Using immersive virtual reality with autistic pupils: moving towards greater inclusion and co-participation through ethical practices

Abstract

Purpose

The potential of head mounted displays based virtual reality (HMD-based VR) for autistic groups has been well documented. However, the deployment and application of this technology, especially in schools, has been extremely limited. One of the main criticisms in this field has been the lack of involvement from practitioners in research on educational approaches for autistic populations and the gap between research and practice in real-life settings.

Design/methodology/approach

This conceptual article focuses on our research in a UK-based special needs school that sought to examine the effects and potential use of VR-HMDs, while seeking to establish best practices for safe and ethical application using this technology. This draws upon ethical and participatory research guidance, including British Educational Research Association and Autism Participatory Research.

Findings

The authors make recommendations on planning and implementing a participatory, safe and ethical approach to researching the use of VR-HMDs in special needs schools and engaging with the priorities of autistic children and young people and their teachers.

Originality/value

This conceptual article provides an initial first consideration of ways we can better include autistic people and their views in research that is with and about them. The value in this will mean we are able to better support autistic groups moving ahead using VR HMD-based technologies. Without this paradigm shift and including autistic people (and their stakeholders) the field might continue to build initiatives around medical-based models of disabilities rather that what the community need/want.

Citation

Newbutt, N. and Bradley, R. (2022), "Using immersive virtual reality with autistic pupils: moving towards greater inclusion and co-participation through ethical practices", Journal of Enabling Technologies, Vol. 16 No. 2, pp. 124-140. https://doi.org/10.1108/JET-01-2022-0010

1. Introduction

1.1 Autism

Autism is a lifelong neurodevelopmental condition that affects how a person perceives, communicates and interacts with the world. Current data suggest that as many as 1 in 54 children in the United States of America are on the autism spectrum (Russell et al., 2014) while other studies suggest a figure of between 1 in 68 to 1 in 100 in the general population (Özerk, 2016). However, these rates and numbers do not fully capture either the individual nature of the condition or the different types of support required to enable this group to achieve their potential. Autism is characterised by significant and lasting differences (compared to typical development) in social communications and interaction, restricted and repetitive patterns of behavior, interests or activities and sensory perception and responses (APA, 2013). Revisions to the DSM criteria (in 2013) for diagnosing autism conditions specifically referring to sensory issues. Within the DSM-5 diagnosis criteria there are specific mention of sensory issues including: “odd responses to sensory input” and “hyper-or hypo-reactivity to sensory input or unusual interest in sensory aspects of environment, such as apparent indifference to pain/heat/cold, adverse response to specific sounds […]” (APA, 2013, p. 50). In addition to sensory issues, autistic people can sometimes struggle to make connections with how others are feeling; sometimes referred to as Theory of Mind (ToM). This is a broad, complex, and multifaceted construct, defined as the ability to attribute mental states (beliefs, desires, intentions) to oneself and to others, making it difficult to explain and predict behaviour (Rosello et al., 2020). As such there are very specific mention to the ways autistic people can and do experience the world around them. However, and despite these characteristics (often defined by a diagnosis and diagnostic criteria) there has been a move towards researchers focussing on the strengths of autistic individuals in recent years (Courchesne et al., 2020; Lee et al., 2020). For example, research has highlighted parental perspectives of strengths of their autistic children as being: (1) technical abilities; (2) attention to detail; (3) trustworthiness; (4) kindness; (5) expertise in a specific area; (6) creative skills; (7) strong sense of morality; and (8) loyalty (Lee et al., 2020). Based on this, specific educational programs and clubs have been designed (Jones et al., 2021) to help target and focus on these strengths. The field of research involving technology for autistic groups also builds on this strength-based approach (Bölte et al., 2010). Not only can technology support autistic people in very specific ways, it can also speak to many key autistic-strengths (Frauenberger, 2015). This is something we return to in the next section.

1.1.1 Technology and autistic groups

Growth in autism research necessitates corresponding attention to autism research ethics, including ethical and meaningful inclusion of diverse participants (Cascio et al., 2020).

