Agentic education · Learning science

Actionable companion · Cognitive load & attention

Learning with agency: Biosensor Real-Time Affective Analytics in Virtual and Mixed Reality Medical Education Serious Games: Cohort Study

AI-assisted editorial implementation guide, separate from the original research. Proposed actions are not claims or recommendations attributed to the source authors. No independent peer review of this guide is asserted.

Read original paperSearch paper agentCC BY 4.0

Evidence and context

Title and abstract screening; stage labels describe reported settings, not verified numerical ages. Design labels require full-text appraisal. Reported stage(s): Higher education. Health-professions learning must be distinguished from K–12 instruction.

Original authors’ abstract, reproduced under the linked source license:

Background The role of emotion is crucial to the learning process, as it is linked to motivation, interest, and attention. Affective states are expressed in the brain and in overall biological activity. Biosignals, like heart rate (HR), electrodermal activity (EDA), and electroencephalography (EEG) are physiological expressions affected by emotional state. Analyzing these biosignal recordings can point to a person’s emotional state. Contemporary medical education has progressed extensively towards diverse learning resources using virtual reality (VR) and mixed reality (MR) applications. Objective This paper aims to study the efficacy of wearable biosensors for affect detection in a learning process involving a serious game in the Microsoft HoloLens VR/MR platform. Methods A wearable array of sensors recording HR, EDA, and EEG signals was deployed during 2 educational activities conducted by 11 participants of diverse educational level (undergraduate, postgraduate, and specialist neurosurgeon doctors). The first scenario was a conventional virtual patient case used for establishing the personal biosignal baselines for the participant. The second was a case in a VR/MR environment regarding neuroanatomy. The affective measures that we recorded were EEG (theta/beta ratio and alpha rhythm), HR, and EDA. Results Results were recorded and aggregated across all 3 groups. Average EEG ratios of the virtual patient (VP) versus the MR serious game cases were recorded at 3.49 (SD 0.82) versus 3.23 (SD 0.94) for students, 2.59 (SD 0.96) versus 2.90 (SD 1.78) for neurosurgeons, and 2.33 (SD 0.26) versus 2.56 (SD 0.62) for postgraduate medical students. Average alpha rhythm of the VP versus the MR serious game cases were recorded at 7.77 (SD 1.62) μV versus 8.42 (SD 2.56) μV for students, 7.03 (SD 2.19) μV versus 7.15 (SD 1.86) μV for neurosurgeons, and 11.84 (SD 6.15) μV versus 9.55 (SD 3.12) μV for postgraduate medical students. Average HR of the VP versus the MR serious game cases were recorded at 87 (SD 13) versus 86 (SD 12) bpm for students, 81 (SD 7) versus 83 (SD 7) bpm for neurosurgeons, and 81 (SD 7) versus 77 (SD 6) bpm for postgraduate medical students. Average EDA of the VP versus the MR serious game cases were recorded at 1.198 (SD 1.467) μS versus 4.097 (SD 2.79) μS for students, 1.890 (SD 2.269) μS versus 5.407 (SD 5.391) μS for neurosurgeons, and 0.739 (SD 0.509) μS versus 2.498 (SD 1.72) μS for postgraduate medical students. The variations of these metrics have been correlated with existing theoretical interpretations regarding educationally relevant affective analytics, such as engagement and educational focus. Conclusions These results demonstrate that this novel sensor configuration can lead to credible affective state detection and can be used in platforms like intelligent tutoring systems for providing real-time, evidence-based, affective learning analytics using VR/MR-deployed medical education resources.

What kind of evidence is this?

Intelligent tutoring & adaptive learning · Design requires full-text appraisal

Stage: Higher education. Screening labels describe reported context and study-design language. They are not a risk-of-bias rating or proof of effectiveness for an age band.

