Age differences in eating traits associated with risk of obesity: The role of interoception

Abstract

In younger adults, the factors driving eating behaviours associated with an increased risk of obesity have been well-researched. However, far less is known about whether the same factors apply in later life. Therefore, the over-arching aim of this thesis was to evaluate whether the same mechanisms driving eating behaviours in younger adults translate to older cohorts. The current thesis explores the extent to which mechanisms such as interoception, may be driving eating behaviours associated with obesity in young and older adults. Specifically, the studies included here identify that age-related alterations in interoceptive processing may be linked with changes in eating behaviour. The current thesis draws from a range of scientific approaches to achieve its aims including neuroimaging, biobehavioural, and statistical approaches. Throughout the thesis, interoception is operationalised using neuroimaging methods, a novel paradigm (developed within an active inference framework), and self-report methods. Chapter 2 presents a systematic review that considered the effects of age on interceptive processes relevant to eating behaviour. Specifically, studies that examined the effects of age on various interoceptive domains such as appetite sensations, cardioception, interoceptive sensibility, and orosensory were examined. However, across all domains, there were various limitations that need to be addressed before recommendations can be made. A roadmap was proposed for future studies, consisting of longitudinal, population-based research, paradigms that can differentiate belief from sensation-driven interoception, multidimensional research designs, and mechanistic links to eating behaviour. Chapter 3 investigated the resting-state functional connectivity of three major brain networks previously linked with hedonic eating and interoceptive processes. Twenty-one younger (aged 19-34 years, BMI range: 18-31) and twenty older (aged 60-73 years, BMI range: 19-32) adults completed the Three Factor Eating Questionnaire and underwent a resting state fMRI scan. The analysis found that older adults reported lower levels of disinhibited eating and had weaker connectivity in the frontoparietal (FPN) and default mode (DMN) networks. Additionally, disinhibited eating was associated with weaker connectivity in the FPN and DMN – effects that were absent in the younger sample. Importantly, these effects could not be explained by differences in habitual diet. These findings provided preliminary evidence that appetite changes in older adults may be associated with differences in the engagement of brain networks underlying executive functioning, attentional control, and interoception. The findings should be replicated in a larger sample. Chapter 4 explored inferential cues and integrative processes involved in eating styles associated with increased risk of obesity, within a predictive coding framework. Fifty-four young adults aged 18 to 32 completed a novel paradigm which assesses the degree to which individuals rely on expectations versus sensations when determining their satiety. Participants scoring high on the restrained eating scale were more prone to depend on prior expectations and acquire an "illusionary" sense of fullness when postprandial sensations were unanticipated (after consumption of a sucralose drink [incongruent condition], but not the glucose drink [congruent condition]). In contrast, high scorers on the disinhibited eating scale showed higher levels of "rebound hunger" and were more sensitive to changes in blood sugar. A high disinhibited eating score was associated with a greater sensitivity to incongruent-unexpected interoceptive states. Regardless of the type of beverage, those scoring high in disinhibited eating were less certain about their ability to predict satiety using visual cues and had fewer specific satiety expectations. The plan was to determine whether older adults differentially relied on expectations versus sensations. It was predicted that older adults would report lower disinhibition and rely to a greater extent on prior expectations compared to younger adults. Unfortunately, due to the COVID 19 pandemic, where older adults were particularly vulnerable to serious disease, we were unable to recruit older adults into this study. Chapter 5 examined whether general (accuracy and attention) and specific (hunger and satiety) self-reported interoception and eating traits mediated the association between age and BMI. A large sample (N= 1006, aged 18-80 years) completed the online survey. Despite being more overweight, older adults reported lower interoceptive attention, hunger drive, emotional overeating, food responsivity, and enjoyment of food. In contrast, compared to younger adults, older adults reported a higher interoceptive accuracy, and a similar responsivity to satiety. Two indirect pathways positively mediated the link between age and BMI: (1) age → interoceptive attention satiety responsivity → emotional eating → BMI and (2) age → interoceptive attention → satiety responsivity → food responsivity → BMI. However, a stronger antagonistic indirect pathway was also present: age → interoceptive attention → hunger drive → emotional eating → BMI. The findings indicated that reduced interoceptive attention in older adults could be harnessed to protect against weight gain by lowering hunger and the propensity towards eating behaviours associated with obesity. Overall, the evidence presented in this thesis indicates that (1) weight gain in older and young adults may be driven by different underlying processes, (2) interoceptive attention may be an important variable to consider when studying the differences in older and younger adult eating behaviour, and (3) eating behaviour research would benefit from moving away from studying undergraduate samples to exploring factors contributing to weight gain in community samples such as older adults over 60 years old.