In recent years, there has been growing interest in the use of virtual reality (VR) and augmented reality (AR) in anatomy education. This follows a decline in overall anatomy teaching time, as well as reductions in cadaver-based teaching and available laboratory hours (Drake et al., 2009). This decline in cadaver usage has created a demand for new and efficient tools to teach anatomical relationships.

Immersive and interactive 3D models can be displayed to students using VR/AR technologies. In VR, a user is fully immersed and feels present in a virtual environment. In AR, virtual objects (like anatomical models) are superimposed onto the user’s view of the real world. Models can be displayed on an individual basis through devices including mobiles, desktops and head-mounted devices (HMDs), and to wider audiences with stereoscopic projectors and screen-based AR systems.

Several institutions have trialled and used VR/AR technology as part of a multimodal, multimedia approach to anatomy teaching and learning. However, uptake of the technology is still limited. This can be due to concerns regarding the purchase cost and software development, as well as uncertainty over usability or efficacy of the technology. This brief review aims to examine the effectiveness of VR/AR in anatomy teaching, both in terms of academic results and student perceptions.

A literature review was conducted using PubMed, Scopus and Google Scholar. Search terms included “virtual reality”, “augmented reality”, “mixed reality”, “head mounted device”, “anatomy” and “students”. To maintain temporal validity, papers prior to 2013 were excluded. The last search was conducted on 29^th June 2018.

The characteristics of the included studies (n=9) are summarised in Appendix 1.

Head Mounted Devices

HMDs aim to create a sense of presence by realistically immersing the user in a shifting virtual environment displayed on head-mounted screens.

In this review, all three studies consulted found no significant differences in the post-test results of HMDs compared with the control alternatives: 2D images (Ekstrand et al., 2018), online textbooks (Stepan et al., 2017), AR tablets, and conventional tablets (Moro et al., 2017). Significant improvement between pre- and post-tests was observed in one study (Ekstrand et al., 2018) but not in another (Stepan et al., 2017). Adverse side effects were measured using a questionnaire in one study, and they found the VR group using HMDs experienced higher levels of side effects - including significantly more dizziness, blurred vision and general discomfort - than the AR tablet or tablet groups (Moro et al., 2017).

In student perception surveys, HMDs received significantly higher ratings than online textbooks for motivation, attention, confidence, satisfaction, interest and usefulness, but there was no significant difference for relevance or ease of use (Stepan et al., 2017). No significant differences were found between HMD and AR/tablet groups, though all the teaching methods had high ratings on how understandable or enjoyable the technology was and its usefulness as a supplementary teaching method (Moro et al., 2017). Compared to 2D images, HMDs received higher ratings for improved confidence and future use in teaching (Ekstrand et al., 2018).

Stereoscopic Projectors and Screens

Stereoscopy is the display of different images to each eye, creating a sense of depth. This can enable more effective perception and understanding of spatial relationships in anatomy. It is utilised in presentations to large audiences through specialist projectors and filters, requiring the users to wear polarising glasses (Kockro et al., 2015; Cui et al., 2017).

Students using a 3D stereoscopic model or 2D images both significantly improved in results between pre- and post-tests (p<0.001), with 3D stereoscopy achieving significantly higher post-test results (p=0.0033) (Cui et al., 2017). In the 3D stereoscopic group, students with low spatial ability performed as well as those with high (p=0.0899), but in the 2D group, they had significantly lower scores than those with high spatial ability (p=0.0259) (Cui et al., 2017). However, a study (Kockro et al., 2015) on students who received a 2D or stereoscopic 3D presentation on neuroanatomy reported no significant differences in post-test results (Kockro et al., 2015). Alternative research (de Faria et al., 2016) found that students instructed using 3D stereoscopic or 3D non-stereoscopic models performed significantly better in post-tests than those instructed using standard 2D images (p<0.05). However, there was no significant difference between stereoscopic and non-stereoscopic scores (p>0.05) (de Faria et al., 2016).

All three of the studies reviewed received positive feedback (de Faria et al., 2016; Cui et al., 2017) and ratings (Kockro et al., 2015; Cui et al., 2017) in student perception surveys. For one of the studies (Kockro et al., 2015), this is included significantly higher ratings from the 3D stereoscopic group for spatial understanding, future teaching, effectiveness and enjoyability than the 2D group (p<0.01) (Kockro et al., 2015). However, there were some reports of eyestrain and software limitations (de Faria et al., 2016).