The last fifty years has seen an emergence and growth of research interested in designing, developing and evaluating technologies targeted at autistic children (Newbutt, 2019). To date much of this research has targeted the role and use of a range of technologies to ameliorate what have been seen as core deficits. This includes areas such as social functioning, or to reduce “problematic behaviours” (Spiel et al., 2017, 2019). This narrowing of research focus has reinforced the view of autism as a medical deficit in need of “correction”, resulting in a lack of design technologies that are geared towards the interests and needs of autistic children; even taking their views into account. This approach has meant that autism research has historically been done “on” autistic participants, rather than “with” them (Oliver, 2013). This lack of ambition to provide agency or participation for autistic children in technology-based research, reflects a wider disconnect between the autism community (autistic people, their parents and family members) and academics and funding bodies (Woods and Waltz, 2019). A study by Pellicano and colleagues (Pellicano et al., 2013) found a large gap in priorities between research that was valued most highly by autistic people (promoting education, independence and support) and the most highly funded and studied (genetic and biological). Dissatisfaction was also expressed by autistic groups over the levels of engagement in the research process and disagreements about the interpretation of research findings. This runs the risk of ignoring and excluding the lived-experiences of autistic individuals and failing to engage them in co-production of research that has relevance to their lives.

Steps to address these issues have been made more recently in the area and development of technologies (from screen-based multimedia to immersive head-mounted displays). This field has tackled issues of design and development methodologies (Schmidt et al., 2008) in addition to ethical approaches to research in this area (Parsons et al., 2020). The area of design methodologies (broadly speaking) has begun to move away from researching about autistic people, moving towards an approach that includes autistic people in the design, development and evaluation of technologies (Fletcher-Watson et al., 2019; Walmsley et al., 2018). This move not only encapsulates the design and development of technologies, but more broadly includes the voices of autistic people in the process of research.

However, this is not to underestimate the challenges and difficulties such an approach can have and is well summarised by Parsons and colleagues (Parsons et al., 2020), who suggest:

co-construction within the technology and autism field may create special challenges because it is not always clear what the best answers or processes are and, therefore, who has the necessary expertise: technology tools develop and change swiftly, as do the expectations from the contexts of their use.

Furthermore, for people working in the field of autism, education and technology, there appears to be limited specific guidance to help inform them about ethical issues (by which we mean inclusive and informed research). Although the British Educational Research Association (British Educational Research Association, 2018) provides relevant guidance for ethical considerations in educational research, including targeted advice for working with vulnerable populations, it does not provide specific guidance for autistic populations. Furthermore, the autism specific guidance that has been developed (Pellicano et al., 2017; Gowen et al., 2020) contains generalised advice that does not take account of specific issues and challenges for researching technology in schools with autistic young children.

1.1.2 Co-participation in research with young autistic people

Within the area of research and autism, there have been calls to better include and involve autistic groups in research that is situated with and for them (Jivraj et al., 2014; Chown et al., 2017; MacLeod et al., 2014). Indeed, recent research suggest that the published evidence-based for “participatory autism research is rare” (den Houting et al., 2021, p. 148). Building on this, Fletcher-Watson et al. (2019) suggest that: “participatory research methods connect researchers with relevant communities to achieve shared goals” (p. 943), however recognise that there remains “evidence of community dissatisfaction” when it comes to autistic communities and research in this domain. This is a theme also located in findings presented by Parsons et al. (2020), who found that research with autistic groups and technology could be more enlightened by involving the whole community (verbal and minimally verbal; children and adults). Additionally, Parsons and colleagues found that there were several issues that meant autistic people report not feeling included or are not able to meaningfully participate in research. Issues were identified as: who directs the research, who knows best, who are the gatekeepers, and how is technology designed to be inclusive. Through engaging with, and placing central, the stakeholders in their research (autistic people, teachers, parents), Parsons et al. (2020) also indicated that the “processes and experiences of engagement and participation were valued as much as (if not more than) any possible, more formalised, indications of ‘outcomes’” (p. 223). This starts to suggest that not only is co-participation valuable for developing meaningful research for autistic people, but also that the experience of being involved is seen as equally valuable. Moreover, Parsons and colleagues found that autistic people reported wanting to be more involved in the decision-making and design of technology with/for them. Therefore, and taken together, co-participation of research in the area of autism needs to include a greater sense of shared goals, include a wider-range of people from the autism community, ensure meaningful inclusion, enable autistic-driven research aims, and locate ways to engage them in the process of research. Therefore, one of the aims of this article is to document and provide ways we have achieved some of these co-participation criteria in our research with virtual reality-based technologies for autistic groups. We do not suggest they are complete, but due the dearth of work reported this in the field, we proffer this as a way for future researchers and practitioners to engage and empower autistic young people.

