Conventional performance metrics fail to offer high-resolution evaluation of learning and memory during training tasks; the P300 component of the event-related potential (ERP) is a promising tool for enhancing the assessment of training quality in virtual environments, but this technique is yet to be investigated. A driver training simulator and scenario were developed to explore the capability of the P300 for this purpose. A user study was conducted with 32 participants divided into two groups objectively determined by driving performance scores, thus enabling observations of the P300 response to be equated to varying levels of learning and memory. Participant electroencephalogram data were recorded during the procedure, which was postprocessed to filter and extract ERPs to capture neural responses to specific events in the virtual training scenario. These were combined to produce a result for each participant, which was then grand averaged to create an overall ERP for each group. Across the eight electrode sites, statistically significant differences were found between the grand average waveforms of the two groups, with high memory retention producing significantly greater peak-to-peak amplitude (U = 9.00, p = 0.045), peak latency (U = 0.00, p < 0.001), and positive area (U = 13.00, p = 0.05) of the waveform than low memory retention. The evidenced relationship between the P300 response and working memory in this context suggests that it has the potential for monitoring learning and memory in stimulus-driven virtual training systems.
- EEG, virtual environments, learning, memory, driving simulator, human-computer interaction