Dynamics of cortical morphology during childhood development – SulcalGRIDS

SulcalGRIDS is divided into two specific Tasks. In Aim#1, the objective will be to infer the spatiotemporal organization of the cortical folds during typical brain development, in fetuses and young children. In Aim#2, we will then postulate biologically plausible rules governing the geometry of the model, and the relationship between this geometry and the underlying axonal connectivity.

-The first major outcome of SulcalGRIDS is the unique fetal MRI dataset resulting from a sustained collaborative effort between INT and AP-HM (Pr. M.Milh and Pr. N. Girard). More specifically, we built the exceptional database of fetal MRI including the anatomical and diffusion MRIs of more than 1000 fetuses acquired at the Timone hospital between 2008 and 2021 as part of clinical routine, after agreement from the ethics committee of Aix-Marseille University (2022-04-14-003), the CNIL (2209585) and the AP-HM data management service (PADS21-170). From this large database, we identified 381 fetal MRI without any visible anomaly that will serve as controls, and 149 fetus with corpus callosum anomaly. We already initiated the recording of the behavioral assessment of the control fetuses in the MultinormFet project (PHRC-I, PI: Pr. N. Girard, clinicaltrials.gov/study/NCT06081036). We will follow the same approach for the evaluation of children with ACC.

-The first study based on this dataset has been submitted for publication. We implemented the first neurodevelopmental study targeting the cortical trajectory of typically developing perinatal subjects, combining MRIs from both fetal and postnatal participants. We conducted a normative modeling analysis on a sample of 607 subjects, age ranged 24 to 45 weeks post-conception, to observe changes that arise as participants traverse the birth barrier. We observed that the trajectories of global surface area and several volumetric features, including total gray matter, white matter, brainstem, cerebellum and hippocampi, follow distinct but continuous patterns during this transition. The current study demonstrates the presence of unique neurodevelopmental patterns for several structural features during the perinatal period, and confirms that not all features are affected in the same way as they cross the birth barrier.

-In parallel, the PhD of Maxime Dieudonné resulted in critical advances regarding the estimation of the depth of cortical sulci. The interactions between the biological mechanisms underlying brain morphogenesis remain largely elusive. Sulcal depth is used in an increasing number of fundamental, methodological and clinically-oriented publications. All the methods from the literature have been designed and evaluated on adult brains only. The potential influence of global brain size on the sulcal depth estimation obtained at each vertex of the cortical surface was not considered in any of these methods. In this present work, we contribute to this field of research by 1) providing the first quantitative characterization of the influence of brain size on sulcal depth estimations; 2) introducing a formalization of the problem, and a scale-invariant sulcal depth estimation method; 3) proposing a validation framework and sharing with the community the benchmark data.

1-An international collaboration allowing us to run large-scale multicentric studies on fetal MRI acquired in clinical settings. We did submit a proposal to the ERA-NET NEURON Call for Joint Transnational Research Projects on Neurodevelopmental Disorders. The title is Multi-centric study of Fetal Abnormal Cortical Trajectory with standardized and privacy-preserving method on fetal MRI, the coordinator is Elisenda Eixarch from Fundació Clínic per a la Recerca Biomédica (FCRB) in Barcelona, Spain. I am (G.Auzias) the PI of the French partner. The project is running until 2025.

2-The PHRC multinormfet (PI:N.Girard) has been funded in 2023. The main aim of this project is to compute and make available to the community normative curves of the evolution of brain tissue volumes derived from over 500 fetal MRI scans acquired as part of a clinical routine on 7 different MRI scanners, spread across 4 university hospital centers. To ensure that our normative curves are representative of normal development, we will: 1) Include a large number of patients from several centers, thereby guaranteeing good statistical robustness and taking into account potential biases linked to acquisition conditions. 2) Include only images of fetuses showing normal postnatal development. To do this, we propose to include only MRIs of fetuses acquired before 2019 for which the babies are now aged 3 years or more, and to call the parents back to obtain an assessment of development at 3 years. We have set up multidisciplinary teams in each of the 4 partner centers, bringing together all the necessary medical skills, i.e. radiologists, neuropaediatricians and obstetricians. These curves, which are representative of normal brain development, will eventually be used to develop tools to help diagnose scanned fetuses, by providing a quantitative characterization of the normality of volumetric measurements.

-

SulcalGRIDS aims at elaborating a new model of cortical morphogenesis that could explain how the human cerebral cortex shape complexity emerges during development. Such a model would be instrumental to develop novel biomarkers of early atypical development and to better understand the pathophysiology of neurodevelopmental disorders.
We will build on our recent works on specific cortical points, the sulcal pits, that correspond to singularities from which sulci emerge during in utero and early post-natal cortical maturation phases. We developed an algorithm to extract sulcal pits, and provide fine-grain information about both their spatial location and the milestones of their emergence. Grounded on these advances, SulcalGRIDS introduces a new spatiotemporal model of cortical development allowing to precisely characterize the dynamics of patterning. SulcalGRIDS is divided into two specific Tasks.
In Task#1, the objective will be to infer the spatiotemporal organization of the sulcal pits during typical brain development, in fetuses and young children.
In Task#2, we will then postulate biologically plausible rules governing the geometry of the model, and the relationship between this geometry and the underlying axonal connectivity. We will test these assumptions using neurodevelopmental pathologies known to affect both the shape and wiring of human cortex during ante-natal or early post-natal maturation: Corpus Callosum Agenesis (CCA) in fetuses and Autism Spectrum Disorders (ASD) in children.

To run his interdisciplinary project, G Auzias will benefit from the joint efforts of two research teams of the Institut de Neurosciences de la Timone (INT). SCALP (C Deruelle) is a neuroscience team with a long experience in neurodevelopmental disorders. MeCA (O Coulon) is a computational anatomy team with a solid experience in methods for cortical morphometry, including for fetal data (J Lefèvre). Both teams have a strong partnership with the Neuroradiology Unit of CHU Timone (Pr. N Girard) where all clinical MR protocols are performed with specific guidelines, including fetuses with CCA that are followed after birth by Pr. M.Milh. This project will benefit from the image processing tools dedicated to fetal MRI data in the team lead by Pr. F Rousseau. SulcalGRIDS will have access to the INT Neurocomputing Center, with a large-scale cluster (548 cores, 170To storage) and two computing engineers specialized in neuroimaging.

This project aims at bringing together research in neuroscience and data analysis, and G Auzias has the background to coordinate such an interdisciplinary effort. This is a strategic scientific operation in the emergence of a major biomedical imaging research site on the Timone site (Fac. of Medicine and CHU, high-Field clinical and preclinical MRI, primatology).
It is also timely as (i) the structuration of neuroimaging data is a priority for INT and will be supported by the recent integration (April 2018) of a new research engineer dedicated to this task; (ii) collaboration with the clinical radiology unit (Pr. N Girard) is now fully mature to organize a systematic collections of MRI data in pathological conditions; (iii) several international teams have expressed interest in our innovative sulcal morphometry methods (e.g. ENIGMA consortium).
By unveiling the rules governing human brain folding, SulcalGRIDS is a basic science project attacking a long-standing mystery in brain sciences. It is critical to elucidate the functional organization of large-scale brain networks and to better characterize early atypical brain development associated with many debilitating developmental pathologies and in particular Autism.