1.1.3 Virtual reality and head-mounted displays for autistic groups

Research in the field of autism and virtual reality has existed since 1996, when some of the first research was reported by Strickland et al. (1996). This work, along with others looking at a range of technologies for autistic groups (Anderson et al., 2019) helped to provide an early interest in this research area. The coupling of evolving and innovative technologies with increasing rates of autism (Yeargin-Allsopp et al., 2003) have led to a growth in research and interest. While earlier work in this area included virtual reality head-mounted displays (Strickland et al., 1996), research soon turned to more affordable, deployable and manageable virtual reality technologies (VRTs) such as virtual environments (Parsons et al., 2004), virtual worlds (Kandalaft et al., 2013), collaborative virtual environments, virtual video modelling (Smith et al., 2014) and single-user environments (Irish, 2013). These VRTs were all investigated with autistic groups as they are considered to hold tremendous promise for individuals with autism due to the fit and natural affinity to computers and technology for this population; the ability of VRTs to simulate real-world situations and contexts in safe and predictable ways; and finally, the ability to control, shape, and tailor VRT interfaces to participant needs and local contexts (Parsons, 2016).

More recently, and since 2016, one specific technology increasingly utilised with autistic groups is that of head-mounted display based virtual reality (HMD-based VR) (McCleery et al., 2020; Newbutt et al., 2016; Parish-Morris et al., 2018; Schmidt et al., 2019). Key reasons for this format of technology being used and evaluated with autistic groups include: (1) the “fit” between a virtual environment (i.e. not real, but life like), leading to a reduction in cognitive overload (Newbutt et al., 2017); (2) a heightened sense of presence leading to possible transference of knowledge from the virtual to real (Parsons, 2016); (3) providing a safe and predictable environment in which to test experiences out (Parsons et al., 2004); and (4) an environment where real-life consequences can be removed (Parsons and Cobb, 2011). First investigated with autistic children during the 1990s by Strickland and colleagues (Strickland et al., 1996), HMD-based VR research has seen a significant resurgence and interest in research since the mid-late 2010s (Bradley and Newbutt, 2018). These studies reflect the dichotomy found in much VR research with autistic groups. Namely, using HMD-Based VR in experimental conditions to better understand core cognitive, social and neurological differences (Mundy et al., 2016) or using this technology to support the learning of skills that will have relevance and practical use in “real world” contexts (Adjorlu et al., 2017). However, when related to children, issues of confusion and susceptibility (real vs virtual) and social exclusion can arise with the application of VR technologies. Studies carried out by Segovia and Bailenson (2009), have found heightened levels of feeling presence in VR; so much so that users believed their VR-based experiences to be real the following week. While on the one hand this demonstrates the unique power and potential of VR, on the other, it highlights that users can view their VR-based experiences as “real”. This finding should be contextualised with the very young children they worked with (i.e. 4–5 year olds), and so older children might not report similar experiences. Ethical issues related to the psychological, physical and social impact of this technology therefore must be carefully considered (Kenwright, 2018). In addition to ethical concerns, issues of stability and young people persist in this field. A recent study by Tychsen and Foeller (2020) found that young children (aged 5–6) tolerated a fully immersive 3D virtual reality game play without inducing significant post-VR postural instability or maladaption. Taken together the findings here suggest that the psychological and social impacts are greater than those of physical experience. In fact, Tychsen and Foeller go so far as to state that their results “suggest that manufacturer warnings against the use of this technology in young children are overstated” (p. 157). Therefore, in the work reflected upon in this manuscript (and reported in Newbutt et al., 2020), we developed safe and ethical working practices with younger autistic children using HMD-VR.

