Abstract
Study Type: Evidence Map / Evidence Mapping Study Background Pediatric developmental brain disorders include a heterogeneous group of conditions such as neurodevelopmental disorders (NDD), congenital brain malformations (CBM), and inherited metabolic diseases (IEM). Despite diverse etiologies, these disorders often involve convergent neurodevelopmental pathways affecting brain structure and function. Quantitative neuroimaging provides important tools for investigating these mechanisms by enabling measurement of structural, functional, and metabolic brain alterations. In recent years, artificial intelligence (AI) has increasingly been applied to neuroimaging analysis. Machine learning and deep learning methods enable automated feature extraction, disease classification, prognosis prediction, and image generation. Although the number of studies has grown rapidly, the research landscape remains fragmented across disorders, imaging modalities, and analytical approaches. A comprehensive overview of methodological trends and research distribution in this area is currently lacking. Evidence mapping provides a structured approach to characterize the scope and distribution of research without conducting formal effectiveness synthesis. This study aims to map the existing literature on AI applications in quantitative neuroimaging of pediatric developmental brain disorders over the past decade. Objective The objective of this study is to describe the research landscape of artificial intelligence applied to quantitative neuroimaging in pediatric developmental brain disorders between 2015 and 2025. Specifically, the study will characterize publication trends, AI methodologies, imaging modalities, clinical tasks, and key methodological characteristics of the included studies. Data Sources A systematic search will be conducted in the following databases: PubMed Embase IEEE Xplore The search will include studies published between January 2015 and December 2025. Eligible publications include peer-reviewed journal articles and conference papers reporting original research. Only English-language publications will be included. Study Selection Records retrieved from the databases will undergo title and abstract screening followed by full-text review. Screening will be performed independently by two reviewers, with disagreements resolved through discussion. Data Extraction Included studies will be coded according to a predefined framework including: bibliographic information (publication year, journal, country) AI methodology and algorithms used imaging modality (e.g., MRI, PET, EEG, ultrasound, or multimodal imaging) primary clinical task addressed by the model study characteristics such as population type, sample size, study design, external validation, multi-center data, code availability, and explainability analysis. Evidence Mapping Extracted data will be synthesized using descriptive statistics and visual mapping techniques to illustrate patterns in research activity and identify potential evidence gaps within the field.