Study design and setting

The Alfi VR trial is an unblinded, randomised controlled trial with two parallel groups and equal allocation ratio (1:1). The effects of the VR intervention compared with a usual care control will be assessed preintervention (week 1) and immediately postintervention (week 9). The trial will be conducted and reported in accordance with the Consolidated Standards of Reporting Trials statement.

The trial will be run across two urban, Australian community-based settings. Data collection and intervention sessions will be run at participating government and independent schools across metropolitan Melbourne and at Monash University. The trial commenced in February 2020, with an estimated completion date of August 2022.

Patient and public involvement

The public were able to provide feedback on the intervention during a proof-of-concept trial of Alfi VR. Proof-of-concept phases represent a key step in ensuring the appropriateness of the intervention prior to undertaking a pilot study. Typically developing teenagers from Melbourne, Australia were invited to participate in three Alfi training sessions, after which they provided feedback on the enjoyability, usability and accessibility of the training programme. This feedback was incorporated into the latest build of Alfi VR, which is being used for this trial, and was informed the timing and number of sessions in the current trial. Details of dissemination of results to participants and the public can be found under the Ethics and dissemination section.

Eligibility criteria

The study is targeted at adolescents, which is defined by the WHO as children between 13 and 19 years of age.21 Children will be eligible to participate in the trial if they are (1) aged between 13 years and 17 years 11 months at the time of enrolment into the study and (2) have elevated ADHD symptoms. Elevated symptoms are defined as t-scores above 58 on either the inattention or hyperactive/impulsive scales, or a t-score of above 58 on either the ADHD inattentive or ADHD hyperactive/impulsive Diagnostic and Statistical Manual of Mental Disorders: Text Revision (DSMIV-TR) scales of the Conners-3 Parent Rating Scale. Children will be excluded if they have (1) a diagnosis of developmental delay or borderline intellectual delay (defined as a Full Scale IQ of less than 80 on the Kaufman Brief Intelligence Test - Second Edition(KBIT-2)); (2) a history of neurological impairment including epilepsy, acquired brain injury or any history of seizures; and (3) any sensory or motor impairment or comorbid diagnosis that may prevent them from completing the intervention or understanding study instructions (eg, visual impairment, paralysis, severe obsessive-compulsive disorder). Screening questions pertaining to all exclusion criteria will be included in the demographic questionnaire, and clarification from parents/caregivers will be sought where responses are ambiguous or the level of impairment is unclear. Details of current medication use will be obtained through the demographic questionnaire. Children who are on a stable dosage will not be excluded from enrolling in the study; however, changes to dose or commencement of medication will result in withdrawal from the trial (see the Discontinuation criteria section).

Intervention

The intervention is an inhibitory control training intervention called Alfi VR. The Alfi VR intervention uses immersive VR technology to deliver a game-based version of an anticipated response stop-signal task.22 This paradigm has been used in studies of children with ADHD previously, with children with ADHD showing deficits on this task when compared with typically developing (TD) controls.23 Anticipated response versions of the stop-signal task require a fixed go response initiation time, rather than a speeded response as in reaction time versions of the stop-signal task. This affords greater control over the timing of the go response, and by extension the timing of stop cues relative to the go response.24 Furthermore this paradigm has been well characterised as targeting behavioural response inhibition.24 The intervention is graded in difficulty and includes task conditions that train reactive and proactive inhibition,25–30 as well as selective inhibition.31–35 As such, the Alfi VR intervention is immersive and challenging and targets core cognitive processes involved in inhibitory control throughout the course of training.

Adolescents are equipped with an HTC Vive VR headset and two controllers, with base stations mounted on stands. Within the programme users play the role of a wizard who must protect a magic crystal from dragon attacks by casting spells. Users are required to hold down the triggers on both controllers at the start of each trial. In Go trials, users release the triggers to cast spells at certain times to defuse the dragon’s fireball attacks (figure 1A). At the start of the intervention, the dragon’s fireball attacks occur on both sides, requiring users to release both triggers on the controllers. As the intervention progresses, the Stop trials are introduced requiring the user to inhibit their primed response to simultaneously release both fireballs on a proportion of trials (25%). In the first block that includes the Stop trials, wizards are introduced who are also able to cast spells and defuse the fireballs. When a wizard has defused a fireball, users must inhibit any initiated responses (eg, trigger releases), otherwise the fireball will become active once more (figure 1B). The timing of the wizard spell represents the stop signal delay (SSD). The SSD represents the amount of time between the onset of the Go signal and the Stop signal, with a longer SSD often associated with increased errors.36 The tracking method is used to adjust the SSD, increasing and decreasing by 33 ms (approximately two screen refreshes) depending on the correct or incorrect responding on the previous Stop trial.24

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Trial structure and timing. 1a. Go trials; 1b. Bimanual stop trials; 1c. Selective stop trials; 1d. Stop trials with cues (bimanual shown). L = Left; R = Right.

As the levels increase in difficulty, the fireballs may also occur from either the right or left side only, requiring one trigger to be held down while the other is released (selective inhibition; figure 1C), and cues are introduced which indicate the increased likelihood of the next trial being a stop trial (proactive inhibition; figure 1D). The use of a bimanual and selective paradigm requires the participant to attend to multiple streams of information simultaneously, increasing the ecological validity of the training. Furthermore, the introduction of cues in later trials (indicating the increased likelihood of a stop trial) allows for proactive inhibition to be trained in addition to reactive inhibition.

By providing variability in trial types and having an adaptive SSD, the potential for maintaining participant engagement is maximised. These variables ensure that the task remains challenging without becoming too difficult. Participants receive feedback on all responses, which includes corrective feedback when participants respond incorrectly or fail to respond. To provide motivation throughout training sessions, participants can monitor their progress and ingame statistics and receive virtual ingame rewards that can be used to personalise certain aspects of gameplay.