Research in the field of HMD-based VR holds promise and potential applied in the right ways with autistic populations. However, work in this field has tended to neglect the views and input of autistic users in many parts of their work. Areas specifically neglected are: (1) formation of research questions; (2) involvement in evaluation; (3) experiences of wearing HMD hardware (given the sensory specific nature related to an autism diagnosis); (4) preferences to the type of HMD; (5) feedback on responses to virtual reality experiences (i.e. interfaces); (6) auditory experiences best suited to autistic children; and (7) where and how they would see this technology best impacting their lives.

To-date there is very limited, if any, specific guidance around the safe and ethical working practices associated with these wearable technologies. This in itself is problematic and raises serious questions about the future of a field that relies on the experiences of autistic people using HMD-based VR, along with teachers and researchers working in this area. This is reinforced by a recent systematic literature review by Van Mechelen et al. (2020), who suggest that: “while ethics is frequently mentioned [in research studies], the literature remains underdeveloped in a number of areas including […] the reporting of formal ethical approval procedures, and the extent to which design and participation ethics is dealt with”.

This leads us to identify two key aspects that reflect the gaps found in this field as it relates to technology adoption in schools and involvement of pupils in this process:

1. How can we fully involve autistic children and in the deployment of HMD-based VR in classrooms?

2. What steps should be taken by teachers and researchers to ensure safe and responsible use of HMD-based VR with autistic children?

Having gained experience of working ethically in this space for several years (see, Newbutt et al., 2016, 2017, 2020; Bradley and Newbutt, 2018) we are in a position to attempt to answer these questions and share where we see opportunities, best practice and limitations to research in this area -- as it informs potential practice and applied use in classrooms. We use a recent case study on HMD-based VR in schools as context for an analysis of the processes and ethical concerns that need to be considered when using this technology with autistic children.

2. Case study

2.1 Case study location and context

We undertook an examination of a range of HMD-based VR across four schools (two primary, one secondary and one special) in the United Kingdom. Working with a range of pupils, teachers and leadership teams we explored the views and experiences of autistic and non-autistic pupils (n = 43) on the uses and preferences of this technology within their setting. Across the four schools we set up and deployed three types of HMDs (with associated VR experiences) and systematically considered research questions located around the preference of HMD-VR, types of HMDs, possible negative effects, and how the pupils felt HMD-based VR might be most useful for them. Before summarising our findings, we next turn to the specifics of the project in addition to contextualising the case study school in this article.

For the purpose of the current article, we focus on a specific school from our previous study (Newbutt et al., 2020). We do so because this was the school we initially worked with when developing our ethical approval processes between the University and the user groups (autistic children and teachers). We refer to the school as Appletree School (a nom de plume). Appletree is a Special School located in the South-West of England and supports 143 pupils from the age of six through nineteen in a setting that is specifically designed to support the unique nature of autistic pupils. Plate 1 highlights the space and set up of the HMD-based VR equipment used in this case study and the school.

We carefully examined the views of pupils on the potential and preferences of three HMDs (shown in Plate 2). The HMDs were chosen to represent a range of devices that schools may consider for use in their settings, differentiated by cost and functionality. This included: (1) low-tech (cardboard device); (2) mid-tech (Class VR); and (3) high-tech (HTC Vive). For our study we worked with n = 12 pupils of which 3 were female and 9 were male. Table 1 provides an overview of the school demographic in addition to the pupils we worked with.

2.2 Case study ethical processes

We next turn to the ethical procedures, processes and timeline used in the case study as we considered a range of issues and perspectives in our research. We will explore and present: (1) process maps and timelines; (2) ethical review documentation; (3) detailed checklists we used when introducing and implement the HMD-based VR in the school; and (4) post-study dissemination.

