A review of study titled Evaluation of an Immersive Virtual Reality Safety Training Used to Teach Pedestrian Skills to Children With Autismr

Dixon, D. R., Miyake, C. J., Nohelty, K., Novack, M. N., & Granpeesheh, D. (2020). Evaluation of an immersive virtual reality safety training used to teach pedestrian skills to children with autism spectrum disorder. Behavior Analysis in Practice, 13, 631-640. https://doi.org/10.1007/s40617-019-00401-1

Reviewed by:
Frank Cicero, PhD, BCBA
Seton Hall University and the Association for Science in Autism Treatment

Why research this topic?

Procedures to teach specific skills for individuals with autism Characteristics of autism spectrum disorder significantly increase risks of injury from everyday activities in the home and community. Common activities such as swimming, commuting to school, household chores (e.g., cooking), and crossing streets pose dangers to individuals with autism despite education and exposure. Dixon and colleagues referenced some recent research on safety training for individuals with autism that largely focused on the effects of direct behavioral teaching methods such as video modeling, live modeling, role play, corrective feedback, prompting, and reinforcement. When it comes to safety skills programming, a consideration that must be taken into account is whether to conduct teaching trials in contrived settings, such as a classroom, or natural settings in which the safety skill is actually required (i.e., in an actual intersection when teaching safe street crossing). The decision is a tradeoff between ensuring safety throughout the learning process (maximized through contrived settings) and generalizing learned skills to real-life settings (maximized through natural settings). Whereas some skills (e.g., street crossing) are dangerous to teach in natural settings, other skills such as how to respond to a house fire, are impossible or unethical to teach within actual situations. Research, however, has demonstrated that safety skills taught within contrived settings require additional training within real-life situations in order to promote generalization (i.e., Bergstrom et al., 2014). This lengthens and complicates the training program.

The answer to this issue may lie in the recent popularity of virtual reality educational programs, which are being used to teach children with autism social, communication, daily living, and cognitive skills (e.g., McMahon et al., 2020; Genova, et al., 2021; Cox et al., 2017). Because virtual reality programs can safely portray elements of the natural environment, they are also being used to teach safety skills such as street crossing. Although studies have shown increases in street crossing skills within the virtual reality environment (Josman et al., 2008; Strickland et al., 2007), limitations were noted when it came to generalization to natural settings. In a systematic review of virtual reality for teaching social skills, Miller and Bugnariu (2016) found that the level of immersion of the virtual reality system correlated with successful skill acquisition. High immersion virtual reality (noted by limited signals indicating devices, involvement of more than two senses, use of a head-mounted device, high visual resolution, and full body motion capture) led to the best outcome.

What did the researchers do?

Dixon et al. (2020) investigated the effects of high immersion virtual reality training on street crossing skills in children with autism. The researchers hypothesized that a high-immersion environment would create an effective and safe teaching setting while maximizing generalization to real-life settings. Participants included three children with autism ages four, six, and ten. Diagnoses were confirmed with the Pervasive Developmental Disorder Behavior Inventory (PDDBI). All participants used spoken language to communicate, received behavior analytic center-based services at least three times per week, had pre-requisite skills of direction following and paying attention, lacked any visual or auditory impairments, tolerated the head-mounted virtual reality device after a brief desensitization period, and demonstrated a lack of street crossing skills in baseline. Baseline sessions and natural environment probes were conducted in unmodified real-life streets within the community.

Virtual reality training sessions were conducted in a training room in a clinic. High-immersion virtual reality programming was employed using an Oculus Rift (Version CV1; 2016) headset and sensors. The headset was connected to a laptop so that the researchers could see what the participant was simultaneously viewing. The immersive virtual reality environment consisted of multiple examples of 360-degree videos of active streets within the local community. All videos were recorded by the experimenters. Data was collected on three targets, the first of which was looking left and right within five seconds of being presented with a street crossing video. The second target was a correct response to the question “Is there a car moving?” The third target was a correct response to the question “Is it safe to cross?” Because the experimenters were able to view the videos simultaneously on a laptop, they were able to judge correct and incorrect responses. The mastery criterion was defined as 100% correct across two consecutive virtual reality training sessions and three natural environment probes.

