57 - Structural connectome analysis in a piglet model of chronic hypoxia.

Poster Number: 57
Publication Number: 57.446

Van Lam, Children's National Hospital, Washington, DC, United States; Jingang Li, Children's National Health System, Washington, DC, United States; Artur Agaronyan, Howard University, College Park, MD, United States; Stephen Lin, Howard University, Washington, DC, United States; Stephen Xu, George Washington University School of Medicine and Health Sciences, Bethesda, MD, United States; Ameya Sinha, College of William and Mary, Edina, MN, United States; Tsang-Wei Tu, Howard University College of Medicine, Washington, DC, United States; Nobuyuki Ishibashi, Children's National Health System, Washington, DC, United States

Background: Children with congenital heart diseases (CHD) are at high risk of white matter injuries and neurodevelopment affecting their brain maturation. The pig brain likely resembles the human brain in anatomy and development, comparing to commonly used small laboratory animals. We have previously demonstrated that a piglet model of developmental hypoxemia (Hx) share similar structural and cellular abnormalities in the CHD infant including reduction in brain size, white matter volumes, gyrification, and subventricular zone width.

Objective: Connectome analysis is an advanced, unbiased computation method to study comprehensive brain development and a wide range of neurodevelopmental disorders. Thus, this study investigates the impact of Hx on the pig brain connectivity using structural connectome.

Design/Methods: High-resolution diffusion tensor images were collected from postnatal pig brains in normoxia (Nx) and Hx (n=4 each) and used as inputs to Mrtrix3 to perform tractography. The process follows a series of denoising, determining white-matter and grey matter boundary, and reconstructing fiber orientation distribution maps. An in-house 3D piglet brain atlases with 68 cortical gray matter regions were then used to generate binary and weighted connectivity matrices. Each brain matrix was compared to thousand random networks to validate their biological network properties.

Results: Several network parameters were extracted from the matrices and analyzed using a Brain Connectivity toolbox in MATLAB. Clustering coefficient and path length in brain networks of both groups was significantly higher than in random networks (student’s test, p< 0.0001) suggesting the high capacity of exchanging information. Network efficiency though was preserved in Nx and Hx group, they both obtained small-word characteristics (σ >1). Therefore, brain regions are densely connected through relatively few intermediate steps. Normoxia brains achieved higher rich club coefficient than Hx brains (1.26 > 1.1, paired t-test, p< 0.01) with 19 rich club regions comparing to 10 in Hx. Sseveral loss of connection between cortical regions vs. basal ganglia or vs. limbic system was also found in Hx group, demonstrating that developmental hypoxia alters global network topology in the pig brain.

Conclusion(s):

Various clinical neuroimaging studies have shown alterations of brain connectivity in newborns and infants with CHD. Thus, our results indicate that network-based connectomics approach in the piglet brain improves our understanding of cellular and molecular events underlying brain network alterations observed largely in children with CHD.