Researchers will observe and monitor participant progress throughout the session via the computer monitor. Training sessions take 20 min to complete, with each block within the session lasting approximately 6 min. The individual blocks comprise a 3:1 ratio of Go trials (75%) to Stop trials (25%). Participants will complete the Alfi programme twice per week for 8 weeks, totalling 16 training sessions. The dates and times of all training sessions are recorded to monitor treatment adherence, with treatment compliance defined as having completed 12 of the 16 allocated training sessions (75%).

Discontinuation criteria

Participants may be withdrawn from the trial if they (1) commence any medication or therapy for the purpose of addressing ADHD symptoms or if there are any changes to their current treatment (eg, increase in medication dosage); (2) experience recurrent or persistent symptoms of cybersickness; or (3) experience a serious adverse event.

Outcomes

Sample size calculation

A priori power analysis was conducted in G*Power V.3.1 to determine the necessary sample size required to detect significant changes in inhibitory control from pretraining (baseline) to post-training between groups. An estimated effect size of 0.3 was used based on the results of a small proof-of-concept study of the Alfi intervention in typically developing adolescents conducted by the authors. A total sample size of 97 was required for a power of 90%. The study will therefore aim to recruit 100 participants, with 50 per group assuming a 1:1 allocation ratio. This is consistent with previous trials evaluating cognitive training interventions in individuals with ADHD.43

Recruitment

Participants will be recruited through a number of avenues. The primary recruitment pathway will be through local government and independent schools. Researchers will contact schools within 40 km of the Monash University Clayton campus. A sampling matrix will be used to ensure that participants are recruited from a range of geographical regions across metropolitan Melbourne and to ensure diversity within the sample with regard to socioeconomic status. Interested schools will be provided with study information packs to disseminate to students and their families.

Participants will additionally be recruited via ADHD-related websites, community groups and ADHD-related social media groups. Due to the requirement for intervention sessions to take place at schools or at Monash University, it is anticipated that participants will predominantly reside in metropolitan Melbourne.

Interested families will contact the research team directly via phone or email and will be given the opportunity to discuss the study further. If families complete the informed consent form, they will be invited to participate in screening. Screening will take place online, with parents emailed a link to the demographic questionnaire and the Conners-3 Parent Rating Scale, which will be used to confirm clinical diagnosis of ADHD. Should children meet all the inclusion criteria and none of the exclusion criteria, they will be accepted into the trial.

Participant timeline/procedure

The timing of screening, randomisation, assessments and intervention can be seen in figure 2. The pretraining intervention will take place a week prior to the commencement of the intervention. The intervention period will run for 8 weeks, with the post-training assessment taking place a week after completion of the intervention. Refer to figure 2 for the intervention and assessment schedule.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Intervention and assessment schedule.

Allocation

Participants will be randomised following completion of the pretesting assessment. Block randomisation will be used, with blocks of four. As there are significant sex differences within ADHD,44 45 randomisation will be stratified by sex to ensure an even distribution of boys and girls between the intervention and control groups. Randomisation will be completed by a member of the research team using R. Group allocation will be communicated to participants and their families and recorded in the study database.

Concealment

Group allocation will be completed by an independent researcher who is not a member of the trial study team to reduce bias. A randomisation sequence will be generated in RStudio and printed onto folded cards with the group allocation listed on the inside. Following enrolment into the study, the participants’ study identification will be provided to the independent researcher, who will record it on the outside of the card and provide it to the trial researcher in a sealed opaque envelope. Following completion of the pretraining assessment, the envelope will be opened and the participant assigned to their allocated group.

Statistical methods

Data obtained during screening will be analysed to determine study eligibility; no further interim analysis will be conducted. At the completion of the trial, study data will be analysed using an intention-to-treat approach, meaning that outcome data for all participants will be included regardless of treatment compliance. Primary outcome data will be analysed in SPSS version 27 using analysis of covariance (ANCOVA) to examine changes in response inhibition from pretraining to post-training between groups. ANCOVA will allow for inclusion of covariates to minimise the impact of extraneous variables such as age, ASD traits (as measured by the Social Responsiveness Scale, 2nd Edition (SRS-2)) or cognitive function (as measured by the KBIT-2), where indicated. Multivariate analysis of covariance will be used to analyse differences in secondary outcome measures, with post-hoc tests completed for any significant results.

After completion of the primary intention-to-treat analysis, a sensitivity analysis using a per-protocol approach will be conducted to assess the robustness of the results and to determine whether there are any differences in outcomes between participants who met the criteria for programme compliance and those who did not.46 The management of missing data will be dependent on the amount of data missing and the pattern of missing data.47

Data collection/data management

Data will be collected by researchers trained in the administration and scoring of the study assessment measures. In week 1 families will be emailed links to the pretraining surveys for completion by the participant and a parent. These surveys take approximately 25 min to complete and responses can be saved and the survey resumed at a later time if required. Participants will concurrently complete the pretraining assessment session at week 1. This session will take place face-to-face and will be facilitated by two researchers. Following the 8-week intervention period, families will be emailed links to the post-training surveys and participants will concurrently complete the face-to-face post-training assessment session. These surveys must be completed within 2 weeks of completing the intervention. Study progress will be recorded for each participant in the study database.

Participant data will comprise a combination of electronic and paper-based records. Electronic data will be recorded and stored within a secure Research Electronic Data Capture (REDCap) database. Paper-based data will be identified by participant number only and will be stored securely and retained in accordance with relevant government regulations. As the study is unblinded a data monitoring committee was not deemed necessary. This role will be undertaken by the study team.