Applications of mixed reality technology in neurosurgery: a narrative review

Neurosurgery demands precise localization and real-time visualization, yet conventional workflows are limited by 2D imaging, alignment drift from intraoperative brain shift, and “hand-eye separation” when looking away to monitors. Mixed reality (MR) blends virtual reality’s (VR) immersive 3D with augmented reality’s (AR) in situ overlays to register multimodal data as holographic models within the surgeon’s field of view, promising improved spatial understanding and guidance. We conducted a structured, non-systematic search of PubMed, Web of Science (WOS), and China National Knowledge Infrastructure (CNKI) for publications from 2015 to 2025. English and Chinese keywords included “mixed reality,” “augmented reality,” “head-mounted display,” “neurosurgery,” “neuroanatomy,” and “neurosurgical navigation,” combined with Boolean operators. We focused on MR studies related to neurosurgical clinical practice, education, telemedicine, or doctor-patient communication that were available as peer-reviewed full texts. AR-only series were retained when the authors explicitly described head‑mounted, in‑situ overlays that are conceptually overlapping with MR platforms, whereas VR‑only and non‑neurosurgical work were excluded. Records judged relevant by the authors were synthesized qualitatively by application domain, with particular attention to randomized controlled trials (RCTs) and multicenter studies where available. The search yielded 2456 records; after deduplication and screening, 118 studies were included. Most addressed clinical practice—preoperative planning and intraoperative navigation (n = 84, 84/118, 71.2%); education (neuroanatomy and skills training) accounted for 33 studies (33/118, 28.0%); telemedicine was rare (n = 1, 1/118, 0.8%); and no neurosurgical doctor-patient communication studies were found. Clinically, MR supports patient-specific 3D visualization, surgical corridor simulation, and in-view navigation that reduces gaze shifts; several reports describe registration or targeting errors typically in the 1–3 mm range, comparable to conventional navigation systems. In education and training, MR enhances spatial cognition, engagement, and task performance, including flipped-classroom integrations; however, most evidence is short-term and small-sample. Key cross-cutting limitations include alignment vulnerability to brain shift and repositioning, restricted field of view/comfort of current head-mounted displays (HMDs), operating-room environment interference, and a predominance of small single-center studies. MR has been explored across the neurosurgical continuum—from education and preoperative planning to intraoperative guidance and remote collaboration—offering intuitive, in-situ three-dimensional (3D) support. However, most current evidence remains early-stage and derived predominantly from small single-center studies. Broader adoption will depend on technical maturation, multicenter comparative research, and health-economic evaluation.