As a starting point, we were committed to involving members of both the school and autism community as partners in the research process (Gowen et al., 2020). Appletree School worked closely with the researchers to establish practices and an environment that would be appropriate when researching with the pupils and staff. In addition, we discussed the proposed project with an autistic adult who acted as a mentor and supported the formulation of research questions, ethical procedures and ethical review documents that lead to institutional review approval. Peter (a pseudonym) was an autistic adult who kindly offered input based on his experience of working with technology and in technology-based environments. Peter was known to the lead author as an employee of a charity supporting autistic people. He was chosen as he worked with, and used, technology in his professional and personal life. As Peter was working with a charity that supported autistic individuals through technology innovation, including VR, he held very specific and pertinent knowledge in this arena. This coupled with his autistic insight, provided a lens and set of views that informed the research questions, software choice, and ethical insights to using HMD-based VR with this specific user group. This input was critical not only to ensure we were considering the right questions to ask, but also who should be answering them and the best way to achieve this, e.g. in the school setting rather than a research lab. Possible challenges relating to the use of VR-HMDs by autistic pupils were discussed with the mentor and these considerations consequently fed into ethical procedures and documentation.

The following section will firstly detail the timescales that the research took followed by processes we planned to utilise (including part of the institutional ethical review process). This will conclude with the protocols we followed in practice, and post-project dissemination. Providing a clear map of our work and associated timescales and details, will enable us to reflect and offer insights that currently are not available to others working in this field. Figure 1 provides details into the “stages” we followed from our idea conception (fed by gaps in literature and previous work reported in Bradley and Newbutt (2018) through to undertaking research in Appletree School. Evident are the number of intentional steps taken to form a relationship with the school, engaging autistic input, ensuring parents saw/understood the proposed work (to enable informed consent), and careful approach taken to technology acclimation.

As articulated in Figure 1, the institutional ethical process falls out of a set of meetings including the school (leadership teams) and an autistic mentor (Peter). These meetings enabled a set of ethical review documents to be drafted for full institutional review. When preparing this document, details of the school could be shared along with key insights and information provided by Peter. Within the ethical review documentation, we articulated our plans to work safely, with intent and in collaboration with the school and teachers. We were guided by the British Educational Research Association (British Educational Research Association, 2018) criteria in a range of areas. These included:

1. Ensuring clear materials for our participants (classified as vulnerable) and their parents/carers to enable informed consent/assent. These were developed in conjunction with teachers and the Trust ethical review team. This ensured we “fully explored ways in which [participants] can be supported to participate with assent in the research”. This involved the use of information leaflets (that were predominantly visual) and video-based materials to enable a complete idea about what both parents and children were “signing up for”.

2. When designing and executing our research plans, we ensured we were clearly “put[ting] participants at ease and avoid[ing] making excessive demands on them” (p. 19). This, as BERA highlights, is important as “the more vulnerable the participants, the greater the responsibilities of the researcher for their protection”. Therefore, building in an intentional process to our work represented reassurances in the ethical review process. Similarly, we designed materials including questionnaires to be suitable for the pupils we worked with; recognising the visual nature of our learners (Benton et al., 2012). This meant the use of cartoon faces/emojis (representing expressions as to how they might feel) alongside a ranking scale (from 1–4). See Figure 2 for an example of both the co-designed questionnaire (co-designed with teachers and autistic children) and the information leaflet design.

Providing the above details in the ethical review documentation, including participation consent forms (coupled with the information leaflets), aimed to articulate the research benefits to a lay audience. In addition, the consent forms included formal (required) information presented clear and easy to understand terminology/language. In the ethical review documentation, the intentional process was articulated as a “map” of activities; outlining the stages and intent at each step. The purpose of this was to share details related to how we were planning to work in practice; ensuring the safety of the pupils. This diagram, Figure 3, is taken from the ethical review document, and highlights our intentional process when planning our research.

As seen in Figure 3 our methodical process emphasised ensuring the health and safety of the pupils. It did so by being explicit in our observations of possible side effects, signs of being stressed or worried and followed by checks with individual pupils throughout the day. This process further extrapolated into a set of aspects we followed in practice. This was not part of the ethical review process, but a set of steps we designed to ensure consistency and intentionality during the research. This is illustrated in Figure 3. Notable is the inclusion of processes for slowly acclimatising the pupils from lower and less immersive equipment (part 1), leading up to fully immersive experiences (part 2). This was in addition to the intentional steps to support “checkpoints” offering the chance for pupils to end the experience/study should they wish. Both of these points and stages proved vitally important when working with our autistic pupils in supporting possible sensory and physical reactions to the HMD-based VR. Within our work, and even before the ethical review process, Peter was able to point out this issue; helping to identify this aspect in addition to sharing ideas and ways we could overcome this in our research.

