Mapping the asynchrony of cortical maturation in the infant brain: A MRI multi-parametric clustering approach

Jessica Lebenberg, Jean-François Mangin, Bertrand Thirion, Cyril Poupon, Lucie Hertz-Pannier, François Leroy, Parvaneh Adibpour, Ghislaine Dehaene-Lambertz, Jessica Dubois

TL;DR

This study uses multi-parametric MRI to map the asynchronous maturation of the infant brain, revealing distinct patterns of development across cortical regions. By combining T1, T2, and diffusion MRI data, researchers identified clusters of voxels with similar microstructural properties, providing a novel in vivo approach to understanding early brain development.

What Problem Does This Paper Address?

Understanding how the infant brain matures is crucial for comprehending cognitive development and identifying potential developmental disorders. While post-mortem studies have provided insights into cortical maturation, in vivo investigations in healthy infants have been limited, especially during the critical first few months of life. This period is characterized by rapid changes in cortical microstructure, including dendritic arborization and myelination, which are thought to underpin emerging cognitive abilities. However, the asynchronous nature of these changes across different brain regions presents a significant challenge for mapping and understanding this complex developmental process. Traditional MRI approaches often struggle with the low contrast and complex folding of the immature infant brain, and existing methods for mapping maturation have limitations in terms of spatial resolution, acquisition time, and the ability to capture the multi-faceted nature of microstructural development.

Methods

The researchers employed a multi-parametric Magnetic Resonance Imaging (MRI) approach to map cortical maturation. This involved:

  • Quantitative MRI Acquisition: They acquired T1, T2, and diffusion tensor imaging (DTI) data using echo-planar imaging (EPI) sequences to minimize acquisition time for sleeping infants.
  • Distortion Correction: A crucial step involved correcting for geometric distortions inherent in EPI images, which are particularly problematic in the infant brain due to its developing structure and skull interface. This was achieved through a non-linear registration strategy that aligned the distorted parametric maps to undistorted anatomical T2-weighted images.
  • Cortical Ribbon Identification: A semi-automatic procedure was used to precisely identify the cortical ribbon (the outermost layer of the brain) on both anatomical and parametric maps, accounting for the thinness and variability of the infant cortex.
  • Univariate Analysis: Individual T1, T2, and axial diffusivity ($\lambda_{//}$) maps were analyzed to observe regional differences in maturation.
  • Inter-Subject Registration: To enable group-level analysis, a two-step registration strategy (DISCO + DARTEL) was used to align the brains of different infants, focusing on sulcal landmarks.
  • Multi-parametric Clustering: The core of the study involved a Gaussian Mixture Model (GMM) clustering algorithm applied to the T1, T2, and $\lambda_{//}$ parameters across all cortical voxels. This approach leverages the complementary information from different MRI parameters to group voxels with similar microstructural properties, avoiding reliance on pre-defined regions of interest. The clustering was optimized by systematically evaluating different covariance matrix models, parameter combinations, and numbers of clusters using Silhouette coefficients and cluster volume variability.

Datasets

The study utilized MRI data from 17 healthy term-born infants (10 boys, 7 girls) aged between 3 and 21 weeks (maturational age). No new benchmark datasets were introduced; the study focused on analyzing existing infant MRI data with novel methods.

Key Insights and Ideas

  • Asynchronous Cortical Maturation: The study strongly supports the idea that cortical maturation is not a uniform process but rather occurs asynchronously across different brain regions.
  • Multi-parametric Approach Superiority: Combining multiple MRI parameters (T1, T2, and $\lambda_{//}$) provided a more comprehensive and robust picture of cortical maturation than analyzing single parameters alone. This approach captured distinct microstructural profiles that would be missed by univariate methods.
  • Clustering as a Powerful Tool: The GMM clustering algorithm effectively identified distinct groups of cortical voxels with similar developmental trajectories, offering a data-driven way to parcellate the cortex based on maturation.
  • Primary Regions Mature First: Similar to findings from post-mortem studies of myelination, the in vivo data indicated that primary sensorimotor areas mature earlier than associative cortices.
  • Individual Maturation Trajectories: While group-level trends were clear, the study also highlighted significant inter-individual variability, suggesting that each infant brain may follow a unique maturation pathway.

