Table 2.

The characteristics of reviewed articles (n = 17)

Authors	Sample size	Sample description (Sex- age)	Experimental intervention	Virtual reality implementation	Study design	Session details (number of sessions, duration)	Cognitive outcomes (memory, attention, executive function (daily skills), learning)	Assessment time	Assessment Score	Results	Reported limitations
De Luca et al. 2019, Italy [23]	1	16-year-old boy	VR training using an innovative tool, namely BTS-Nirvana (BTS-N)	BTS-N is a medical device based on VR, this is the first device using a two-dimensional flat-screen projection system with optoelectronic infrared sensors, through which the patient can simply interact by his movements	Pre and post test	24 session, each session 40 min	Attention and visual-spatial exploration	Baseline and post intervention	Authors found a significant increase of attention processes, by the improvement of MTCM [from T0 with a rapidity score of 5 c¼49, 2–3 SD and accuracy score of 34 c¼124, 7 > 3 ds to T3 with a rapidity score of 19 c¼49 and accuracy score of 68 c¼124,7 > 3 ds]	This case-study showed that VR could be helpful to potentiate cognitive and adaptive behavioral (with regard to attention process, spatial cognition, and visual-motor integration	Epidemiological bias Impossibility of causal inference Generalization, and over-interpretation Small sample sizes Cost of virtual reality devices
Wang et al. 2013, Canada [24]	4	AgeM = 6.7 yrs., Male = 3, Female = 1	Virtual reality training programs	This program is based on a two-dimensional flat screen projection system. This system has motion-capture capabilities, where a tracking camera is able to capture and project a child’s image and motions on screen in real-time	Pre and post test	4–6 sessions	Contextual processing of objects	Baseline, middle of intervention and a two-week follow up session	The results demonstrate improvements in contextual processing ability from baseline to treatment for each child, with average increases from 15% (Child 2) to 46% (Child 4). All children maintained a high level of performance at the two-week follow-up assessment	All children demonstrated statistically significant improvements in contextual processing and cognitive flexibility. Mixed results were found on the control test and changes in context-related behaviors	Small sample sizes The lack of multiple, independent assessors Time of follow up duration is limit Potential eye damage
Lamash et al. 2017 Israel [25]	56	AgeM = 14.58 yrs., Male = 29, Female = 4	VAP-S	V AP-S—is a virtual environment software that operates on a laptop or desktop and requires the use of keyboard arrows and a mouse	Case–control study	8-sessions	Executive function: teach shopping task form a supermarket	Baseline and post intervention	The results show that among the intervention group, significant improvements were found in the attention component (P < 0.01) and in the executive functions component (P < 0.01); the intervention group showed a significant improvement in all the accuracy indices in shopping task (F (1, S4) = 14.23 P < 0.01)	The results show improvement of the intervention group compared to the control group in several indices, indicating great promise for intervention programs based on virtual technology for improving the independency and community participation with ASD	No limitation was reported
Adjorlu et al. 2019, Denmark [26]	5	18 to 22 yrs., Male = 5	VR for teaching money skills	The VR money training intervention was developed using Unity 3D and C# scripting. The application was developed to run on the HTC Vive VR hardware The virtual coins and bills were designed using textures from images of real danish money; The coins and bills could be grabbed by the player using the grab button on the HTC Vive controller and released agains by releasing the same button	Pre and post test	5 sessions for 10–15 min each one	Executive function: teach money skills	Baseline and post intervention	Results showed that the maximum improvement was score from 9 pre-test to 30 post-test and minimum score was 0 to 0. Students C and D illustrated very small improvements (from 3 to 8 and 0 to 2 correct purchases)	Four out of the five participants showed some improvement in their money skills after five training sessions with the VR application	No limitation was reported
Adjorlu et al. 2017, Denmark [27]	9	12 to 15 yrs., Male = 8, Female = 1	VR based supermarket shopping training system	The VR intervention was developed using Autodesk Maya and Unity. HTC Vive was chosen to run the application due to its effective room scale tracking. The signals are than captured via the infrared sensors placed on the VIVE head-mounted display	Case–control study	7-sessions	Executive function: teach shopping task form a supermarket	Baseline and post intervention	The treatment group efficiency score was 100% during both the baseline- and post-treatment assessments. On the contrary, the control group effectiveness decreased from 97% in the baseline assessment to 91% in the final assessment	The study indicates some positive effects of a head-mounted display-based VR simulation to train DLS of individuals	The weight of devices
