The early postnatal period is characterized by a robust growth of cortical gray matter compared with white matter, as well as by a posterior to anterior regional specificity of cortical gray matter growth. Sexual dimorphisms are generally similar to those found in later stages of brain development, indicating that adult patterns of sexual dimorphism arise before birth and persist throughout postnatal brain development. The patterns of cerebral asymmetry in this period of brain development are different from those observed in older children and adults, suggesting that adult patterns of cerebral asymmetry arise after birth.
Overall brain volume in the neonatal period is ~35% that of normal adults scanned on the same scanner (Mortamet et al., 2005
), indicating that there is enormous overall growth of the brain between birth and adulthood. Early results from follow-up scans at 1 year of age suggest that total brain size increases 100% in the first year of life to ~73% of adult size (our unpublished results).
In the first weeks after birth, there is a robust increase in cortical gray matter volumes, likely the result of exuberant development of new synapses in the cortex, which has been observed in human and nonhuman primates (Huttenlocher, 1979
; Bourgeois and Rakic, 1993
; Bourgeois et al., 1994
; Huttenlocher and Dabholkar, 1997
). Interestingly, there is regional specificity to this gray matter growth, with occipital and parietal regions growing significantly more than prefrontal regions. This pattern suggests that gray matter is maturing faster in the sensory and motor regions than in the prefrontal regions, reflecting the rapid maturation of visual and motor functions relative to the executive functions of the prefrontal cortex in the early postnatal period (Kagen and Herschkowitz, 2005
). This pattern of gray matter development is also consistent with a similar posterior to anterior maturation of cortical white matter tracts, with the occipital poles myelinating before the frontotemporal poles (Sampaio and Truwit, 2001
The regional differences in gray matter growth is consistent with previous studies of synapse development in the human brain, in which synapse numbers peak at 3–4 months in the occipital cortex and at 4–5 years in the prefrontal cortex (Huttenlocher and Dabholkar, 1997
). In nonhuman primates, there does not appear to be regional differences in synapse development (Bourgeois and Rakic, 1993
; Bourgeois et al., 1994
), and whether there are real regional differences in humans has been debated (Rakic et al., 1994
). Although overall volume gray matter changes are an indirect measure of synapse proliferation, our results are supportive of regional differences in synapse development in humans.
Sexual dimorphism is present in the neonatal brain. After birth, males have ~9% larger ICVs than females. This difference in ICV is similar to that observed in adults, 11.9–14.6% (Gur et al., 1999
; Raz et al., 2004
). Infant males had 10% more cortical gray matter than females, which is similar to the 10–11% differences in gray matter observed in older children (Reiss et al., 1996
; Giedd et al., 1999
) and the 7–10% differences noted in adults (Gur et al., 1999
; Allen et al., 2003
). Males also had 6% more cortical white matter, similar to the 7.5% differences in older children (Reiss et al., 1996
) but less than the 17–15% differences found in adults (Gur et al., 1999
; Allen et al., 2003
). Sexual dimorphism in head circumference has been observed as early as the second trimester in human and nonhuman primates on ultrasound (Joffe et al., 2005
). Our study suggests that adult patterns of sexual dimorphisms of overall ICV and cortical gray matter volume are present at birth, whereas adult patterns of cortical white matter and lateral ventricle volumes develop after birth.
