همایش بین المللی هوش مصنوعی

صفحه اصلی / دومین همایش بین المللی هوش مصنوعی

Personalized Serious Games for ADHD: A Systematic Review and Evidence Synthesis

نویسندگان :

Mohammad Cheraghi¹ Milad Hassani² Niloofar Mirzaei Chahardeh³

1- دانشگاه آزاد اسلامی واحد تهران مرکزی 2- دانشگاه تهران 3- دانشگاه آزاد اسلامی واحد علوم و تحقیقات

کلمات کلیدی :

ADHD،Serious Games،Personalization،Adaptive Algorithms،Machine Learning

چکیده :

Abstract— Attention-Deficit/Hyperactivity Disorder (ADHD) is a neurodevelopmental condition commonly affecting 5-7% of children worldwide, characterized by inattention, hyperactivity, and impulsivity. Serious games have arisen as entertaining digital treatments, and personalization has been put forward as a way of enhancing their therapeutic effect through adapting tasks to the individual's needs. This systematic review synthesizes evidence from 20 studies assessing personalized or adaptive serious games for children and adolescents with ADHD published between January 2011 and August 2024. Searching across five databases resulted in 72 unique records, while 20 met the inclusion criteria. Personalization mechanisms - usually enabled by AI or machine-learning models - included adaptive difficulty, individual feedback, and tailored reinforcement schedules. Compared to nonadaptive controls, personalized games yielded small-to-moderate improvements in attention and working memory, with heterogeneous yet generally positive effects on behavioral symptoms. Engagement levels were consistently high, and adverse events were rare. However, overall certainty of evidence was reduced due to methodological constraints such as small sample sizes, limited blinding, and diversity in outcome measures, leading to a moderate-to-high risk of bias across studies. Personalized serious games therefore represent a potentially promising adjunctive tool in ADHD interventions, especially in cognitive enhancement. Future studies should prioritize using standardized outcome measures, longer follow-up duration, transparent reporting of the adaptive algorithm, and rigorously designed trials that will enable replication and meta-analytic synthesis.