Leveraging AI-Driven Neuroimaging Biomarkers for Early Detection and Social Function Prediction in Autism Spectrum
Evgenia Gkintoni1, Maria Panagioti2, Stephanos P Vassilopoulos1
1Department of Educational Sciences and Social Work, University of Patras, 26504 Patras, Greece.
Healthcare (Basel, Switzerland)
|August 14, 2025
Summary
Artificial intelligence (AI) combined with neuroimaging shows promise for early autism spectrum disorder (ASD) detection. Machine learning models achieve high accuracy, identifying critical developmental windows for intervention.
Area of Science:
- Neuroscience
- Computer Science
- Developmental Pediatrics
Background:
- Systematic review of artificial intelligence (AI) applications in neuroimaging for autism spectrum disorder (ASD).
- Addresses AI's role in biomarker optimization, modality integration, and prediction of developmental trajectories.
- Focuses on enhancing multimodal data for earlier ASD detection and improved outcomes.
Purpose of the Study:
- To examine AI applications in neuroimaging for autism spectrum disorder (ASD).
- To assess AI's potential in early detection, prognosis, and intervention efficacy measurement.
- To identify challenges and opportunities for clinical translation of AI-driven neuroimaging in ASD.
Main Methods:
- Comprehensive systematic literature search across 8 databases following PRISMA guidelines.
- Analysis of 146 selected studies from an initial 1872 records.
- Systematic analysis to address research questions on AI neuroimaging approaches in ASD.
Main Results:
- AI and neuroimaging demonstrate significant potential for early ASD detection, particularly using electroencephalography (EEG).
- Machine learning classifiers achieved 85-99% diagnostic accuracy using neural features.
- Multimodal approaches and longitudinal analyses enhance predictive capabilities and track developmental trajectories.
Conclusions:
- AI-driven neuroimaging biomarkers offer a promising avenue for detecting ASD before behavioral symptoms manifest.
- Objective measures of intervention efficacy can be provided by AI neuroimaging.
- Standardization, diverse sampling, and clinical validation are crucial for translating AI findings into clinical practice for earlier ASD intervention.