Bian D et al. 2019, USA [28]	23	AgeM = 15.19 yrs., Male = 21, Female = 2	VR driving simulator	Models in the virtual driving environment, such as traffic lights, stop signs, and vehicles, were developed with the modeling tools ESRI CityEngine and Autodesk Maya. The game development platform Unity3D was used to implement the system logic	RCT	1-session, 90 min	Executive Function: teach driving task	Baseline and post intervention	No differences were found in performance data, however, with participants in ES group achieving similar performance as participants in PS group	These findings could support future work into driving simulator technologies, which could provide opportunities to practice driving skills in cost-effective, supportive, and safe environments	Long-term driving training program is needed
Cox et al. 2017, USA [29]	51	AgeM = 17.96 yrs	VRDST	The commercially available DGS-78 VRDS is a realistic driver’s cockpit with side and rear-view mirrors. The driver’s view is projected onto a 2.44 m (8 ft) diameter, 210° curved screen	RCT	8–12 sessions, 60 min each session	Executive Function: teach driving task	Baseline and post-intervention and after 3 months of training (follow up duration)	The general tactical composite score improved differentially across groups (p < 0.010), and a significant covariate (p < 0.001, β = 0.50) indicated that better baseline performance was associated with better post-assessment performance	VRDST significantly improved driving and EF performance over RT. This study demonstrated feasibility and potential efficacy of VRDST for novice ASD drivers	Obtrusive or irritating nature of wearing eye-tracking glasses Time of follow up duration is limit
Peisley et al. 2019,New Zealand [30]	4	6 to 7 yrs	Virtual Week	Participants viewed the Virtual Week board game on a laptop PC monitor and responded using the keyboard. Participants clicked the mouse to roll the die, read aloud the event cards, and made decisions about the daily activities. All children were reading at, or slightly below, their age level and could read the cards without assistance	Pre and post test	45–60 min each session	Memory	Baseline and post intervention	There was a significant main effect of regularity on mean accuracy scores and prospective memory, p = .04, r = .79, but no main effect of PM-task type, p = .59, r = .27, and no significant interaction between regularity and task type, p = .08, r = .65	These results suggest the delivery of positive reinforcement may improve accuracy on PM tasks post reinforcement	Long-term training program is needed
Austin DW et al. 2008, Australia [31]	2	AgeM = 14.50 yrs., Male = 29	VRH	The VRH method uses a head-mounted display to create a non-threatening, virtual reality environment, where the hypnotherapeutic process can be implemented	Pre and post test	4-sessions	Attention	Not mentioned	No Scores are reported	They indicated that they believed it was an effective technique to gain their son’s attention, and this, combined with the fact that the boys found it enjoyable and engaging, led them to believe there is significantly potential for this particular treatment modality	No limitation was reported
Herrera et al. 2008, Spain [32]	2	AgeM = 15.8 yrs	VE	Not mentioned	Pre and post test	Not mentioned	Executive Functions	Baseline and post intervention	The first participant showed considerable progress in structured pretend play, obtaining an improvement of 6.5. The advances shown by the second participant were also in both types of play. In the structured play test, he gained 4.75 points (from 40.3 to 49.8 months)	The results, confirmed by independent observers, showed a significant advance in pretend play abilities after the intervention period	No limitation was reported
Josman et al. 2008, Israel [33]	12	8 to 12 and 14 to 16 yrs., Male = 10, Female = 2	VE	Three keyboard keys (marked on a standard keyboard with round, colored stickers) were used to change the user's viewpoint to the right or to the left or to initiate street crossing. Users who succeeded in safely crossing the street automatically proceeded to the next stage	Case–control study	Not mentioned	Executive Function: teach street-crossing skills	Baseline and post intervention	The maximum stage reached by the research group during Phase A ranged from one to four, with a mean of 2.7 (SD) = 1.2). Using the Wilcoxon test, p < .01 was obtained, demonstrating a significant difference between the two groups	Significant differences were found between the performance of the experimental and control groups within the VE. half of the experimental subjects made considerable improvement	No limitation was reported
Weilun et al. 2011, Singapore [34]	9	AgeM = 16 yrs	Virtual game	Following an iterative and incremental model for software development life cycle, an interactive quiz programme which employs virtual avatar to pose academic-related questions to autistic students is developed. Additionally, drawing on a Sumo wrestling game using robotics agent iRobot Create to hone motor skill is evaluated	Pre and post test	Not mentioned	Learning meaningful academic activities	Not mentioned	No scores were reported	Experimental results showed that the deployment of virtual games hold great potential in motivating and exciting the children in their learning process, as well as providing valuable insights to related rehabilitative industries	No limitation was reported
