[Paper Review] Detailed delineation of the fetal brain in diffusion MRI via multi-task learning

Davood Karimi, Camilo Calixto+7 more|arXiv (Cornell University)|Aug 26, 2024

Advanced Neuroimaging Techniques and ApplicationsMedicine3 citations

TL;DR

This study introduces a multi-task deep learning framework for automated, unified segmentation of fetal brain tissue, white matter tracts, and cortical/subcortical structures directly in diffusion MRI. The method achieves high accuracy—mean Dice scores of 0.865 (tissue), 0.825 (tracts), and 0.819 (parcellation)—enabling reproducible, fast analysis of low-quality fetal dMRI data, advancing fetal neuroimaging and connectivity studies.

ABSTRACT

Diffusion-weighted MRI is increasingly used to study the normal and abnormal development of fetal brain in-utero. Recent studies have shown that dMRI can offer invaluable insights into the neurodevelopmental processes in the fetal stage. However, because of the low data quality and rapid brain development, reliable analysis of fetal dMRI data requires dedicated computational methods that are currently unavailable. The lack of automated methods for fast, accurate, and reproducible data analysis has seriously limited our ability to tap the potential of fetal brain dMRI for medical and scientific applications. In this work, we developed and validated a unified computational framework to (1) segment the brain tissue into white matter, cortical/subcortical gray matter, and cerebrospinal fluid, (2) segment 31 distinct white matter tracts, and (3) parcellate the brain's cortex and delineate the deep gray nuclei and white matter structures into 96 anatomically meaningful regions. We utilized a set of manual, semi-automatic, and automatic approaches to annotate 97 fetal brains. Using these labels, we developed and validated a multi-task deep learning method to perform the three computations. Our evaluations show that the new method can accurately carry out all three tasks, achieving a mean Dice similarity coefficient of 0.865 on tissue segmentation, 0.825 on white matter tract segmentation, and 0.819 on parcellation. The proposed method can greatly advance the field of fetal neuroimaging as it can lead to substantial improvements in fetal brain tractography, tract-specific analysis, and structural connectivity assessment.

Motivation & Objective

• To address the lack of automated, reliable computational methods for analyzing low-quality, rapidly developing fetal diffusion MRI (dMRI) data.
• To develop a unified deep learning framework capable of performing three distinct but related tasks: tissue segmentation, white matter tract segmentation, and cortical/subcortical parcellation.
• To overcome data scarcity and labeling challenges in fetal dMRI by leveraging multi-task learning to improve model generalization and performance.
• To enable accurate, reproducible, and fast processing of fetal dMRI for tractography, structural connectivity assessment, and neurodevelopmental research.

Proposed method

• A multi-task deep learning model was trained to simultaneously perform tissue segmentation, white matter tract segmentation, and cortical parcellation using a shared encoder and task-specific heads.
• The model was trained on a dataset of 97 manually, semi-automatically, and automatically annotated fetal brains across multiple gestational ages.
• Task-specific loss functions were combined using learnable task weights to balance uncertainty and optimize performance across the three tasks.
• The framework used a U-Net-based architecture with multi-scale features to handle the low signal-to-noise ratio and anatomical variability in fetal dMRI.
• Model training incorporated data augmentation and normalization techniques to improve robustness to image intensity variations and motion artifacts.
• The method was validated using both quantitative metrics (Dice similarity coefficient) and expert visual assessment against anatomical landmarks.

Experimental results

Research questions

• RQ1 Can a single deep learning model achieve high-accuracy segmentation of fetal brain tissue, white matter tracts, and cortical regions in diffusion MRI?
• RQ2 How does multi-task learning improve performance on fetal dMRI segmentation when training data is limited and noisy?
• RQ3 To what extent can the model generalize across gestational ages despite rapid brain development and low image quality?
• RQ4 How does the model’s performance compare to existing methods for adult or fetal structural MRI, particularly in parcellation and tract segmentation?
• RQ5 Can the model produce more accurate segmentations than the training labels, especially in regions with noisy or inconsistent annotations?

Key findings

• The proposed multi-task learning model achieved a mean Dice similarity coefficient of 0.865 for tissue segmentation (WM, CGM, SGM, CSF), indicating high accuracy in classifying major brain tissue types.
• For white matter tract segmentation, the method achieved a mean Dice score of 0.825, demonstrating reliable delineation of 31 complex tracts despite low image quality and motion artifacts.
• The parcellation task achieved a mean Dice score of 0.819, with expert review confirming that predictions often corrected local errors present in the training labels.
• The model showed robustness to label noise, producing more anatomically consistent segmentations than the original training annotations in some cases.
• Task weights were learned during training, indicating higher uncertainty in tract segmentation due to the lack of mutual exclusivity among tract labels compared to tissue and parcellation tasks.
• The method enables automated, fast, and reproducible processing of fetal dMRI, significantly advancing fetal neuroimaging for developmental and clinical research.