Brain structure differences in pediatric obesity: cause or consequence?

Obesity in adults and children is associated with alterations in brain structure and function, as assessed by neuroimaging methods. However, the majority of extant literature, although informative, has been limited by reliance on small samples and cross-sectional data. This has made it challenging to establish reliable obesity-associated differences and to distinguish between two possible explanations. One explanation is that observed brain differences precede the development of excess weight and could thereby function as predictors of weight gain and potentially illuminate central mechanisms driving weight gain. An alternative is that observed brain differences occur subsequent to weight gain and instead reflect downstream effects of metabolic or other physiological sequelae of heightened adiposity.

Adise et al. [(1)] address this important issue by leveraging a large set of data from the Adolescent Brain Cognitive Development (ABCD) Study (http://www.abcdstudy.org). They examine longitudinal relationships between body weight and volume of subcortical brain regions implicated in appetite and weight control over a 2-year period spanning pre- and early adolescence. To decrease the likelihood that current adiposity drives observations, they limit their sample to children who had healthy weight at baseline (n = 3614). To determine whether brain or weight differences come first, they test two competing models using linear mixed-effect regression. Importantly, owing to the large sample size, they are able to investigate weight gain with potential clinical relevance because, among the n = 3614 children designated as having healthy weight at baseline, a total of n = 385 (12%) developed overweight or obesity by the 2-year follow-up.

Using sex-stratified analyses that carefully control for potential confounders including maternal education, handedness, and puberty, as well as intracranial volume, Adise et al. find that for girls, but not boys, greater increases in body mass index (BMI) are driven by smaller volumes over time in the bilateral accumbens, amygdala, hippocampus, and thalamus; the right caudate and ventral diencephalon; and the left thalamus (all p < 0.05, false discovery rate corrected). In contrast, they find no evidence in either girls or boys to support the inverse model that increases in BMI over time drive volume changes in regions of interest. These findings generate important insights that advance understanding of central mechanisms involved in weight gain but also inspire questions for future research.

The finding that lower volumes in subcortical regions are associated with later weight gain from pre- to early adolescence in girls suggests that these brain regions may play a functional role in weight gain, perhaps via effects on behavioral pathways such as sensitivity to food reward, food-related impulsivity, and emotion/stress processing, which could indirectly influence eating behaviors. This finding also suggests that markers of development in these regions could perhaps one day be used as indicators of obesity risk to support early intervention and allocation of resources to children at raised susceptibility. However, as alluded to by the authors, because brain and weight changes are only investigated over a short period, further research extending both earlier in development (not possible in ABCD) and later in development (possible as ABCD continues) is necessary to understand whether, for example, smaller region of interest volumes over this period reflect a pattern of attenuated growth (e.g., reduced synaptogenesis) or accelerated reductions in size (e.g., synaptic pruning). As children undergo puberty and adolescence, brain–weight relationships could begin to emerge in boys, trajectories in frontocingulate brain areas mediating cognitive control may become more relevant, and larger rather than smaller accumbens volume may be associated with obesity risk by adulthood [(2)]. Earlier in development, opposite phenomena (e.g., larger accumbens volume) [(3, 4)] may also pertain in association with obesity risk, with trajectories over childhood potentially reflecting a pattern of initial overgrowth and subsequent over-pruning. Datasets spanning wider developmental periods will allow and require testing for nonlinear deviations in brain growth trajectories. Analyses of behavioral variables relevant to body weight in ABCD and other datasets could help establish the functional relevance of the brain volume differences detectable by neuroimaging.

The absence of evidence that increases in BMI over the developmental period studied drive volume changes in subcortical brain regions is encouraging because it suggests negligible negative impacts of heightened weight gain within a population sample. Successful intervention during the prepubertal period might therefore promote normative neurodevelopment. However, measurement of metabolic and inflammatory markers in larger pediatric datasets extending to older ages and including longitudinal data from children who go on to develop metabolic dysregulation will be essential to fully investigate effects of heightened adiposity and its correlates on brain function and structure. Sadler et al. [(5)] includes discussion of relevant extant research and potential mechanistic pathways.

In summary, this valuable contribution by Adise et al. adds to the literature by leveraging a powerful, publicly available dataset of an ongoing cohort study to address critical unanswered questions regarding brain mechanisms in human obesity with potential clinical implications.

Susan Carnell discloses research funding from Eli Lilly, payments from Gettysburg University, and serving as a board member for the Society for the Study of Ingestive Behavior.