Background
Nearly 50 years, overweight/obesity has become a global pandemic, especially in children with a much larger rate of increase [
1]. While the increasing trend of overweight/obesity prevalence has flattened in developed western countries, such increase, however, is still accelerating in most developing Asian countries [
2], with China having the largest number of children and adolescents with excessive body weight [
3]. It is worth noting that children’s overweight/obesity, especially in adolescence, is more likely to persist into adulthood, not only predisposing individuals to a bulk of adverse physical consequences such as metabolic syndrome, cardiovascular disease, type 2 diabetes, and cancers [
1,
4], but also leading to a high risk of lower academic achievement and cognitive impairment [
5].
Executive function (EF), a sub-domain of cognitive capacity, also called cognitive or executive control, refers to a set of top-down neurocognitive mental processes against one’s bottom-up instinct actions (including three core subcomponents: inhibitory control, working memory, and cognitive flexibility), which is necessary to make one’s decisions and engage in one’s purposeful and goal-driven behaviors [
6]. While recent works suggested that overweight/obesity may increase the risk of EF issues [
7,
8], the overweight/obesity-EF relationship in adolescents has not been well studied [
9,
10]. Specifically, some studies reported adolescents with overweight/obesity, or higher body mass index (BMI) value, had poor EF performance [
11,
12], while others showed null results [
13,
14]. For instance, one recent study reported that not overweight/obesity per se exerted negative impact on EF performance [
14]. Importantly, few research has investigated EF impairments across weight spectrum (e.g., underweight, normal, overweight/obesity) [
15] in adolescents, especially those with underweight condition. In the meantime, most studies had a limited sample size (often by case–control study design), only used the BMI indicator (less sensitive to body fat [
16]) to assess overweight/obesity condition, and often neglected the group with underweight state that likely present with a lower EF score [
17]. These limitations likely lead to the above-mentioned equivocal findings, hence hindering a comprehensive understanding of the weight-EF link.
More importantly, one possible reason for the aforementioned conflicting findings is that the weight-EF association is likely contingent on some factors. In other words, weight may relate to EF only among certain adolescents. Negative emotions (e.g., depression) frequently occur in adolescence, a period known for significant changes in the body, the brain and the mind, as well as the social environment around them [
18]. Adolescents with weight concerns may tend to experience negative self-image and more maladaptive behaviors (e.g., unhealthy eating, physical inactivity, and more screen exposure) and, in turn EF problems [
19], alternatively, negative emotions may serve to exacerbate the weight issue [
20], which gave us a hint that negative emotions might moderate the weight-EF relationship. Moreover, one adult research has shown that the EF difference between obesity and normal weight group was more prominent in patients with major depressive disorder compared to healthy controls [
21]. However, much of the current literature only focused on the solitary effect of weight condition on EF in adolescents [
11‐
14], making it difficult to determine the extent to which the weight-EF association may be contingent on negative emotions. Given that a growing number of adolescents suffer from weight problems, it is crucial to identify modifiable targets for the EF improvement.
To address these gaps, the current study aimed to determine whether overall and abdominal weight spectrum (including underweight, normal, and overweight/obesity) significantly associated with EF impairment and to identity whether negative emotions (e.g., depression, anxiety, and stress) moderate the weight-EF association in adolescents. We hypothesized that both overweight/obesity and underweight condition would significantly associate with EF impairment in adolescents, especially abdominal non-normal weight state. We also hypothesized that all three types of negative emotions would significantly moderate the weight-EF association. That is, overweight and underweight condition would significantly associated with EF problem only among those experiencing negative emotions.
Discussion
To our knowledge, this is the first study to examine whether negative emotions moderate the weight-EF link in adolescents, taking into consideration both overall and abdominal weight spectrum. There are several important findings. Firstly, we observed abdominal overweight, rather than overall overweight, significantly associated with EF problem. Furthermore, we identified a statistically significant moderating effect of depression in the relationship of abdominal overweight and EF issue. Moreover, we also found abdominal thinness significantly associated with EF impairment.
