The efficacy of adaptive training in immersive virtual reality for a fine motor skill task

Abstract Immersive virtual reality (IVR) offers novel and promising ways of continuously adapting training content and difficulty to the individual trainee, thus paving the way for an improved fit between training content and trainee needs. The objective of the present study was to investigate the efficiency of utilising self-efficacy and performance measures to continuously adapt training content to the individual trainee. Using a preregistered, between-subjects experiment, 130 participants were randomly assigned to receive virtual training content that was either based on behavioural measures at the beginning of the study (fixed training) or that continuously adapted to the behaviour of the trainee (adaptive training). The results revealed no significant difference between the groups for either performance or self-efficacy, suggesting that the additional development required for fully adaptive training may be unwarranted in some cases. Further research should investigate when the additional complexity of adaptive training is outweighed by enhanced efficiency. That said, results revealed an overall beneficial effect of IVR-based training. However, while IVR had an overall positive effect on performance, transfer was only observed to a limited extent. Specifically, participants improved in both accuracy ( d = −0.41) and speed ( d = −0.43) on a virtual performance test, while performance on a real equivalent (i.e., transfer of skill) showed improved accuracy ( d = −0.25) but reduced speed ( d = 0.17). In other words, the study demonstrates that performance measures in IVR should not necessarily be expected to transfer to similar tasks outside IVR without a potential loss in performance, emphasising the need for future studies to include measures of skill transfer when investigating IVR-based training.