The study was conducted within a nonconcurrent multiple baseline design across the three participants. In nonconcurrent multiple baseline designs, which are often used when conducting research in applied settings, an intervention is initiated to several participants in the same way but at different moments in time. If similar effects are shown across the participants even though the interventions were not initiated simultaneously, it would lend evidence to the effectiveness of the intervention. After conducting baseline probes and ensuring that the participants tolerated the wearing of the headsets through implementing desensitization sessions, high immersion virtual reality training sessions were initiated. Desensitization sessions consisted of having the participants become familiar with wearing the headsets while playing preferred virtual reality games. Sessions lasted three-five minutes and were required only two to three times. Sessions were conducted over two to three days within a one-week period. Individual sessions lasted between three-five minutes in duration and consisted of between three and eight teaching trials. Up to four sessions were conducted per day. For each trial, a ten second video was shown to the participant through the high-immersive virtual reality headset. Later in the training, videos with audio distractors and longer length videos (four to five minutes in duration) were introduced. Teaching consisted of viewing of the virtual reality videos, flexible prompt fading (prompts chosen by the experimenter as needed), verbal praise, and tangible reinforcement.

What did the researchers find?

During baseline, all three participants demonstrated low and stable responding at 50% or less correct. Quick skill acquisition was then achieved within the virtual reality training sessions, with participants meeting mastery criteria in the virtual reality environment within three to five sessions. Unfortunately, none of the participants demonstrated generalization to the natural environment after initial mastery of virtual reality. After longer duration videos with audio distractors were introduced into the high immersion virtual reality program (replacing the initial ten-second videos) all three participants were then able to meet the mastery criterion in the natural environment.

What are the strengths and limitations of the study?

The current study improved on the current state of research by demonstrating that successful and generalized street crossing skills can be taught to children with autism through the use of highly immersive virtual reality technology (along with prompting and reinforcement) in the absence of teaching trials in the natural environment. Training length was short in duration and participants did not show an aversion to the technology. Interobserver agreement data were high across targets and participants.

The authors self-identified a few limitations of their study that must be considered when interpreting the data and study implications. The original videos of 10 seconds in length were ineffective in promoting generalized skills despite producing mastery-level responses within virtual reality training sessions. The authors speculate that the onset of short videos became discriminative stimuli for correct responses, whereas longer-length videos better reflected real-life conditions. In the future, they would recommend only using the longer length videos with audio distractions in the virtual reality sessions. Another potential limitation of the current treatment is the possibility that children may not easily tolerate the virtual reality headset. A more significant limitation regards the targets that were used in the study, which consisted of verbal responses of whether or not it was safe to cross the street rather than actual street-crossing behavior. Unfortunately, significant software development would have been needed to use physical street crossing targets. A final caution involves the use of virtual reality headset devices with children, the safety of which had not yet been properly researched in this population at the time of the study (Gent, 2016). An additional potential limitation, which was not identified by the researchers, involves the reliance on a nonconcurrent multiple baseline design. Although a valid single subject research design, it is not as strong a design as a traditional multiple baseline design when it comes to accounting for interfering variables that could have affected the interpretation of the data.

What do the results mean?