Two other key features of this case study are important to highlight. Firstly, that Part 2 in Figure 4 involved the HTC Vive HMD device which included a cabled headset (cabled to a laptop). This meant we needed to be aware of trip hazards and the pupils becoming tangled up in the cable as they navigated the 3D spaces wearing the headset (some standing and movement was required to interact with the environment). While not specific to autistic groups this is something generally an issue with this format/type of HMD. Second, the changes in the fidelity of 3D graphics between the low- and high-tech devices meant the level of immersion and realism increased; meaning that sense of presence would become more powerful and perhaps stressful and/or overwhelming for some pupils (Ostrolenk et al., 2019). Thus, we needed to be fully aware of these important issues alongside the weight of the devices. As the pupils progressed using a cardboard device to the HTC Vive device, they also experienced an increase in weight on their heads and covering their faces/eyes.

Completing the process map of this work (Figure 1) was feedback and dissemination we provided after the data collection, analysis and initial findings. This included a presentation of findings to the teachers along with a study newsletter sent to the parents and children. The study newsletter contained two-pages of information related to the study, the outcomes, what was learned and next steps. The Trust also published a lay summary, in conjunction with the researchers, of the outcomes from the work on their website for dissemination to the teaching and wider autistic-community within and beyond the wider school network (Figure 5).

Dissemination also led to meaningful strategic impacts at the school that we had not originally envisaged. The findings and outcomes from the research were used to feed into the school's five-year strategic plan in addition to feeding policy on the adoption and use of HMD-based VR in the school and a set of teacher training events to share best practice and to ensure appropriate training (helping to bridge the technology to adoption gap).

3. Discussion and lessons learned

We have provided case study details outlining a participatory-based approach to developing research involving autistic children reporting their experiences of using HMD-based VR in their school. While we did not experience issues related to pupils withdrawing from the study, feeling overwhelmed or motion sickness (that can sometimes be associated with HMD use (Sharples et al., 2008)), we did identify several key considerations that should help to inform practitioners attempting to work in this space. We also suggest that our careful and staged working process, including the voices of autistic people from the outset, meant we were able to limit possible issues and concerns when deploying this technology.

It is also important to highlight the differences in the HMD technology we deployed in our work. As highlighted in Plate 2 we used a range of tools, ranging from low-tech (low immersion) to high-tech (high immersion). This was to support a gradual and staged exposure to HMDs and the virtual experiences therein to the users. However, the differences in experiences were substantial. From simply tours of a moon surface (via a smart phone and cardboard HMD) through to a fully immersive fun fair with interactive components, motor/hand control inputs and six-degrees of freedom (i.e. rotational movement around x, y, z axis and translational movement moving forward or backward, left or right, and up or down).

We next outline our recommendations in five areas: (1) research in and with schools; (2) enable and ensure autistic participation; (3) research tools and measures; (4) HMD-VR safety and for whom VR is perhaps most suitable; and (5) suggestions for applied use of HMD-based VR in schools.

3.1 Research in schools and with schools

The research to practice gap within the field of autism education research has been well documented (Guldberg, 2017). This case study highlighted how developing key relationships with the staff, pupils and leadership at Appletree School was critical in translating the findings into practical and meaningful outcomes. The continued use of the HMD-based VR, the ways in which it is being used and the development of a trust-wide policy to support this shows the impact of the research in this setting. By placing autistic and teacher voices and views at the centre of our work, we were able to provide powerful messages and insights to leadership and teacher populations of this school in a way that enabled them to “see” the potential of this specific technology with their pupils. This, we argue, is a major step forward in this field that for so long has attempted to develop relationships with schools with limited success (Konza, 2011; Newbutt et al., 2020; Southgate et al., 2019).

3.2 Enable and ensure participation of autistic pupils

While we made conscious efforts to include all stakeholders in our study, we feel that this field would be better informed by involving the staff and pupils in more meaningful ways at all stages of the research process (Pellicano et al., 2017). Firstly, by involving them in the formation of research priorities and involving them in discussions around technology use in their school. Secondly, collaborating more fully on the research questions and identifying key themes that both staff and pupils and aspects they were interested in investigating. Finally, through involving staff and pupils more fully in the evaluation and dissemination of the research will help to ensure collaboration on the interpretation and meaning of the results. This should lead to communicating findings to the right audience and in the right way. This is something we attempted, but more focus should be applied to this part of any research in this area.

