Overweight is not associated with cortical thickness alterations in children

Abstract

Introduction: Several studies report an association between body mass index (BMI) and cortical thickness in adults. Some studies demonstrate diffuse cortical thinning in obesity, while others report effects in areas that are associated with self-regulation, such as lateral prefrontal cortex.

Methods: This study used multilevel modeling of data from the NIH Pediatric MRI Data Repository, a mixed longitudinal and cross-sectional database, to examine the relationship between cortical thickness and body weight in children. Cortical thickness was computed at 81,942 vertices of 716 MRI scans from 378 children aged between 4 and 18 years. Body mass index Z score for age was computed for each participant. We performed vertex-wise statistical analysis of the relationship between cortical thickness and BMI, accounting for age and gender. In addition, cortical thickness was extracted from regions of interest in prefrontal cortex and insula.

Results: No significant association between cortical thickness and BMI was found, either by statistical parametric mapping or by region of interest analysis. RESULTS remained negative when the analysis was restricted to children aged 12-18.

Conclusions: The correlation between BMI and cortical thickness was not found in this large pediatric sample. The association between BMI and cortical thinning develops after adolescence. This has implications for the nature of the relationship between brain anatomy and weight gain.

Keywords: MRI; adolescence; body mass index; childhood; cortical development; cortical thickness; gray matter; obesity.

Figure 1

(A) Unthresholded correlations between BMI Z-Score for age and cortical thickness across the entire brain, shown here for comparison. Typically T-values of around 4 or greater are needed to reach significance. (B) Positive correlations between BMI Z-Score for age and cortical thickness across the entire brain corrected for multiple comparisons with FDR q = 0.05 (C) Negative correlations between BMI Z-Score for age and cortical corrected for multiple comparisons with FDR q = 0.05.

Figure 2

(A) Unthresholded correlations between age and cortical thickness across the entire brain, shown here for comparison. (B) Positive correlations between age and cortical thickness across the entire brain corrected for multiple comparisons with FDR q = 0.05. (C) Negative correlations between age and cortical thickness across the entire brain corrected for multiple comparisons with FDR q = 0.05.

Figure 3

(A) Unthresholded correlations between gender (male—female) and cortical thickness across the entire brain, shown here for comparison. (B) Positive correlations (male—female) between gender and cortical thickness across the entire brain corrected for multiple comparisons with FDR q = 0.05. (C) Negative correlations (female—male) between gender and cortical thickness across the entire brain corrected for multiple comparisons with FDR q = 0.05.

Figure 4

(A) Unthresholded correlations between BMI Z-Score for age and cortical thickness across the entire brain, shown here for comparison. Typically T-values of around 4 or greater are needed to reach significance. (B) Positive correlations between BMI Z-Score for age and cortical thickness across the entire brain corrected for multiple comparisons with FDR q = 0.3 (C) Negative correlations between BMI Z-Score for age and cortical corrected for multiple comparisons with FDR q = 0.3.