The Effects of Obstructive Sleep Apnea Syndrome on the Dentate Gyrus and Learning and Memory in Children

Abstract

Obstructive sleep apnea syndrome (OSAS) is associated with intermittent hypoxia and sleep loss. In children, impairments of cognitive function are important manifestations, but the underlying pathology is unknown. We hypothesized that OSAS would affect the dentate gyrus, a hippocampal subdivision essential to neurogenesis and cognition, and that this impact would further affect cognitive function in children. In children with OSAS (n = 11) and control subjects (n = 12; age and sex matched), we performed diffusion tensor imaging and structural MRI, polysomnography, and neuropsychological assessments. We found that OSAS was associated with decreased mean diffusivity of the left dentate gyrus (p = 0.002; false discovery rate corrected; adjusting for sex, age, and body mass index), showing a large effect size (partial η2 = 0.491), but not with any other structural measures across the brain. Decreased dentate gyrus mean diffusivity correlated with a higher apnea hypopnea index (Spearman's r = -0.50, p = 0.008) and a greater arousal index (r = -0.44, p = 0.017). OSAS did not significantly affect neuropsychological measures (p values >0.5); however, a lower verbal learning score correlated with lower dentate gyrus mean diffusivity (r = 0.54, p = 0.004). Path analysis demonstrated that dentate gyrus mean diffusivity mediates the impact of OSAS on verbal learning capacity. Finally, the diagnostic accuracy of a regression model based on dentate gyrus mean diffusivity reached 85.8% (cross validated). This study demonstrates a likely pathway of effects of OSAS on neurocognitive function in children, as well as potential utility of the dentate gyrus mean diffusivity as an early marker of brain pathology in children with OSAS.SIGNIFICANCE STATEMENT In this study we investigate the relationships between dentate gyrus structure, hippocampus-dependent cognition, and obstructive sleep apnea syndrome (OSAS). We demonstrate lower mean diffusivity of the dentate gyrus in children with OSAS, which correlates with a lower verbal learning and memory score. This study provides new evidence of disrupted microstructure of the dentate gyrus in children with OSAS that may help explain some of the neurocognitive deficits described in these children.

Keywords: diffusion anisotropy; hippocampus; magnetic resonance imaging.

Copyright © 2017 the authors 0270-6474/17/374280-09$15.00/0.

Figures

Figure 1.

Significant effects of OSAS on dentate gyrus microstructure and cognition in children. A, Violin plot of MD of the left dentate gyrus in OSAS subjects and controls (box represents the interquartile range; white dot, median; thin line, 95% confidence interval; density plot width, frequency). P value indicates significance of the group difference; general linear model; False Discovery Rate controlled. B, Representative segmentation of hippocampal subdivisions based on T1-weighted images using an automated segmentation method. C, An exploratory analysis showing effects (effect sizes, partial eta squared) of OSAS on hippocampal mean diffusivity and volumes across subdivisions. D, For all subjects (OSAS subjects and controls), a decrease in dentate gyrus mean diffusivity correlates with an increase in apnea hypopnea index, an increase in arousal index, and a decrease in verbal learning (International Shopping List Test). DG, Dentate gyrus; HT, hippocampal tail; Fim, fimbria; HATA, hippocampus, amygdala transition area; Sb, subiculum; Pre-s, presubiculum; Par-s, parasubiculum.

Figure 2.

Path analysis shows that decreased left dentate gyrus mean diffusivity mediates the effect of OSAS on learning and memory capacity. Parameter estimates are shown with p values (in brackets). For OSAS subjects/controls, diagnosis was used; for verbal learning and memory, the ISLT was used. The red dashed arrow denotes the ACME. Covariates in the mediation model are not shown (age, sex, and IQ).

Figure 3.

Voxelwise whole-brain analyses on gray and white matter. A, No significant group differences were found in fractional anisotropy or mean diffusivity of the white matter (track-based spatial statistics; green represents a group-mean white matter “skeleton” map) or in gray matter intensity (voxel-based morphometry) across the whole brain at corrected p < 0.05. B, Voxel-based morphometry analysis shows that a lower dentate gyrus mean diffusivity correlates with a smaller volume in the left hippocampus, amygdala, striatum, and thalamus. The scatter plot is based on relative signal intensity (in arbitrary units) extracted from the ventral hippocampus around the dentate gyrus from VBM analysis. Amy, amygdala; cb, cerebellum; DG, dentate gyrus; GM, gray matter; hipp, hippocampus; i.c., insula cortex; thal, thalamus; WM, white matter.

Figure 4.

Dentate gyrus mean diffusivity predicts diagnosis of OSAS. An ROC plot shows sensitivity and specificity in the prediction of diagnosis using a logistic regression model with dentate gyrus mean diffusivity and BMI. The ROC analysis was cross-validated (leave-one-out cross-validation). AUC, Area under curve.