The uses of cognitive training technologies in the treatment of autism spectrum disorders.
Computer brain games teach kids to win the game, not to live the skill.
01Research in Context
What this study did
Cicchetti et al. (2014) looked at every paper they could find on computer brain games for kids with autism.
They did not run new kids; they simply told the story the studies told.
The team asked: do the games boost thinking skills only on the screen, or also at home and school?
What they found
The games almost always improved scores inside the same computer task.
Real-life changes—like talking more or finishing homework—were small or missing.
In short: the kids got better at the game, not at life.
How this fits with other research
Lee et al. (2025) found a similar gap. Paper pictures helped reading more than screen pictures, even when both looked alike.
Gilroy et al. (2023) and Fleury et al. (2018) ran real trials comparing high-tech tablets with low-tech cards for talking. Both tools worked the same, as long as adults used ABA steps.
Put together, the pattern is clear: the teaching method matters more than the shiny device.
Why it matters
Before you buy that new app, check if it has real-life practice built in. Pair any screen task with hands-on trials in the kitchen, classroom, or playground. Measure what the learner does there, not just what they do on the tablet. If you see no change after two weeks, shift time back to live coaching and skip the game.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Run one quick probe of the target skill in the natural setting before the first app session, then again after one week—drop the app if real-life scores stay flat.
02At a glance
03Original abstract
In this review, we focus on research that has used technology to provide cognitive training - i.e. to improve performance on some measurable aspect of behaviour - in individuals with autism spectrum disorders. We review technology-enhanced interventions that target three different cognitive domains: (a) emotion and face recognition, (b) language and literacy, and (c) social skills. The interventions reviewed allow for interaction through different modes, including point-and-click and eye-gaze contingent software, and are delivered through diverse implementations, including virtual reality and robotics. In each case, we examine the evidence of the degree of post-training improvement observed following the intervention, including evidence of transfer to altered behaviour in ecologically valid contexts. We conclude that a number of technological interventions have found that observed improvements within the computerised training paradigm fail to generalise to altered behaviour in more naturalistic settings, which may result from problems that people with autism spectrum disorders experience in generalising and extrapolating knowledge. However, we also point to several promising findings in this area. We discuss possible directions for future work.
Autism : the international journal of research and practice, 2014 · doi:10.1177/1362361313499827