Dealing with Autism Spectrum Disorders: Journey from Traditional Methods to Artificial Intelligence

Area of Science:

• Neuroscience and Developmental Psychology
• Medical Informatics and Artificial Intelligence

Background:

• Autism spectrum disorder (ASD) affects social interaction and communication, with global prevalence around 1% in children.
• Historically termed 'autism,' it's now an umbrella term for neurodevelopmental disorders (NDDs) with diverse presentations.
• Early identification and intervention are crucial for individuals with ASD.

Purpose of the Study:

• To evaluate traditional methods and explore the effectiveness of artificial intelligence (AI) in diagnosing and treating ASD.
• To enhance the accuracy of ASD diagnosis and treatment through AI applications.

Main Methods:

• A narrative review of approximately 100 research articles published between 2010 and 2024.
• Articles were selected based on defined inclusion and exclusion criteria.
• The review synthesized findings across medical, occupational, vocational, psychological, family, and rehabilitation engineering domains.

Main Results:

• AI demonstrates a significant role in identifying early indicators of ASD, aligning with UN Sustainable Development Goals.
• AI analyses of genetic, neuroimaging, behavioral, and electronic health record data enable early and individualized ASD detection.
• Machine learning (ML) algorithms achieve high accuracy in differentiating ASD from neurotypical development and other disorders, supporting timely interventions.

Conclusions:

• AI-driven interventions, including virtual reality and robot-assisted therapies, enhance social and communication skills in individuals with ASD.
• AI offers potential for improved diagnostic accuracy, telehealth implementation, and customized treatment plans for ASD.
• Challenges such as data privacy and interpretability need to be addressed for widespread AI adoption in ASD care.