Key Results

  • The clustering analysis identified 5 distinct clusters of cortical voxels, each representing a different microstructural and maturational stage.
  • These clusters revealed three main patterns of maturation: two immature patterns (age-related decreases in T1 and $\lambda_{//}$), two mature patterns (age-related increases), and one intermediate pattern.
  • Primary sensorimotor regions (e.g., visual cortex V1, auditory cortex A1, somatosensory cortex S1, motor cortex M1) were consistently found to be the most mature regions across infants.
  • Associative regions, such as the superior temporal gyrus and inferior frontal gyrus, showed delayed maturation.
  • The study demonstrated that the proposed registration and clustering methods were robust enough to reveal these subtle differences in maturation, even with the challenges of infant brain imaging.

Limitations and Future Work

  • Spatial Resolution: The spatial resolution of the parametric maps (1.8mm isotropic) limited the ability to distinguish between cortical layers. Future work with ultra-high field imaging (e.g., 7T) could provide finer-grained mapping.
  • Partial Volume Effects: While efforts were made to isolate the cortical ribbon, some contamination from adjacent white matter voxels was possible. However, the authors argue this is unlikely to significantly alter the results given the synchronous maturation of cortex and white matter.
  • Number of Clusters: Selecting the optimal number of clusters can be challenging. While 5 clusters provided a good balance, the authors acknowledge that other numbers might be relevant for different developmental periods.
  • Future Directions: The authors suggest extending this multi-parametric approach to other developmental periods and exploring the relationship between in vivo cortical maturation and the development of functional systems and cognitive abilities. Further studies with larger cohorts and longitudinal data are needed to confirm these findings and investigate the influence of factors like brain growth and cortical folding.

Summary updated

Abstract

Highlights • The cortical maturation was studied in infants between 1 and 5 months of age. • Diffusion and relaxometry MRI were analyzed in univariate and multivariate manner. • Clustering analyses provided reliable results both at the subject and group levels. • Distinct maturation profiles were shown across regions of the infant cortex. Abstract While the main neural networks are in place at term birth, intense changes in cortical microstructure occur during early infancy with the development of dendritic arborization, synaptogenesis and fiber myelination . These maturational processes are thought to relate to behavioral acquisitions and the development of cognitive abilities. Nevertheless, in vivo investigations of such relationships are still lacking in healthy infants. To bridge this gap, we aimed to study the cortical maturation using non-invasive Magnetic Resonance Imaging , over a largely unexplored period (1–5 post-natal months). In a first univariate step, we focused on different quantitative parameters: longitudinal relaxation time (T1), transverse relaxation time (T2), and axial diffusivity from diffusion tensor imaging (λ // ) These individual maps, acquired with echo-planar imaging to limit the acquisition time, showed spatial distortions that were first corrected to reliably match the thin cortical ribbon identified on high-resolution T2-weighted images. Averaged maps were also computed over the infants group to summarize the parameter characteristics during early infancy. In a second step, we considered a multi-parametric approach that leverages parameters complementarity, avoids reliance on pre-defined regions of interest, and does not require spatial constraints. Our clustering strategy allowed us to group cortical voxels over all infants in 5 clusters with distinct microstructural T1 and λ // properties The cluster maps over individual cortical surfaces and over the group were in sound agreement with benchmark post mortem studies of sub-cortical white matter myelination, showing a progressive maturation of 1) primary sensori-motor areas, 2) adjacent unimodal associative cortices, and 3) higher-order associative regions. This study thus opens a consistent approach to study cortical maturation in vivo . Graphical abstract Download : Download high-res image (246KB) Download : Download full-size image Previous article in issue Next article in issue
BibTeX
@article{lebenberg2019mapping,
  author = {Jessica Lebenberg and Jean-François Mangin and Bertrand Thirion and Cyril Poupon and Lucie Hertz-Pannier and François Leroy and Parvaneh Adibpour and Ghislaine Dehaene-Lambertz and Jessica Dubois},
  title = {{Mapping the asynchrony of cortical maturation in the infant brain: A MRI multi-parametric clustering approach}},
  journal = {NeuroImage},
  year = {2019},
  doi = {10.1016/j.neuroimage.2018.07.022},
}
Figure