Global Network Organization of the Fetal Functional Connectome

Abstract

Recent advances in brain imaging have enabled non-invasive in vivo assessment of the fetal brain. Characterizing brain development in healthy fetuses provides baseline measures for identifying deviations in brain function in high-risk clinical groups. We examined 110 resting state MRI data sets from fetuses at 19 to 40 weeks’ gestation. Using graph-theoretic techniques, we characterized global organizational features of the fetal functional connectome and their prenatal trajectories. Topological features related to network integration (i.e., global efficiency) and segregation (i.e., clustering) were assessed. Fetal networks exhibited small-world topology, showing high clustering and short average path length relative to reference networks. Likewise, fetal networks’ quantitative small world indices met criteria for small-worldness (σ > 1, ω = [−0.5 0.5]). Along with this, fetal networks demonstrated global and local efficiency, economy, and modularity. A right-tailed degree distribution, suggesting the presence of central areas that are more highly connected to other regions, was also observed. Metrics, however, were not static during gestation; measures associated with segregation—local efficiency and modularity—decreased with advancing gestational age. Altogether, these suggest that the neural circuitry underpinning the brain’s ability to segregate and integrate information exists as early as the late 2nd trimester of pregnancy and reorganizes during the prenatal period. Significance statement. Mounting evidence for the fetal origins of some neurodevelopmental disorders underscores the importance of identifying features of healthy fetal brain functional development. Alterations in prenatal brain connectomics may serve as early markers for identifying fetal-onset neurodevelopmental disorders, which in turn provide improved surveillance of at-risk fetuses and support the initiation of early interventions.

Functional Connectivity-Derived Optimal Gestational-Age Cut Points for Fetal Brain Network Maturity

“…RS-fMRI data were collected as part of a prospective, longitudinal study at Children’s National in Washington DC investigating pre- and postnatal brain development in the setting of complex congenital heart disease. We previously described global network topology in this cohort in a recently published study [ 5 ]. Conventional T2 for all fetuses revealed structurally normal brains.…”

“…Fetal RS-fMRI data were preprocessed to attenuate the effects of noise on the measured blood oxygen level dependent (BOLD) signals. These steps were previously described in [ 5 , 18 ]. EPI images were slice time corrected, followed by removal of the first four RS volumes to allow for magnetic gradients to stabilize.…”

“…The timing of developmental events can help differentiate aberrant from neurotypical processes. Thus, perinatal RS-fMRI aims to detect when functional networks emerge (e.g., appearance of visual and sensorimotor RSNs) and to identify when networks mature (e.g., become more similar to newborn architecture) [ 3 , 4 , 5 , 6 ].…”

“…Systems-wide changes in network integration and segregation associated with brain maturation have been described using graph-theoretic techniques, a quantitative framework for describing complex networks such as the brain [ 10 , 11 ]. Studies have shown the emergence of small-world network topology as early as the second trimester and reduced segregation with advancing GA based on decreasing modularity, normalized clustering, and local efficiency [ 5 , 12 , 13 ]. While trends have been described, it is unclear when the transition to mature network patterns occurs.…”

“…The goal of our study was to estimate optimal gestational-age (GA) cut points, or transition periods, where we can subdivide fetal groups based on age-associated changes in network topology and perform meaningful between-group comparisons. In this study, cut points were defined for global functional network metrics that we have recently described; these metrics are normalized clustering and path length, global and local efficiency, modularity, and the small-world index, metrics we have previously described in this fetal cohort [ 5 ]. We selected GAs between 28 and 33 weeks as potential cut points based on previous studies reporting transitional electrophysiologic events [ 16 , 17 ] and resting-state studies that demonstrated significant connectivity changes around this period [ 4 , 9 ].…”