The left hemisphere-greater than-right hemisphere asymmetry present after birth is the opposite of the right-greater than-left asymmetry observed in older children and adults (Caviness et al., 1996
; Giedd et al., 1996
; Nopoulos et al., 2000
; Good et al., 2001
; Matsuzawa et al., 2001
; Raz et al., 2004
). The magnitude of cerebral asymmetries in the early postnatal period is also greater than the asymmetries observed in older children and adults, which are modest (Reiss et al., 1996
; Gur et al., 1999
; Matsuzawa et al., 2001
). For example, although we found a 5% left-greater than-right difference in hemispheric gray matter in the neonatal brain, Gur et al. (1999)
found a 0.19% right-greater than-left difference in adults. There was a significant asymmetry in cortical white matter in our cohort, whereas white matter volumes are symmetric in adults (Gur et al., 1999
Adult patterns of overall fronto-occipital asymmetry or torque are not present in the neonate. Although we did find the expected right-greater than-left asymmetry in the occipital region, the expected prefrontal region asymmetries are not present. Prefrontal asymmetry appears to develop after birth, and this lack of asymmetry may reflect a relative “immaturity” of the prefrontal regions compared with occipital regions, consistent with the slower trajectory of prefrontal gray matter growth pattern versus posterior cortical regions. Interestingly, the magnitude of asymmetry in each region was significantly larger in females, in contrast to studies in adults that found no gender difference in the magnitude of torque (Raz et al., 2004
Cerebral asymmetries are present at birth in humans, indicating they arise during prenatal brain development. A previous ultrasound study found that this left-greater than-right hemisphere asymmetry is present as early as 20–22 weeks gestational age (Hering-Hanit et al., 2001
). Left-greater than-right asymmetry of the lateral ventricles has also been observed in the fetus using ultrasound (Achiron et al., 1997
). In a postmortem study of fetal and neonatal tissue, the right frontal region was found to be larger than the left, whereas the left occipital region was larger than the right (Weinberger et al., 1982
). This contrasts with our finding that the left prefrontal region is the same volume or even larger than the right.
Asymmetry patterns present in the neonatal brain are likely attributable to genetic programs that operate during prenatal brain development, because gene expression in the human embryonic cortex is asymmetric as early as 12 weeks (Sun et al., 2005
; Sun and Walsh, 2006
). Adult patterns of asymmetry are also likely the result of genetic programming and environmental influences on postnatal brain development. A recent study of cerebral asymmetry in 14-year-old children who were born very preterm found normal fronto-occipital asymmetry, suggesting that environmental insults to the developing brain at this very early age does not affect adult asymmetry patterns (Lancefield et al., 2006
). Children with post-traumatic stress disorder have a loss of normal frontal lobe asymmetry, suggesting that postnatal stress may influence development of adult patterns of asymmetry (Carrion et al., 2001
Although the automatic segmentation approach used is the best available for reproducible, highly automated segmentation with good validity given the processing needs of a study with a large number of subjects, there are a few limitations to this approach. There appears to be a slight overestimation of the region of early white matter myelination, especially in the region of the lower anterior part of the internal capsule. Binary classification (i.e., tissue label maps) may misclassify tissue in regions that do not exhibit a clear assignment to gray or white matter. This problem will be addressed in future studies by calculating regional tissue volumes through integration of tissue probability values, produced by the currently applied EMS instead of hard labels. Such a procedure might also better cope with partial voluming at the very thin cortical gray matter layer, which in label color maps leaves the impression of overestimation of gray matter.
The current parcellation is based on a Talairach-based box parcellation of the brain volume with separation of subcortical regions, brain stem, and cerebellum. Our results provide information about overall regional growth in the cortex but does not provide detailed information about specific lobes or subregions within the cortex. Growth trajectories of cortical lobes, such as those defined in the semiautomated approach of Nishida et al. (2006)
, may differ from the results presented in this study, but the overall pattern of posterior-greater than-anterior growth should be consistent.
In summary, brain development in the first weeks of life is characterized by rapid cortical gray matter growth that is regionally specific. Overall, sexual dimorphism of brain size at this age is similar to that found in older children and adults, although there are some subtle differences in patterns of cerebral gray and white matter volumes. The pattern of cerebral asymmetry in neonates is opposite to that of older children and adults and suggests that adult patterns of cerebral asymmetry are the result of ongoing neurodevelopmental processes active after birth, driven by genetic programming as well by experience. There is much to learn about human brain development in the first years of life. We are currently following this cohort with longitudinal imaging and developmental assessments at ages 1 and 2 years.