Poster B111, Sunday, March 26, 8:00 – 10:00 am, Pacific Concourse
Subclinical Eating Disorder Traits are Correlated with Cortical Structure in Regions Associated with Food Perception and Food Reward
Emily Richard1, Cynthia Peng1, Esha Mehta1, Caylynn Yao1, Annchen Knodt2, Ahmad Hariri2, Gregory Wallace1; 1The George Washington University, 2Duke University
Although many studies have found atypical brain structure in clinically-diagnosed eating disorders, no studies to date have explored individual differences in brain structure as a function of subclinical eating disorder traits. Therefore, we seek to identify neural endophenotypes in large and relatively unconfounded subclinical samples. In the current study we correlate variation in cortical thickness with scores on two Eating Disorder Inventory-3 (EDI-3) subtests: Bulimia and Drive for Thinness. 456 young adults (313 female:143 male) self-reported drive for thinness traits, and 247 young adults (169 female:78 male) self-reported bulimia traits and provided one anatomic MRI scan. The CIVET brain-imaging pipeline (v2.0) and SurfStat were used to derive vertex-level cortical thickness values and complete analyses. There were significant negative correlations between drive for thinness traits and cortical thickness in the right insula, and between bulimia traits and cortical thickness in the bilateral insula, left posterior parietal, left somatosensory, and right orbitofrontal cortices (FWE-corrected ps<.05). Furthermore, in subjects showing higher EDI-3 traits, there was diminished correlation between these key regions and the rest of cortex (FWE-corrected ps<.05) compared to those with lower EDI-3 traits. Strikingly, self-ratings of EDI-3 traits were correlated with thickness in distinct cortical regions (e.g., insula and orbitofrontal cortex) that are crucial to food perception and food reward. Furthermore, higher levels of these traits negatively modulated anatomical coupling between these regions and other portions of cortex. These findings complement the clinical literature, and provide additional evidence that these neural signatures can serve as informative endophenotypes for future genetic studies.
Topic Area: NEUROANATOMY