The developing Human Connectome Project (dHCP) automated resting-state functional processing framework for newborn...

Area of Science:

• Neuroimaging
• Developmental Neuroscience
• Computational Neuroscience

Background:

• The developing Human Connectome Project (dHCP) aims to map brain development in early life.
• Neonatal neuroimaging presents unique challenges, including motion and low signal-to-noise ratio (SNR).
• Existing preprocessing pipelines often struggle with the complexities of neonatal resting-state fMRI data.

Purpose of the Study:

• To present an automated, robust, and quality-assured pipeline for minimally preprocessing neonatal resting-state fMRI data.
• To address specific challenges in neonatal data, such as head motion and image distortion.
• To improve the reliability and quality of functional brain network analysis in neonates.

Main Methods:

• Development of an automated preprocessing pipeline integrating slice-to-volume motion correction and distortion correction.
• Implementation of a robust multimodal registration approach.
• Application of Independent Component Analysis (ICA)-based denoising and an automated quality control (QC) framework.
• Evaluation on a large cohort of dHCP subjects.

Main Results:

• The pipeline demonstrates robust performance with low failure rates and high quality assurance.
• Significant reduction in motion-related artifacts and distortions was achieved.
• Substantial improvements in SNR and detection of high-quality resting-state networks (RSNs) were observed.
• The processing refinements enhance the analysis of neonatal brain connectivity.

Conclusions:

• The developed automated pipeline effectively addresses challenges in neonatal fMRI preprocessing.
• The pipeline enhances data quality, leading to more reliable identification of brain networks in neonates.
• This work provides a valuable tool for advancing research in early brain development and connectomics.