ACTing Minds: An ACT-based Videogame for Mental Health

This trial evaluates a brief, scalable approach to supporting common mental health difficulties by embedding Acceptance and Commitment Therapy (ACT) processes in an engaging interactive experience. ACT targets psychological flexibility, the capacity to make values-guided choices while openly contacting difficult thoughts and feelings. ACTing Minds integrates core ACT processes (openness and acceptance, cognitive defusion, present-moment awareness, values clarification and values-guided action) within a structured, single-session videogame.

The objective is to determine whether ACTing Minds yields greater improvement in psychological flexibility and psychological distress than a neutral, engagement-matched game, and whether those process-level changes are accompanied by improvements in wellbeing and related functioning. The trial specifies two co-primary outcomes: psychological flexibility, assessed with the CompACT total score, and general psychological distress, assessed with the DASS-21 total score. For each co-primary outcome, the prespecified primary estimand is the between-group difference in change from baseline to short-term (10-13 days) follow-up. Secondary estimands examine change in CompACT and DASS-21 subscales, subjective wellbeing, social connectedness, health-related quality of life, and autonomic regulation indexed by heart-rate variability derived from short seated ECG recordings.

The study uses a two-arm, parallel-group, randomised design with 1:1 allocation at a single university laboratory site. Randomisation is computer-generated with concealed allocation and implemented after baseline assessment to prevent foreknowledge of assignment. Participants are masked to condition, and blinding integrity is assessed at follow-up by asking participants to guess their allocation. Both arms follow identical procedures and time on task; the comparator is a neutral commercial game selected to match the ACT intervention for duration and engagement while avoiding therapeutic content.

Assessments occur at baseline, immediately after the single session, and 10-13 days post-intervention. Questionnaires are administered electronically with built-in range and completeness checks to minimise entry errors. ECG is recorded seated for about five minutes per timepoint (Lead II). Heart-rate variability processing follows standard steps: signal inspection, identification and correction of artefacts, R-peak verification, derivation of time- and frequency-domain indices, and natural-log transformation of skewed metrics prior to modelling. Mean heart rate (beats per minute) is retained to contextualise autonomic indices.

The statistical analysis plan is specified as follows. Primary and secondary outcomes will be analysed using linear mixed-effects models with fixed effects of Condition, Timepoint and their interaction, and a random intercept for participant to accommodate within-person correlation and incomplete follow-up under maximum likelihood. Residual covariance structures (for example, autoregressive versus unstructured) will be compared using information criteria, retaining the better-fitting specification for each outcome family. The primary contrasts estimate group differences in change from baseline to short-term follow-up on the two co-primary outcomes, CompACT Total and DASS-21 Total; post-session contrasts and 95% confidence intervals will also be reported. Heart-rate variability variables will be analysed on the natural-log scale. Two-sided tests with an alpha of .05 will be used, with attention to confidence intervals and effect sizes. Analyses follow an intention-to-treat principle using all available data; planned sensitivity checks will include alternative covariance structures and, if warranted, supplementary multiple-imputation analyses for outcomes with materially higher missingness.

Data quality procedures include scripted data pipelines with version control, reproducible derivation of analysis datasets from raw sources, and automated checks for ranges, outliers and internal consistency across timepoints. The research team maintains a study data dictionary that defines each variable, coding and scoring rules, and any transformations applied. Access to identifiable data is role-based and restricted to authorised personnel; de-identified analysis files are stored on secure university systems. Adverse events are monitored during the laboratory session and at follow-up, and any serious events are managed according to institutional policy. Given the single-session, minimal-risk behavioural design, an independent data monitoring committee is not required.

The target sample size is approximately 300 participants, providing precise estimates of group differences and allowing robust sensitivity analyses. On completion of the trial and after primary publication, de-identified individual participant data, a data dictionary and analysis code will be shared on a public repository, subject to ethical and legal safeguards. Findings will be disseminated through peer-reviewed publications, conference presentations and lay summaries suitable for non-specialist audiences.