Related Experiment Video
Updated: May 27, 2026
07:49
Time-lapse Live Imaging of Clonally Related Neural Progenitor Cells in the Developing Zebrafish Forebrain
Published on: April 6, 2011
Spatio-temporal reconstruction of early brain developmental trajectories via self-supervised learning
Chenglin Ning1, Tianshu Zheng1, Yiwei Chen1
1Key Laboratory for Biomedical Engineering of Ministry of Education, Department of Biomedical Engineering, College of Biomedical Engineering & Instrument Science, Zhejiang University, Hangzhou 310027, China.
Medical Image Analysis
|May 25, 2026
Summary
STRIDER reconstructs complete brain development trajectories from sparse MRI data, improving neurodevelopmental research and autism risk prediction. This novel framework enhances understanding of early brain changes and cognitive development.
Area of Science:
- Neuroscience
- Developmental Biology
- Artificial Intelligence
Background:
- Understanding early brain development is vital for identifying neurodevelopmental disorders.
- Longitudinal magnetic resonance imaging (MRI) studies face challenges with incomplete data due to acquisition difficulties.
- Existing methods struggle to reconstruct complete brain developmental trajectories from sparse longitudinal data.
Purpose of the Study:
- To introduce STRIDER, a novel framework for reconstructing complete brain developmental trajectories from extremely sparse longitudinal MRI data.
- To address the challenge of incomplete follow-up data in early-life neurodevelopmental studies.
- To enable accurate estimation of brain markers at past and future time points.
Main Methods:
- STRIDER integrates graph neural networks within a transformer architecture.
- The framework treats brain development reconstruction as a multivariate time series imputation problem.
- It utilizes spatial and temporal graphs with a Temporal Spatial Layer (TSL) to aggregate information from sparse observations.
Main Results:
- STRIDER significantly reduces mean absolute error (MAE) by 18.4%-54.6% compared to baseline methods.
- Imputed trajectories enhance cognitive prediction accuracy by 46.5%-110.7%.
- Brain age gap (BAG) derived from STRIDER data shows improved prediction for autism spectrum disorder (ASD) risk (48.7%-101.3% increase).
Conclusions:
- STRIDER effectively reconstructs complete and heterogeneous early brain developmental trajectories from sparse longitudinal data.
- The framework demonstrates robust performance and generalizability across different cohorts.
- STRIDER offers significant potential for advancing neurodevelopmental research and clinical applications, including early risk identification for ASD.