We observed a significant association of abdominal, not overall, overweight with EF problems, particularly in the areas of inhibit, emotion control, shift, working memory, and monitor. This finding was consistent with two prior findings indicating that waist circumstance or WHtR was more sensitive than BMI indicator when examining the weight-EF association [
33,
34]. Consistent with our findings, prior studies also suggested overweight had pronounced effect on the deficits of inhibit control, working memory, cognitive flexibility (i.e., shift in our study), as well as high-order EF (e.g., reason, problem-solve, and plan) [
9]. The underlying mechanism of the overweight-EF link may be attributed to chronic low-grade inflammation. Extensive studies have documented that overweight, especially abdominal adiposity (a marker of visceral fat), can induce systemic low-grade inflammation through activating pro-inflammatory processes in adipose tissue by secreting chemokines and in gut microbiota by altering intestinal permeability and releasing endotoxins [
35,
36]. Inflammation can result in decreased brain volume, reduced white matter integrity, atrophied grey matter, and lower regional blood flow, all of which likely damage EF if occurred in EF-related regions, e.g., frontal, temporal and occipital cortices [
7].
Furthermore, we also found depression significantly moderated the link of abdominal overweight and EF impairment in adolescents, especially in the domains of emotional control, working memory, and initiate. Our finding corroborates with one adult research showing that obesity condition interacts with major depressive disorder to influence EF performance [
21]. Research showed that individuals who suffered from overweight were more vulnerable to depression due to body dissatisfaction and peer stigma than those who had normal weight [
37]. Meanwhile, persons with depressed symptoms were more likely to gain weight and had poor metabolic profile given increased unhealthy diet, sleep disorder, and less physical activity [
38]. A recent longitudinal study also demonstrated that higher weight predicted more severe depressive symptoms and worse EF via higher levels of IL-6, and depressive symptoms also predicted IL-6 increase, and all observed associations were unique to depression but not anxious [
39]. Adolescence is widely recognized as a stage of heightened emotional reactions [
18], and depressive status measured by DASS-21 in our study likely predispose these adolescents to clinical disorder if not properly addressed, hence warranting much more attention.
Moreover, we observed a quadratic relation between WHtR and EF profile, similar to one recent finding showing a U-shape relationship of BMI with brain volume in children [
17]. Besides aforementioned abdominal overweight, we also found abdominal thinness significantly associated with EF impairment in adolescents, with, however, other studies reporting null findings, possibly due to the BMI indicator being used [
13]. Because of undernutrition in low income region, or self-pressure to restrict dieting fulfilling the thin beauty ideal [
40] (oftentimes by unhealthy eating [
41]), individuals with abnormal thinness/underweight probably have lower than desired levels of nutritional and biochemical molecule for optimal EF, e.g., protein, micronutrient, IGF-1 and brain derived neurotrophic factor [
42,
43]. Future studies using more sensitive indicators such as body composition and biochemical markers are needed to confirm and extend our finding.
Our study neither found a significant association of overall overweight and EF, nor observed a moderating effect of depression on that association. It is possible that the BMI indicator incorporating both lean and fat mass is an insufficient marker of body fat. Although one adult study found stress significantly moderated the abdominal obesity and EF association [
44], our study didn’t find stress and anxiety as statistically significant moderators. Our post-hoc analysis indicating that anxiety and stress were not associated with abdominal overweight may be one of the reasons. More mechanism researches of the association between different types of negative emotions and abdominal overweight in adolescents should be conducted in the future.
Limitations
The present study had some limitations. Firstly, our sample was from one of the underdeveloped cities in China with a relatively low obesity prevalence. More samples with diverse socioeconomic status should be performed in future. Secondly, due to low rate of obesity and clinical EF impairment, we utilized the overweight and sub-clinical EF cutoffs. Although the effects estimated were at magnitudes that may be considered sub-clinical in adolescents, these findings are important at the population level. Thirdly, adolescent lifestyle behaviors relied on self-report. Future studies should use more objective measurements, e.g., using actiwatch to assess night sleep duration and physical activity, to reduce reporting bias. Fourthly, we did not collect the adolescents’ perspectives on their EF which will influence our interpretation of the findings. Fifthly, that we performed lots of regressions may increase the type I errors, which should be taken into consideration when interpreting the findings. And lastly, given the cross-sectional nature, we cannot determine the causal association, and cannot exclude the effects influenced by other unmeasured or unknown confounders.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.