Artificial Intelligence-Based Interventions for the Management of Attention-Deficit/Hyperactivity Disorder (ADHD): A Systematic Review of Current Evidence

Abstract

Background:
Attention-deficit/hyperactivity disorder (ADHD) is a prevalent neurodevelopmental disorder associated with impairments in attention, executive function, and behavior. Pharmacologic and behavioral therapies face challenges related to incomplete symptom control, adherence, and side effects. Artificial intelligence (AI)-based interventions have emerged as novel tools to personalize and optimize ADHD management.

Objective:
To systematically review the current evidence on AI-based interventions for ADHD symptom relief, assessing their efficacy, safety, and clinical applicability.

Methods:
A comprehensive literature search was conducted across PubMed, Scopus, Web of Science, PsycINFO, Cochrane Library, and IEEE Xplore for articles published between 2010 and 2024. Inclusion criteria encompassed interventional studies evaluating AI-based therapeutic interventions for ADHD. Data were extracted on study design, sample size, intervention type, outcomes, and safety. Risk of bias was assessed using ROB-2 for randomized controlled trials.

Results:
From 342 identified studies, eight interventional studies met inclusion criteria, including randomized controlled trials, clinical trials, and pilot studies. AI interventions included FDA-approved digital therapeutics, gamified cognitive training, AI-assisted neurofeedback, chatbot-based coaching, and robotic agents. Across studies, AI-based interventions demonstrated improvements in attention regulation, executive function, and ADHD symptoms. Safety profiles were favorable with no serious adverse events reported.

Conclusions:
AI-driven interventions represent a promising adjunctive approach for ADHD symptom management. While early evidence supports their clinical utility, further large-scale randomized trials are necessary to validate long-term effectiveness, optimize personalization, and assess real-world scalability.