To our knowledge, this is the first study of neonatal brain structure in children at high genetic risk for schizophrenia. And while this study has limitations, we unexpectedly found that male high-risk neonates have significantly larger intracranial, CSF, gray matter, and lateral ventricle volumes than male comparison infants, while female high-risk neonates have brain volumes similar to those of female comparison infants. Our findings suggest that prenatal and early neonatal brain development is abnormal in males at genetic risk for schizophrenia and that some brain structure endophenotypes associated with risk for schizophrenia may be present at birth.
Schizophrenia is increasingly considered a disorder of cortical connectivity, evidenced by lower than normal levels of cortical synaptic markers (30
), spine density (31
), and dendritic complexity (32
) in postmortem studies and by low gray matter volume in imaging studies. It is unclear when the abnormalities observed in patients with schizophrenia or their unaffected first-degree relatives arise. In the human cortex, synapse development occurs rapidly in the first years of life; the number of synapses plateaus prior to puberty and then regresses after puberty (33
). This general pattern is evident in the development of cortical gray matter volume, which increases rapidly in the first 2 years of life (26
) and decreases after puberty (36
). Our finding of larger cortical gray matter volumes in male neonates at risk for schizophrenia suggests that abnormalities of gray matter may arise early in brain development. In the first 2 years of life, there is enormous growth of cortical gray matter, which increases 185% from birth (35
). This period of rapid growth of synaptic connections and dendritic complexity would seem a likely period in which gray matter abnormalities might become evident in high-risk children.
There is evidence in our study that male high-risk children have larger brains than comparison subjects. This is a somewhat unexpected finding and must be considered in light of the small number of subjects. However, this pattern of enlargement is somewhat reminiscent of the larger neonatal brain volumes, especially larger gray matter volumes associated with larger lateral ventricle volumes, observed in children with prenatal mild ventriculomegaly (37
). The larger cortical gray matter volumes are also reminiscent of the macrocephaly observed in autism (38
), perhaps implicating similar early developmental trajectories in these disorders, which have some overlapping behavioral features. A recent twin study we conducted indicates that in the neonatal period, the heritability of gray matter volume is 0.56, less than that observed in adults, while the heritability of lateral ventricle volume is 0.71, higher than that observed in adults (39
). The larger gray matter volumes observed in this cohort may reflect genetic and environmental risk factors, while the enlarged ventricles may reflect more genetic risk.
It is tempting to speculate that in some forms of schizophrenia, larger than normal neonatal gray matter is an early indication of abnormal connectivity in the developing cortex, an abnormal connectivity that persists throughout brain development and may ultimately result in low cortical gray matter volumes as developmental trajectories play out over childhood. For example, low gray matter volumes are seen not only in adults with schizophrenia but also in children with early-onset schizophrenia (24
). Of course, much larger groups and longitudinal studies are needed to determine this with any certainty. Adolescence, a period of significant synaptic reorganization and elimination (33
), also represents another period of brain development and a period in which abnormalities in gray matter may likely emerge. Alternatively, gray matter abnormalities could also occur after the onset and during the course of the illness (36
). This question turns on whether gray matter abnormalities occur as a consequence of altered neurodevelopment or pathophysiologic processes inherent in the disorder.
Several of the genes that may increase risk for schizophrenia, including NRG1, erbB4, DISC1, and DTNBP1, have roles in synapse development and plasticity (7
), the effects of which might become apparent in the period of rapid postnatal synapse development. Finally, studies in twins suggest that the heritability of gray matter volume increases with age (39
), additional evidence that genetic effects on gray matter development would become more evident through postnatal brain development.
Diffusion tensor imaging has revealed widespread abnormalities of white matter in people with schizophrenia and those at genetic high risk; these abnormalities include major white matter tracts of the corpus callosum and internal capsule (11
). We detected no group differences in diffusion tensor imaging properties in any of the white matter tracts studied. This suggests that white matter abnormalities may emerge after the neonatal period. It appears that major white matter tracts are established at birth, while myelination and maturation of diffusion properties in these tracts occur rapidly over the first 2 years of life, with a more gradual attainment of adult levels thereafter (28
). The first 2 years of life, with their rapid growth and maturation, may be a period when white matter abnormalities associated with schizophrenia risk emerge.
A previous retrospective prenatal ultrasound study showed that lateral ventricle width in the offspring of mothers with schizophrenia was not greater than normal (42
). We found no differences in prenatal ventricle width, supporting the findings of Clarke et al. (42
), although we did find larger than normal neonatal ventricle volume in males. This suggests that the higher lateral ventricle volume observed in high-risk neonates may arise after the 32-week ultrasound or that the two-dimensional ultrasound is not a sensitive measure of prenatal lateral ventricle volume.
This study has several limitations. While the number of subjects is large in terms of the difficulty in recruiting and imaging this high-risk group, it is small given the heterogeneity of schizophrenia (and risk for schizophrenia) and the relative inconsistencies of imaging findings in older subjects. With our study group size, we had 0.8 power to detect differences in gray matter volume of about 10% (43
). The magnitude of volume differences in adults with schizophrenia is typically less than 10%, and it is even less in studies of unaffected siblings at genetic risk. Most of the mothers of the high-risk subjects in this study took antipsychotics during pregnancy. Previous studies have shown that antipsychotics do not appear to significantly increase the risk of malformations or significant pregnancy outcomes (44
), and studies in nonhuman primates suggest that chronic antipsychotic exposure can decrease cortical gray matter volumes (45
), although the effect of antipsychotics on prenatal cortical development in humans is unknown. Our finding of no large differences in neonatal brain structure or white matter integrity in the female offspring of mothers with schizophrenia suggests that antipsychotics during pregnancy do not have a major impact on brain structure, although gender-specific effects cannot be ruled out. The mothers with schizophrenia were also more likely to have smoked tobacco and to have used illicit substances during pregnancy, and each of these factors could have confounded the results, although the expected direction of effects would be toward smaller overall head and brain sizes (46
). As already noted, we did not find significant differences in volumes between the high-risk neonates exposed to maternal smoking and those who were not exposed. Otherwise, the high-risk and comparison groups were reasonably well matched for pre- and perinatal factors. The absence of detectable abnormalities in the female high-risk neonates in this study does not mean prenatal brain development is normal, as more subtle, undetectable differences in gray and white matter development may be present. Finally, we would expect only about 10% of the high-risk neonates to ultimately develop schizophrenia, although observed abnormalities may represent intermediate phenotypes of risk.
In summary, male neonates at high risk for schizophrenia appear to have larger than normal gray matter, CSF, and lateral ventricle volumes, and this appears to be the first evidence that neonatal brain development is abnormal in individuals at genetic risk for schizophrenia. It is suggested that the period of rapid cortical synapse development and gray matter volume growth in the first 2 years of postnatal life represents an important period in which the abnormalities of “connectivity” thought to underlie risk for schizophrenia may emerge. Neuropsychological studies of children genetically at risk for schizophrenia have revealed a variety of neuropsychological and motor abnormalities that are apparent early in life (47
); the structural abnormalities that underlie them may also arise in the first years of life. We are following this cohort through this period of rapid postnatal brain development to study how these structural brain abnormalities progress after birth.