Self et al. 2007, USA [35]	8	6 to 12 yrs., Male = 6, Female = 2	VE	Not mentioned	RCT	20-sessions, 30 min each session	Executive Function: teach how to escape the fire and survive	Baseline, middle of intervention and post intervention	No scores were reported	Both groups improved in their learning and transfer of safety skills. The VR group, however, learned these skills in considerably less time	No limitation was reported
Saiano et al. 2015, Italy [36]	6	19 to 44 yrs., Male = 7	VE	The experimental apparatus included a video projector, displaying a virtual reality environment on a 2 m × 2 m screen. The participants were required to stand in front of the screen, at a distance of approximately 2 m. Instead of using mouse and keyboard to interact with the VE, they used a markerless motion capture device (Microsoft Kinect), placed below the screen to record the subjects’ full-body movements in 3D space	Pre and post test	10-sessions, 45 min each session	Executive Function: teach street-crossing skills	Baseline and post intervention	The ability to follow the street signs, we found that subjects significantly increased (p = 0.0042; paired-samples t-test) their average speed from T0 to T1, of an amount ranging from 40 to 100%; they found no significant changes in path length, figural distance, and composition index	The six subjects who completed the protocol easily learned the simple body gestures required to interact with the VE. Both parents and caregivers reported a significant improvement in the subjects’ street crossing performance	Small sample sizes The team only made an indirect assessment of transfer of the learned skills to real life
Dixon et al. 2019, USA [37]	3	AgeM = 6.6 yrs., Male = 3	VE	Not mentioned	Pre and post test	3 and 5 min, an average of 5.46 trials; up to 2 to 3 days in a 1-week period	Executive Function: teach street-crossing skills	Baseline and post intervention	For each participant in the study, low, stable scores were observed during baseline and an increase in scores was seen after each VR training condition	Findings suggest that immersive VR is a promising medium for the delivery of safety skills training to individuals with ASD	Lack of testing in the natural environment Only the skill of identifying whether a street is safe to cross was trained and evaluated Supervised usage and limited access to such technology until larger studies demonstrate its safety
Wade et al. 2017, The USA [38]	Pilot 1: 7 Pilot 2: 9	Pilot 1: AgeM = 16 yrs., Male = 4, Female = 3 Pilot 2: AgeM = 15.26 yrs., Male = 9	VADIA	Users interact with the system via a Logitech G27 controller, which features a steering wheel, pedal board, and gear shifter, although this last item was not utilized in the presented studies; The G27 controller mounts conveniently onto a car-like bucket seat that is positioned in front of a flat panel LCD monitor displaying the driving environment	Case–control study	Pilot1: 90 min, 1 session Pilot 2: 60 min, 6 session	Attention	Baseline and post intervention	Study 1 demonstrates statistically significant performance differences between individuals with (N = 7) and without ASD (N = 7) with regards to the number of turning-related driving errors (p < 0.01) Study 2 shows that both the performance-based feedback group (N = 9) and combined performance- and gaze-sensitive feedback group (N = 8) achieved statistically significant reductions in driving errors following training (p < 0.05)	In this study, the use of virtual reality-based systems has significant effects on people's cognition index but researcher team will assess changes in performance based on best practice clinical and on-road evaluation metrics	Small sample sizes Limit the statistical power of the analyses and subsequently the generalizability of the results Age and driving experience are not sufficiently controlled (confounders are not completely controlled) Potential addiction
Bozgeyikli et al. 2016, USA [39]	15	AgeM = 25.4 yrs	VR4VR	The system is composed of the following hardware components: a Head Mounted Display (HMD); an optical motion tracking system with 12 cameras; a large 180° curved curtain screen; controllers; tangible objects equipped with optical markers that can be tracked real time by the system; and a tablet computer for remote control panel for the job coaches	Pre and post test	Not mentioned	Executive function: teach money skills, cleaning, vocational training, shelving	Not mentioned	No scores were reported	Proposed system utilizes six transferrable skill modules within immersive virtual environments for vocational training of individuals with ASD	No limitation was reported

MTCM: The Modified Little Bell Test; SD: Standard Deviation; VAP-S: The Virtual Action Planning Supermarket; VR: Virtual Reality; DLS: Daily Living Skills; RCT: Randomized Control Trial; PS: Performance-Sensitive System; ES: Engagement-Sensitive System; VRDST: VR Driving Simulation Training; PC: Personal Computer; DGS-78: Driver Guidance System; PM: Prospective Memory; VRH: Virtual Reality Hypnosis; VE: Virtual Environment; VADIA: The Virtual Reality Adaptive Driving Intervention Architecture; VR4VR: Virtual Reality system For Vocational Rehabilitation; VADIA: The Virtual Reality Adaptive Driving Intervention Architecture