AR Magic Mirror

The AR Magic Mirror is a system where the user stands in front of a screen displaying their mirror image, whilst virtual images are superimposed onto them on the screen (see Figure 1). In anatomy teaching, these are often radiological or anatomical sections. Through real-time tracking, the images align with the user to represent the precise location of structures within the user’s body (Kugelmann et al., 2018).

In a study (Kugelmann et al., 2018) using the AR Magic Mirror to teach anatomy, a survey was administered to investigate student perceptions of teaching anatomy with the AR Magic Mirror. They received strong positive responses, notably for its usefulness in active learning and 3D understanding. Criticisms of the software revolved around technical problems and suggestions of more content (Kugelmann et al., 2018).

Figure 1: Students using AR Magic Mirror (Kugelmann et al. 2018).

AR Magic Book

Virtual objects can be superimposed onto the pages of an AR Magic Book when it is viewed through the screen of a suitable device, like a tablet or mobile. The physical book contains markings that devices can identify, track and augment content to. This content includes 3D models, animations and videos (Küçük, Kapakin and Göktaş, 2016).

Students studying anatomy with an AR Magic Book performed significantly higher than those using a dissection video in a post-test (p<0.000) and had a significantly lower score distribution (Ferrer-Torregrosa et al., 2016). This is consistent with a study into a mobile AR (mAR) Magic Book, where the post-test performance of students using an mAR Magic Book was significantly higher than those using 2D images (p<0.05) and was associated with significantly lower cognitive load (p<0.05) (Küçük, Kapakin and Göktaş, 2016).

Positive student perceptions of the AR Magic Book were found in both studies (Ferrer-Torregrosa et al., 2016; Küçük, Kapakin and Göktaş, 2016). This included a survey on motivation, autonomous learning and 3D understanding (Ferrer-Torregrosa et al., 2016), and interview responses regarding effectiveness, motivation and interest (Küçük, Kapakin and Göktaş, 2016).

Reported research studies suggest that VR and AR are at least as effective as traditional anatomy teaching methods. However, this generalisation may be unreliable due to the large heterogeneity between studies. As there is a limited number of studies available, this review included various teaching methods that may not be directly comparable. Another confounding factor is the control method used. Whilst there is value in comparing VR/AR to one simple teaching method like 2D diagrams or textbooks, it can be more insightful to also compare it with multiple methods, like standard 3D physical and virtual models. This is demonstrated by an included study (de Faria et al., 2016) in their conclusion that results varied between 2D and 3D models, but not necessarily between stereoscopic and non-stereoscopic (de Faria et al., 2016). Considering the often-high costs of VR/AR hardware acquisition and software development compared with standard online 3D models, this would be valuable information for prospective institutions.

Results may have also been influenced by large variation in study topic, participant degree courses, prior knowledge of students, timing of the study (whether used to introduce or reinforce a topic), form of study (individual or group) and study time. Differences in model quality and features (such as the level of interactivity) may have also impacted results. Furthermore, for many studies, results may be unrepresentative due to small sample sizes.

The number of questions, as well as their difficulty and the presence of any repeats, could have also impacted results. Results offered restricted insight into learning as studies largely used short MCQ tests that reflect factual recall over other forms of understanding, such as that of 3D relationships. For more translatable and insightful findings, all studies would need pre-, post- and long-term recall tests.

Despite the wide heterogeneity, there is a consensus between studies regarding positive student perceptions of VR/AR. In all studies, surveys revealed high levels of student satisfaction - generally higher than control alternatives - demonstrating the benefits of VR/AR for student enjoyment and engagement. It can be observed that VR/AR had a notable impact on spatial understanding, motivation and interest. Other reported areas of benefit included autonomous learning, confidence, cognitive load and overall student perceptions.

These survey results may be subject to sampling bias as participants were often volunteers and so more likely to be motivated and engaged. There may also be author bias, as evaluated systems were often developed by or for the paper authors. The novelty of this new technology may also affect positive perceptions. Moreover, the five-point Likert scales generally used could produce response bias, whilst yes/no responses (Küçük, Kapakin and Göktaş, 2016) are rather restrictive. Study responses are also limited when there are no reported control responses for comparison.

Studies indicate that VR and AR are effective and capable resources for anatomy teaching in terms of both academic achievement and student satisfaction, supporting the use of VR and AR as supplements to current teaching methods. However, further research into the depth and longevity of learning would aid understanding of the extent of their long-term impact on academic achievement. Future investigations into cost-effectiveness would also be helpful.