Regional brain volume reductions relate to facial dysmorphology and neurocognitive function in fetal alcohol spectrum disorders

Abstract

Individuals with heavy prenatal alcohol exposure can experience significant deficits in cognitive and psychosocial functioning and alterations in brain structure that persist into adulthood. In this report, data from 99 participants collected across three sites (Los Angeles and San Diego, California, and Cape Town, South Africa) were analyzed to examine relationships between brain structure, neurocognitive function, facial morphology, and maternal reports of quantities of alcohol consumption during the first trimester. Across study sites, we found highly significant volume reductions in the FASD group for all of the brain regions evaluated. After correcting for scan location, age, and total brain volume, these differences remained significant in some regions of the basal ganglia and diencephalon. In alcohol-exposed subjects, we found that smaller palpebral fissures were significantly associated with reduced volumes in the ventral diencephalon bilaterally, that greater dysmorphology of the philtrum predicted smaller volumes in basal ganglia and diencephalic structures, and that lower IQ scores were associated with both smaller basal ganglia volumes and greater facial dysmorphology. In subjects from South Africa, we found a significant negative correlation between intracranial volume and total number of drinks per week in the first trimester. These results corroborate previous reports that prenatal alcohol exposure is particularly toxic to basal ganglia and diencephalic structures. We extend previous findings by illustrating relationships between specific measures of facial dysmorphology and the volumes of particular subcortical structures, and for the first time show that continuous measures of maternal alcohol consumption during the first trimester relates to overall brain volume reduction.

Figure 1

Subcortical regions of interest. Inferior view (left panel) and lateral view (right panel) of the brain, showing the seven subcortical regions for which we obtained volume measurements using FreeSurfer's automated brain segmentation software.

Figure 2

Calculation of cortical and subcortical volumes with FreeSurfer The pial surface, or gray matter surface (red outline) and the white matter surface (yellow outline) are shown, overlaid over the T1‐weighted image. The segmented cortical volumes are shown in shaded colors. Volumes of the seven subcortical structures of interest were calculated using FreeSurfer's automatic quantification of subcortical structures. Total gray matter volume was calculated by obtaining the volume between the gray and white matter surfaces. Total white matter volume was calculated by subtracting the subcortical and ventricular volumes from the volume bounded by the white matter surface.

Figure 3

Reduced intracranial volume (ICV) in subjects with FASD Box plot illustrating the distribution of total ICV by group. Each circle represents one subject. Scan locations of subjects are coded with different colors. Across groups, subjects with FASD have significantly (P = 0.001) smaller ICV than controls.

Figure 4

Group differences in regional brain volumes after controlling for scan location, age, and overall intracranial volume Inferior view (left panel) and lateral view (right panel) of the brain, showing the regions in which we observed volume reductions in subjects with FASD compared to controls. After controlling for site, age, and overall intracranial volume (ICV) reduction in the exposed population, areas in gray represent regions where group differences were no longer significant. Areas displayed in color represent regions where volume reductions in the FASD group remained significant (P < 0.05, in the left putamen, left total cortical gray matter, right pallidum, and right total cortical gray matter) or retained a trend‐level of significance (P < 0.1, in the left pallidum, left thalamus, left ventral diencephalon, and right putamen). Residual plots corresponding to each one of these regions are displayed on the side.

Figure 5

Relationships between brain volume reductions and palpebral fissure length (PFL) after controlling for scan location, age, sex, and ICV in subjects with FASD Inferior view of the brain, showing the regions in which smaller PFL was associated with smaller subcortical volumes in the FASD group (n = 52). After controlling for site, age, sex, and ICV in multiple regression analyses, areas in gray represent regions where this relationship was no loner significant. Areas displayed in color represent regions where these results remained significant (P < 0.05, in the left and right ventral diencephalon). Residual plots corresponding to these regions are displayed on the side.

Figure 6

Relationships between brain volume reductions and philtrum lipometer scores after controlling for scan location, age, sex, and ICV in subjects with FASD Superior view of the brain, showing the regions in which more pronounced dysmorphology of the philtrum was associated with smaller subcortical volumes in the FASD group (n = 52). After controlling for site, age, sex, and ICV in multiple regression analyses, areas in gray represent regions where this relationship was no loner significant. Areas displayed in color represent regions where these results remained significant (P < 0.05, in the left pallidum and in the left and right thalamus), or retained a trend‐level of significance (P < 0.1, in the right putamen). Residual plots corresponding to each one of these regions are displayed on the side.

Figure 7

Relationships between brain volume reductions and neurocognitive function after controlling for scan location, age, sex, and ICV in subjects with FASD Superior view of the brain, showing the regions in which lower FSIQ scores were associated with smaller subcortical volumes in the FASD group (n = 55). After controlling for site, age, sex, and ICV in multiple regression analyses, areas in gray represent regions where this relationship was no loner significant. Areas displayed in color represent regions where these results remained significant (P < 0.05, in the left and right putamen). Residual plots corresponding to these regions are displayed on the side.

Figure 8

Correlation between total brain volume and level of prenatal alcohol exposure, in subjects from South Africa Scatter plot illustrating a significant negative correlation (P = 0.014, r = −0.501) between ICV and total number of drinks per week during the first trimester, in exposed subjects from SA. Each circle represents one subject (n = 17).