Study methods · source passage

A total of 11 healthy participants took part in our study after providing written informed consent. The participants included medical students (4 participants), medical school postgraduates (3 participants), and neurosurgeons (4 participants). The participants were informed that they would take part in 2 VP scenarios. The 2 scenarios were (1) a simple emergency response VP scenario that is familiar to all medical students beyond the third year of their studies, implemented on a simple web-based platform (OpenLabyrinth), and (2) a neuroanatomy-focused case regarding the ascending and descending pathways of the central nervous system, designed for a specialized neuroanatomy lecture at the graduate/postgraduate level and implemented in the MR HoloLens platform. The choice of scenario for the MR platform (Microsoft HoloLens) was forced to be this specific resource, as it was the only one that was available in scenario-based format. Both scenarios were medical narrative games that were guided by player choice. In the case of the web-based VP game, the users chose their responses from a multiple-choice panel on each of the VP game’s pages. In the MR VP game, the users visualized the actual case through the HoloLens device, seeing and tacitly manipulating the relevant parts of the anatomy as they were described to them by narrative text. Selections in this modality were conducted by hand gestures in relevant locations of the actual anatomy that was presented to the user.

Limitations and uncertainty · source passage

Despite the overall promising results, this work contains some inherent limitations, mostly linked with the novelty of its aim and scope. A significant limitation is the small and diverse group of subjects that were used for the pilot run of the multimodal signal acquisition configuration. It must be emphasized that given the sampling rate of the biosensors, every 1 minute of recording provided 15,360 samples (256 Hz × 60 s) of EEG per channel, 60 samples (1 Hz × 60 s) of HR, and 240 samples (4 Hz × 60 s) of EDA. This high-density data throughout the study allowed for rather definitive biosignal results to be extracted on a per-participant basis. While the extensive data set gathered from participants provided credible results for this feasibility study, obviously there is the need to expand the participant sample in order to explore personalization and statistical verification challenges.

Design the system before the pilot

Document the learner model, mastery rule, hint sequence and adaptation signal. Compare adaptive routing with the same content in a non-adaptive sequence. Check whether early errors lead to persistent low-expectation pathways.

  • Limit simultaneous instructions and unnecessary dialogue. Compare task performance and a suitable mental-effort measure; easier-feeling work is not automatically better learning.
  • Measure engagement separately from achievement. Check whether novelty-driven participation persists and whether it translates to independent performance.

A 90-day district pilot

  1. Days 1–15 · Appraise. Have the curriculum lead and a teacher read the methods and limitations. Record the study design, sample, comparator, outcomes and evidence gaps. Define a local question that the study can help inform.
  2. Days 16–30 · Design. Choose one disciplinary task. Require a first attempt, a documented AI consultation, a critique of the output and an independent explanation. Align the comparison group and scoring rubric with the same objective. Name an accountable owner and set the baseline, success threshold and budget before implementation.
  3. Days 31–60 · Test. Start with a small voluntary cohort, preserve an accessible non-AI route and require teacher review. Use synthetic or de-identified examples; approve data handling and tool terms through district processes.
  4. Days 61–90 · Decide. Blind-score reasoning and accuracy on a new task without AI. Distinguish model examination scores, student perceptions and actual learner gains; only the last directly tests educational benefit. Compare with baseline, document uncertainty and stop or revise if learning worsens, access is unequal or critical errors persist.

The board decision brief

Report the learning problem, evidence quality, participating population, total cost including review time, measured benefit, errors, access differences and remaining uncertainties. Make a dated recommendation to stop, revise or expand. A positive perception survey alone does not establish learning improvement.

Attribution and reuse

Panagiotis E Antoniou; George Arfaras; Niki Pandria; Alkinoos Athanasiou; George Ntakakis; Emmanouil Babatsikos; Vasilis Nigdelis; Panagiotis Bamidis. Biosensor Real-Time Affective Analytics in Virtual and Mixed Reality Medical Education Serious Games: Cohort Study. JMIR Serious Games (2020-09-02). https://doi.org/10.2196/17823

Companion prepared 2026-10-30. Original authorship and DOI are preserved. The abstract is an attributed source extract; the implementation plan is new AI-assisted editorial material, offered under the same CC BY 4.0 license. This guide has not undergone independent peer review.