A Randomized Trial Utilizing EEG Brain Computer Interface to Improve Facial Emotion Recognition in Autistic Adults.
Screen-based facial-emotion drills give autistic clients a short-lived bump in reading faces but the skill fades and stays narrow.
01Research in Context
What this study did
The team pooled every good-quality trial of computer facial-emotion training for autistic people. They found 595 participants across the studies.
All programs used screens to teach happy, sad, angry, and other faces. Some added voice clips or short games.
What they found
Right after the course, learners scored moderately better on face-reading tests. The boost equals about one extra right answer for every three tries.
Months later the gain was gone. Scores also did not rise on wider social tests like starting chats or taking turns.
How this fits with other research
Vasilevska Petrovska et al. (2019) saw large gains that spread to new faces after only 12 hours of training. The new meta shows smaller, non-spreading effects. The difference is Ivana used one tight lab program while the meta mixes many real-world trials.
Mittal et al. (2024) found large social gains with immersive VR. The flat-screen programs in the meta gave only medium face-only gains. VR’s real-life feel may explain the stronger result.
Richman et al. (2001) ran the first RCT in this area. Their medium effect still matches today’s pooled average, showing the field has not moved the needle in two decades.
Why it matters
Use computer emotion games as a quick primer, not a stand-alone fix. Plan brief refresher sessions every few weeks to keep the skill alive. Pair the software with real-peer practice so learners link faces to live conversations.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Schedule five-minute face-review warm-ups at the start of each social group and collect weekly cold-probe data to catch drop-off early.
02At a glance
03Original abstract
A large number of computer-based training programs have been developed as an intervention to help individuals with autism spectrum disorders (ASD) improve their facial emotion recognition ability, as well as social skills. However, it is unclear to what extent these facial emotion training programs can produce beneficial, long-lasting, and generalizable results. Using standard meta-analytic techniques, we investigated the effects of facial emotion training including generalization and maintenance restricted to randomized control trial studies comprising a total of 595 individuals with ASD. Our findings revealed that the intervention resulted in a robust improvement in emotion recognition for individuals receiving training compared with controls. However, while there was also some evidence for generalization of training effects, the small number of studies which conducted follow-ups and assessed social skills reported that improvements were not maintained and there was no evidence for general improvement in social skills. Overall, the analysis revealed a medium effect size in training improvement indicating that facial emotion training may be an effective method for enhancing emotion recognition skills in ASD although more studies are required to assess maintenance of effects and possible general improvements in social skills. LAY SUMMARY: Facial emotion training as an intervention may be a potential way to help improve emotion recognition in autism spectrum disorder (ASD), however robust empirical support for its efficacy has not been sufficiently established. Here, we conducted a meta-analysis of previous studies to summarize the effects of facial emotion training on ASD. Our results show that the training produces a robust improvement in subsequent emotion recognition, while maintenance and generalization effects still need further investigation. To date, no experimentally verified improvements in social skills have been reported.
Journal of autism and developmental disorders, 2025 · doi:10.1002/aur.2565