Data from the current study indicate that important safety skills, such as street crossing behaviors, can be effectively taught to children with autism through the use of virtual reality technology along with prompting and reinforcement. The most important finding from this study was that once four to five minute videos with audio distractors were used within the high immersive virtual reality programming, generalization to natural settings (real-life active streets in the community) was quickly achieved without any need for training in the real-life setting. This ensures safety during training, a concern that is always present when teaching children with autism in natural environments where the potential for danger is a reality until a skill is learned to mastery. When interpreting the results it is important to keep in mind that only verbal responses of whether or not cars were approaching and whether or not it was safe to cross the street were taught and evaluated. Actual street crossing behaviors were not targeted. At this time, however, the safety of using high immersion virtual reality systems with children continues to need investigation. Also, some children may require extensive desensitization trials in order to tolerate the virtual reality headset. Finally, although potentially very effective, virtual reality training may involve expensive devices and software, requires understanding on the part of the trainer on how to properly use the technology, necessitates the recording and editing of carefully designed videos, and must be used along with behavioral teaching techniques such as prompting and reinforcement.

References

Bergstrom, R., Najdowski, A. C., & Tarbox, J. (2014). A systematic replication of teaching children with autism to respond appropriately to lures from strangers. Journal of Applied Behavior Analysis, 47(4), 861-865. https://doi.org/10.1002/jaba.175

Cox, D. J., Brown, T., Ross, V., Moncrief, M., Schmitt, R., Gaffney, G., & Reeve, R. (2017). Can youth with autism spectrum disorder use virtual reality driving simulation training to evaluate and improve driving performance? An exploratory study. Journal of Autism and Developmental Disorders, 47(8), 2544–2555.

Dixon, D. R., Miyake, C. J., Nohelty, K., Novack, M. N., & Granpeesheh, D. (2020). Evaluation of an immersive virtual reality safety training used to teach pedestrian skills to children with autism spectrum disorder. Behavior Analysis in Practice, 13(3), 631-640. https://doi.org/10.1007/s40617-019-00401-1

Genova, H. M., Lancaster, K., Morecraft, J., Haas, M., Edwards, A., DiBenedetto, M., Krch, D., DeLuca, J., & Smith, M. J. (2021). A pilot RCT of virtual reality job interview training in transition-age youth on the autism spectrum. Research in Autism Spectrum Disorders, 89, 101878. https://doi.org/10.1016/j.rasd.2021.101878

Gent, E. (2016, October). Are virtual reality headsets safe for children? A lack of data and guidelines is leaving consumers in the dark about virtual reality’s potential negative side effects for kids. Scientific American. Retrieved from https://www.scientificamerican.com/article/are-virtual-reality-headsets-safe-for-children/

Josman, N., Ben-Chaim, H. M., Friedrich, S., & Weiss, P. L. (2008). Effectiveness of virtual reality for teaching street-crossing skills to children and adolescents with autism. International Journal on Disability and Human Development, 7(1), 49-56. https://doi.org/10.1515/IJDHD.2008.7.1.49

McMahon, D. D., Barrio, B., McMahon, A. K., Tutt, K., & Firestone, J. (2020). Virtual reality exercise games for high school students with intellectual and developmental disabilities. Journal of Special Education Technology, 35(2), 87-96. https://doi.org/10.1177/0162643419836416

Miller, H. L., & Bugnariu, N. L. (2016). Level of immersion in virtual environments impacts the ability to assess and teach social skills in autism spectrum disorder. Cyberpsychology, Behavior, and Social Networking, 19(4), 246-256. https://doi.org/10.1089/cyber.2014.0682

Strickland, D. C., McAllister, D., Coles, C. D., & Osborne, S. (2007). An evolution of virtual reality training designs for children with autism and fetal alcohol spectrum disorders. Topics in Language Disorders, 27(3), 226-241. https://doi.org/10.1097/01.TLD.0000285357.95426.72

Citation for this article:

Cicero, F. (2024). Research synopsis: Evaluation of an immersive virtual reality safety training used to teach pedestrian skills to children with autism spectrum disorder. Science in Autism Treatment, 21(05).

Related ASAT Research Synopses:

Other Related ASAT Articles:

Related ASAT Reviews:

Related Media Watch Letters:

#Community #Home #Safety

Evaluation of an Immersive Virtual Reality Safety Training Used to Teach Pedestrian Skills to Children with Autism Spectrum Disorder