A study using structural magnetic resonance imaging (MRI) data found that brain regions most affected by anorexia nervosa (AN) were the most energetically demanding. These findings were published in the Journal of the American Academy of Child & Adolescent Psychiatry.
Data for this study were sourced from previously published data from 395 girls or women combined with a replication sample of 248 girls or women recruited for this study. The 2 cohorts comprised 120 and 75 girls or women with acutely underweight AN (acAN), 68 and 34 weight-recovered former AN (recAN), and 207 and 139 healthy controls (HC), respectively.
All participants underwent an MRI following an overnight fast, and a virtual histology approach was used to investigate the inter-regional profiles of cortical thickness, cell-specific gene expression, and functional connectomics. The acAN cohort underwent a scan at time point 1 (TP1) within 96 hours of beginning nutritional rehabilitation and at time point 2 (TP2) after restoring >12% of their BMI.
Among the replication sample, the acAN, recAN, and HC cohorts were aged mean 16.9 (SD, 3.7), 21.7 (SD, 3.7), and 18.5 (SD, 4.3) years; minimum lifetime BMI was 14.4 (SD, 1.4), 14.8 (SD, 1.2), and 19.5 (SD, 1.7) kg/m2; and current BMI was 14.8 (SD, 1.4), 20.7 (SD, 2.0), and 20.8 (SD, 2.2) kg/m2, respectively.
At TP1, the acAN cohort had a widespread reduction of cortical thickness compared with HC. Longitudinal increase in cortical thickness related with partial weight restoration at TP2. No significant differences between cohorts were observed at TP2.
In the combined sample using all data, mean cortical thickness in acAN at TP1 was 6.4% lower than HC. At follow-up, the acAN cohort had a 6.4% increase in cortical thickness.
Regions with larger reductions among the acAN cohort at TP1 correlated with larger increases after weight restoration (r, -0.86; P <.00001).
Compared with cortical thickness, the reduction in cortical surface area was considerably smaller (1.6% vs 6.4%).
The virtual histology analysis found that regions of the brain with greater cortical thickness reductions expressed pyramidal cells and oligodendrocytes, which may indicate these cell types are more vulnerable to AN.
In general, the investigators observed that the most greatly affected regions were characterized as hubs due to correlations with functional (r, 0.44; P <.001) and structural (r, 0.22; P <.001) connectivity.
The findings of this study may not be generalizable among older patients with chronic AN.
The study authors concluded, “We confirm sizable and widespread dynamic structural changes in brain morphology in young patients with AN including a rapid normalization already following partial weight restoration in the largest samples to date. The effects are larger than those observed in schizophrenia and not limited to cortical thickness, but also extend to surface area. The virtual histology approach revealed an underlying pattern of affected cell types that is distinct from a number of psychiatric disorders, but resembles the process of normative maturation.”
Disclosure: An author declared affiliations with industry. Please refer to the original article for a full list of disclosures.
Bahnsen K, Bernardoni F, King JA, Patel Y, Paus T, Ehrlich S. Dynamic structural brain changes in anorexia nervosa: a replication study, mega-analysis, and virtual histology approach. J Am Acad Child Adolesc Psychiatry. Published online April 3, 2022. doi:10.1016/j.jaac.2022.03.026