Synopsis

Using diffusion MRI, we investigated white matter microstructure and structural brain connectivity in 11-year old children born with extremely low birth weight (ELBW) in comparison with full-term born children. Microstructural white matter properties were investigated within the tract skeleton, and constrained spherical deconvolution based probabilistic tractography and gray matter parcellation were used to reconstruct structural brain connectivity networks. We found decreased integrity and complexity of the white matter microstructure in ELBW, and increased segregation of the structural brain connectivity networks. In addition, the microstructural changes were associated with the administration of antenatal corticosteroids and with retinopathy of prematurity.

Introduction

Preterm birth rate in Europe varies between 5-10% of all births¹. As the brain rapidly grows and develops during the last trimester of gestation², very preterm infants are at greater risk for abnormalities in brain development³ leading to motor, cognitive, and social deficits such as cerebral palsy, learning disabilities, and autism spectrum disorder^4-7. Therefore, it is important to understand the neurobiological abnormalities related to prematurity, and to recognize the neonatal risk factors leading to neurodevelopmental impairments.

Diffusion-weighted (DW) magnetic resonance imaging (MRI)⁸ has enabled noninvasive investigation of the neural tracts and their microstructural properties in vivo. In this study, we investigated white matter microstructure and structural brain connectivity networks in 11-year old children born with extremely low birth weight (ELBW) from Helsinki region, a geographically defined subgroup of a national ELBW cohort⁹, and compared them to full-term (FT) born children.

Methods

Diffusion MRI data were acquired from 37 ELBW children (aged 11.3±1.2 years, 25 born extremely preterm <28 gestational weeks) and 29 FT children (aged 11.6±0.7 years) with a Philips Achieva 1.5T machine by using 20 gradient orientations and a b‑value of 800 s/mm² with a resolution of 1.75 mm × 1.75 mm × 2 mm. T₁‑weighted MRI data were acquired with a resolution of 0.94 mm × 0.94 mm × 1 mm. The study protocol was approved by the Ethics Committee of Helsinki and Uusimaa Hospital District, Finland. The parents and the child gave written informed consent before the investigation.

The DW data were corrected for subject motion¹⁰, eddy current induced¹¹, and echo planar imaging distortions¹². White matter tract skeleton was reconstructed as introduced in tract-based spatial statistics (TBSS)¹³. Global microstructural changes in fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), radial diffusivity (RD), and coefficient of planarity (CP) were investigated¹⁴.

We investigated the structural brain connectivity networks by using graph theoretical analysis¹⁵. First, we estimated the fiber orientation distributions with constrained spherical deconvolution (CSD)¹⁶ by using up to fourth order spherical harmonics and performed probabilistic streamlines tractography in MRtrix^17-18. Cortical parcellation of the T₁-weighted images was performed in FreeSurfer¹⁹ by using the Desikan-Killiany atlas²⁰, combined with the subcortical gray matter structures segmented with FIRST²¹ in FSL²². Then, the two endpoints of each streamline were assigned to the gray matter areas, resulting in structural connectivity networks, presented in Figure 1, in which the nodes represent the gray matter areas and the edges represent the number of streamlines between the nodes.

We compared global microstructural and network properties in ELBW children to FT control subjects and looked for associations between the properties with the following ante- and neonatal factors: gestational age, birthweight, antenatal corticosteroid administration, twin pregnancy, intrauterine growth restriction, vaginal delivery, respiratory distress syndrome, neonatal sepsis, retinopathy of prematurity stage III or higher, and need of oxygen at the age corresponding to 36 gestational weeks.

Results

We found significantly decreased FA and CP in ELBW children compared to FT children, as shown in Table 1. Furthermore, the ELBW children whose mothers did not receive antenatal corticosteroid had decreased FA compared to those treated, as shown in Figure 2. In addition, the children with retinopathy of prematurity stage III or higher had increased MD and AD compared to those with no or lesser retinopathy degree, as shown in Figure 3. However, there were no significant correlations between gestational age or birth weight and the microstructural measures in ELBW children. No associations were found with the other investigated factors.

We found increased normalized clustering coefficient and local efficiency in ELBW children compared to FT children, as shown in Table 2. No correlations with the gestational age or birth weight were found, but vaginal delivery was associated with increased local efficiency (p=0.035).

Discussion

We found global microstructural and connectivity abnormalities in the preterm group compared to FT children. A decrease in FA in ELBW children suggests decreased white matter integrity or myelination²³, while a decrease in CP suggests a decrease in the complexity of the white matter organization¹⁴. Increased clustering coefficient and local efficiency were found in ELBW children, which indicates increased segregation of the structural brain connectivity networks¹⁵. In addition, the microstructural changes were associated with the administration of antenatal corticosteroids and with the retinopathy of prematurity.

The limitations of the study concern the suboptimal DW data acquired with a low b-value and number of gradient orientations²⁴, and too small field of view missing the superior part of the brain.

Conclusion

Acknowledgements

References