A novel neurophysiological approach to evaluate the impact of virtual training on skills acquisition

Objective evaluation of operational training is challenging, especially when comparing physical training with virtual reality (VR). Behavioural and subjective measures capture performance and workload but provide limited insight into neurocognitive adaptation. We propose and validate an EEG-derived Cognitive Training Index (CTI) integrating multiple training-sensitive neurometrics to (i) track training progression and (ii) compare VR- versus physical-based training during subsequent real-environment execution.&#xD;&#xD;Approach: Participants performed three repetitions of a realistic confined-space industrial maintenance procedure with Entrance, Rolling shutter, and Pipe replacement phases. Execution time, procedural errors, and NASA-TLX were collected after each trial. Wearable EEG was recorded continuously; six neurometrics were computed from cleaned EEG using band-specific Global Field Power over predefined electrode sets. CTI was obtained via Principal Component Analysis at trial level, then analysed segment-wise. Temporal specificity was tested with phase-randomized surrogate analyses.&#xD;&#xD;Main results: Execution time decreased across trials (F = 14.82, p < .001, η²p = 0.35), errors decreased (F = 10.31, p < .001, η²p = 0.27), and NASA-TLX declined (F = 11.94, p < .001, η²p = 0.30). CTI increased with training (F = 18.21, p < .001, η²p = 0.39) and was higher in learning-relevant phases (Entrance and Pipe replacement) while remaining near baseline in Rolling shutter; surrogate analyses confirmed CTI exceeded null expectations. Repeated-measures correlations linked higher CTI to fewer errors (R = -0.42, p < .01) and lower NASA-TLX (R = -0.38, p < .01), with no association with execution time (R = 0.18, p = .12).&#xD;&#xD;Significance: CTI provides a psychologically grounded, multivariate EEG marker that tracks training progression and aligns with operationally meaningful outcomes (accuracy and perceived workload) rather than speed, enabling objective comparison of cognitive adaptation across training modalities.