In this study, we used magnetic resonance imaging (MRI) to identify the different brain phenotypes within apparently healthy children and to evaluate whether these phenotypes had different prenatal characteristics. We included 65 healthy children (mean age, 10 years old) with normal neurological examinations and without structural abnormalities. We performed cluster analyses to identify the different brain phenotypes in the brain MRI images. We performed descriptive analyses, including demographic and perinatal characteristics, to assess the differences between the clusters. We identified two clusters: Cluster 1, or the “small brain phenotype” (n = 44), which was characterized by a global reduction in the brain volumes, with smaller total intracranial volumes (1044.53 ± 68.37 vs. 1200.87 ± 65.92 cm3 (p < 0.001)), total grey-matter volumes (644.65 ± 38.85 vs. 746.79 ± 39.37 cm3 (p < 0.001)), and total white-matter volumes (383.68 ± 40.17 vs. 443.55 ± 36.27 cm3 (p < 0.001)), compared with Cluster 2, or the “normal brain phenotype” (n = 21). Moreover, almost all the brain areas had decreased volumes, except for the ventricles, caudate nuclei, and pallidum areas. The risk of belonging to “the small phenotype” was 82% if the child was preterm, 76% if he/she was born small for his/her gestational age and up to 80% if the mother smoked during the pregnancy. However, preterm birth appears to be the only substantially significant risk factor associated with decreased brain volumes.
Keywords: magnetic resonance imaging; maternal behavior/drug effects; preterm birth; small for gestational age.