Cohort protocol: risk assessment of maternal inflammation and early brain development in infants and young children based on multi-source data modeling

TL;DR

This study assesses how maternal inflammation during pregnancy affects early brain development in children using multi-source data and AI modeling.

Contribution

A novel deep learning model is proposed to evaluate early brain development risks based on gene-image-environment-behavior interactions.

Findings

A cohort of 360 mother-child pairs was followed to assess maternal inflammation's impact on brain development.

Deep learning models were used to integrate multi-source data for early risk assessment of brain development disorders.

The study aims to improve early identification and intervention for developmental issues in infants and young children.

Abstract

Infancy and early childhood are the key stage for the rapid development of brain structure and function, and brain development at this stage has a profound impact on the future intelligence, behavior and health of individuals. A growing body of research suggests that maternal inflammation, as a potential environmental factor, may affect brain development in infants and young children through a variety of mechanisms. Therefore, it is of great significance to evaluate the risk of maternal inflammation to early brain development in infants and young children based on multi-source data modeling to understand the mechanism of early development and prevent brain development disorders.

Between December 2021 and May 2024, 360 pairs of pregnant women and their offspring were recruited into the Xiamen Children's Brain Development Cohort. Pregnant women's exposure during pregnancy was collected through standardized and structured questionnaires and medical records. All children were followed up to 3 years of age. We administered questionnaires, behavioral assessments, and performed neuroimaging. Environmental exposures during infancy and early childhood were collected. Children's cognitive, emotional, and linguistic development was evaluated, and blood samples were obtained for whole-exome sequencing and exposure-related biomarker analysis.

In this study, we used deep learning artificial intelligence to construct an early risk assessment model for infant brain development based on the developmental trajectory and developmental results of early brain structure, function, and connections under the complex interaction of “gene-image-environment-behavior” multi-factors, which can improve the early identification and precise intervention of problems in this period, and improve infants cognitive learning and work performance in childhood, adolescence and even adulthood.

https://www.clinicaltrials.gov/; identifier [NCT05040542].

Figures