3.3 Research tools and methods

The suitability of using pre-existing inventories and questionnaires, not designed explicitly for autistic groups need to be considered carefully and involve autistic groups from the outset. This was very evident from our research where we intended to utilise a pre-existing inventory (Sense of Presence Inventory; SoPI) but noticed the pupils had issues comprehending the questions and challenges in providing answers. Therefore, we chose not to utilise the SoPI, but rather designed our own questionnaire that elicited responses to sense of presence, experience using the HMD, and the potential of HMD-based VR in their lives. We do not propose this as a reworked or validated instrument, but rather a way to enable our participants to express their experiences of HMD-based VR in our study. The questionnaire was co-designed with Appletree school for this study, and was tailored for the age and special nature of this group (Benton et al., 2012; Parsons, 2015). This was supported by feedback from the pupils who were happy to complete the co-designed questionnaire but asked to stop completing the SoPI.

3.4 HMD-VR safety and use in a school for autism

A review of previous studies in this field (i.e. HMD-based VR) by Bradely and Newbutt (Bradley and Newbutt, 2018) revealed that only 4 studies have previously reported either capturing or measuring or asking about negative effects. This indicates that negative effects and issues closely related to health and safety (all contextualised within ethics), have yet to be placed centrally or reported explicitly in this field. This coupled with a lack of autistic input and engagement, means researchers could be exploring and evaluating these technologies without being appropriately guided by important ethical principles; teachers should be aware of this.

This finding and important as many HMD manufactures (i.e. Oculus) recommend use with people over the age of 13 years old. However, by working carefully, and in conjunction with users and their teachers, we successfully deployed HMDs with children as young as six. This does not however mean that others will find the same. Careful and meaningful health and safety protocols should first be established; some of which we report here.

Within our own research we found that the larger of the HMDs (i.e. HTC Vive) provided some challenges to the input mechanisms associated with this device (motor and coordination control) for the very youngest pupils (6–8 year olds). It also appeared to be too heavy for them to use. For the very youngest pupils in schools (<8 years old) we suggest additional measures might be considered to ensure balance and input fidelity. One suggestion is to have users of this age sat while using the device or limiting their experiences/use to lower-tech options reported earlier (i.e. ClassVR or cardboard/smartphone). In addition, and in line with the findings of Tychsen and Foeller (2020) we reported limited, if any, negative physical impact of using the HMD on the young people we worked with (Newbutt et al., 2020).

3.5 Suggestions for using HMD-based VR in schools

Based on the research presented in this article, we have made several suggestions for how researchers might conduct research with HMD-based VR with autistic groups in schools. Work within this field can, and should, only advance with meaningful inclusion of autistic people and placing their input central to research endeavours. Linked to this are questions of how this technology can be used in school to provide benefits and opportunities for autistic pupils. We have alluded to several opportunities, primarily around: the “fit” between VR; a heightened sense of presence (leading to possible transference of knowledge from the virtual to real); providing a safe and predictable environment in which to test experiences; and an environment where real-life consequences can be removed. We have also identified that autistic children suggest using VR may support them in relaxing/feeling calm, being able to explore somewhere virtually before visiting in the real world, and to develop learning opportunities in school (Newbutt et al., 2020). This translates to a range of potential applications that should include employment contexts (experiencing interviews and transitioning to work), independent living (managing living alone), job specific experiences (working in a role with specific competencies), transitions within schools (across years/grades), school visits to local museums/science centres (helping to alleviate anxiety and worries about visiting new places), and curriculum-specific content (tours of geographic and historical spaces, art projects, science enquiry, etc …).

