognitive workload, and the DCL measure indicated that pre-training increases intrinsic cognitive load. We attribute these findings to limited skill transferability between pre-training and VR simulation, individual differences in spatial and psycho-motor abilities, factors related to pre-training design, target cognitive workload type, and other factors. On a positive note, results show the benefit of repeated exposure to the task leading to automaticity-based performance improvement. Several limitations and future directions are also discussed.