Abstract:[Objective] To evaluate the effect of virtual reality (VR) and mixed reality (MR) based digital teaching on the training of postgraduate students with professional degrees and standardized training doctors in neurosurgery, and to determine whether this approach improves learning quality without increasing the cognitive and time burden on trainees. [Methods] A prospective controlled study was conducted at a single tertiary hospital. Twenty postgraduate students with professional degrees and standardized training doctors in neurosurgery, who had completed systematic courses on basic neurosurgical anatomy and clinical theory and had prior experience of clinical observation, were randomly allocated to an observation group (VR/MR digital teaching, n=10) and a control group (conventional teaching, n=10). The two groups were comparable in age, educational background, duration of clinical practice, and baseline neuroanatomical knowledge. In addition to conventional training, the observation group received three-dimensional reconstruction based on patient multimodal imaging, virtual surgical simulation, and MR-based real-virtual fusion intraoperative assistance. Teaching effects were assessed before and after the intervention using a unified closed-book examination, a practical test of three-dimensional anatomical cognition with blinded image datasets, a double-blind surgical planning score rated independently by two senior surgeons, and a 5-point Likert satisfaction questionnaire; the time required to complete each task was also recorded. Normality of quantitative data was assessed by the Shapiro-Wilk test; between-group comparisons of normally distributed data were performed with independent-samples t tests, with Cohen''s d and 95% confidence intervals reported. Satisfaction ratings, being ordinal data on a 5-point scale, were expressed as medians (interquartile ranges) and compared with the Mann-Whitney U test. Bonferroni correction was applied across the six outcomes. [Results] Compared with the control group, the observation group showed greater improvement in theoretical scores (10.70±1.16 vs 7.90±1.37 points, t=4.93, P=0.0001, d=2.21), higher surgical planning scores (91.20±1.32 vs 89.60±0.97 points, t=3.10, P=0.0062, d=1.39), a higher accuracy of anatomical structure identification (91.50%±1.27% vs 89.60%±1.58%, t=2.97, P=0.0083, d=1.33), higher satisfaction ratings [5.0 (4.0-5.0) vs 4.0 (3.2-4.0) points, U=86.0, P=0.0028], and a shorter time for anatomical identification (270.50±6.43 vs 316.00±11.01 s, t=-11.29, P<0.0001), with a reduction of 45.50 s (95%CI -53.97 to -37.03). No significant difference was found in the time spent on surgical planning (16.60±1.17 vs 16.10±1.97 min, P=0.4992), indicating that the quality gains were not achieved at the expense of additional planning time. All five differences other than surgical planning time remained statistically significant after Bonferroni correction. [Conclusion] VR/MR based digital teaching improved theoretical mastery, three-dimensional anatomical cognition, surgical planning ability, and trainee acceptance, and substantially reduced the time required for anatomical identification without increasing the time burden of surgical planning. Trainee acceptance was high, indicating that the approach is feasible within routine residency training. Whether immersive exposure carries risks such as cybersickness or visual fatigue, whether the gains persist after training, and whether the observed effects generalise to other specialties remain to be established. Given the small sample size and the single-centre, short-duration design, these findings should be regarded as preliminary evidence supporting further evaluation of digital reform in neurosurgical education.