3.6 Limitations

Our case study provides a snap-shot of a project we engaged with in an autism-specific school over a period of three months. We have outlined key lessons learnt, including the limitations of our own research, in attempting to bridge the research-practice gap when using HMD-VR with autistic children. It should not be assumed that closing this gap with the methods we propose is either the only way, nor might it work in other schools. We fully understand that this level of engagement and co-production may not always be possible in other settings, e.g. mainstream schools, where multiple priorities exist and time, curriculum and physical space pressures may be greater than those we experienced at Appletree school. Nonetheless, we argue that this case study can serve as a framework for ethical and participatory research in this field which can then be adapted to the context of different educational settings. We also recognise that this work reports on the use of very specific HMD-based VR. Further work is needed to investigate the ever-evolving context of HMDs and devices that are becoming cable-less, smaller and lighter. Finally, we also point out that the characteristics of the pupils are not fully elaborated and that because this groups accepted and used HMD-based VR successfully, it does not follow that all autistic pupils/people would. This is especially true when we consider the full array of autistic people with, sometimes, comorbidity.

4. Conclusion

As research in the field of autism and HMD-based VR moves forward, the development and use with autistic groups needs careful consideration, and urgent attention, as it relates to engaging with and prioritising the views of this community. As Spiel and colleagues (Spiel et al., 2019) suggest “researchers and developers of technologies for autistic children need to carefully reflect on how their work fits into the larger field and how it might contribute to the further marginalisation of autistic children”. Our case study sought to address this issue through establishing an ethical research process that went beyond ideas put forward by Cascio and colleagues (Cascio et al., 2020) who recommend that “researchers can facilitate inclusion by using inclusive terminology, developing accessible communication strategies, or traveling to meet participants”. Our close collaboration with Appletree School and the autistic mentor, Peter, meant that we were able to co-produce ethical processes that embedded the views, experience and knowledge of staff and pupils. However, we recognise research involving immersive technologies, including our own, needs to move beyond the current state of the art (i.e. Cascio et al., 2020). This requires building relationships and trust between researchers, schools and autistic pupils; seeing participants as collaborators, asking and listening to their views at all stages of the process and properly valuing their lived experience when planning, implementing, evaluating and disseminating research. When referring to lived experiences, we mean that researchers need to be sure they are aligning their activities to the groups they are working with, in genuine ethical ways. In the case of working in schools for autistic groups, we suggest this needs to be nuanced and situated around the specific ways autistic individuals experience the world around them. For some this could relate to sensory concerns/needs. For others, factors related to becoming overly immersed and feeling a heightened sense of presence Segovia and Bailenson (2009). Researchers need to be aware that there is no “one size fits all” in this research area. To support this endeavour, genuine research partnerships in this field, as within other areas of research involving autistic children, should mean “joint working between researchers (autistic or non-autistic) and autistic people, family members or practitioners, where research is carried out with or by community members rather than about or for them” (Pellicano et al., 2017). We argue that placing these perspectives central will yield insights to help move this field forward and help to fully and completely establish technology that can be implemented, designed and deployed more successfully than before in schools.

However, recognising that working in educational settings can be “both challenging and confronting as the competing demands of research, teaching and learning raise significant ethical issues” (Guldberg, 2017), we also conclude with a reflection that acknowledges the time and effort required in our case study to engage with stakeholders in meaningful ways, with a view to include them as completely as possible. This will have implications for funding, research site location selection, and the need to consider relationships with a range of stakeholders throughout the entire research journey (from co-design to dissemination and impact).

Figures

Plate 1

Room allocated for the study; highlighting the space, set up and pupils

Plate 2

Examples of the three types of HMDs used in the case study (L) and a pupil using the HTC Vive HMD in Appletree School (R)

Figure 1

Process map outlining key steps from reaching out to schools through ethical approval and dissemination

Figure 2

Example questions from our questionnaire, highlighting the visual nature to support participation in providing representative responses (L) and information leaflet with a highly visual design and links to media (R)

Figure 3

Details taken from our ethical/safety protocols in the case study school

Figure 4

Two-part process when examining the views of autistic pupils using various HMDs

Figure 5

Study Newsletter published via the Trust's website

Table 1

Appletree School demographics

School status	Special educational needs
Pupil admission	All pupils require an autism diagnosis
Age range	4 to 19
Number of pupils in whole school	143 (as of 2020)
Number of pupils involved our study and their age range	Involvement n = 12
	Aged 8–16 (12.4 mean)