Emotional and external eating are common eating behaviours in the general population, while not disordered eating per se they have been linked to overeating, obesity and problems engaging in health behaviours. Theories of emotional and external eating have been around for decades however little is known of the factors that contribute to these eating behaviours. Emotional and external eating tend to co-occur, and high correlations between them have been reported. Some theorists have argued that they are not distinct constructs. The current study aimed to provide further understanding on the nature and distinction between emotional and external eating in a non-clinical sample. The associations of impulsivity, alexithymia and negative affect in emotional and external eating were investigated, utilising a cross-sectional design. These variables have been shown to be related to eating behaviour in clinical samples however there has been limited research in non-clinical samples. Emotional eating was positively associated with alexithymia, negative affect and lack of perseverance, the relationship with urgency was less clear. External eating was indirectly associated with depression through the mediating variable urgency. The results indicated that emotional and external eating do show some similarities in the variables associated with them, however, the pattern of associations were different for the two eating behaviours. It appears from this study that the theoretical distinction between emotional eating and external eating is warranted with emotional eating appearing to be directly associated with problems with affect regulation whereas external eating is indirectly associated with negative affect.

Attentional bias (AB) refers to the tendency to selectively attend to (orientation towards) and/or hold attention on (slowed disengagment from) disorder-relevant stimuli. Females with eating-related concerns are thought to preferentially process threatening stimuli, which in turn is thought to maintain and exacerbate eating concerns. The aim of the present thesis was to explore AB for threatening stimuli in females characterised by restrained, external or emotional eating, and those with high levels of (non-clinical) eating psychopathology. This was carried out with the intention of identifying cognitive processes that contribute to eating behaviours in females, in order to assess the relevance of an attention training (AT) programme for reducing such biases. A pilot study assessed orientation/slowed disengagement, for mood and food words amongst females with high/low levels of restraint. Forty females completed a modified Stroop task with three conditions. Food and mood conditions included sequences of five words ( target food/mood followed by four neutral). The neutral condition consisted of all neutral words. Performance did not significantly differ according to high/low restraint groups. All participants took longest to colour-name word position 2 (demonstrating slowed disengagement lasting one consecutive trial). However, this pattern was also found in the neutral condition. Methodological limitations were then addressed in study one. High/low restrained eaters (n=48) completed a modified Stroop where targets (food, interpersonal threat, animal) were presented prior to four neutral words. Participants were slow to disengage from targets (slowest for word position 2) in all conditions. Patterns of responding indicated that restrained eaters might take longer to disengage (i.e. the carry-over effect from the food word seemed to last longer than one trial). However, more neutral words in the sequence were needed to assess this. As slowed disengagement from animals also arose, a categorical effect may have occurred. Study two explored attention processing of food using modified Stroop and dot probe tasks. In the Stroop task targets (food, interpersonal threat, household objects) were presented prior to six matched neutral words. This task revealed no evidence of AB. No significant pattern of differences between restrained (n=29)/unrestrained eaters (n=31) emerged; however, binge eating scores were significantly negatively correlated with response times. A dot probe task with food/neutral picture pairs also revealed no evidence of AB. Both restrained/unrestrained eaters had negative mean interference scores indicating avoidance of food. None of the following eating behaviours significantly correlated with AB: restraint, disinhibition, external eating, emotional eating and non-clinical eating psychopathology. Study three employed a further modified dot probe task based on image ratings. There was no evidence of AB, and no significant relation between task performance and restrained, emotional or external eating. 2000ms bias scores (assessing disengagement) were significantly negatively correlated with eating psychopathology and age, suggesting that those with high levels of non-clinical eating psychopathology attentionally avoid food stimuli and that younger females are slower to disengage attention from food (although found within a limited age range). Study four employed further modified Stroop and dot probe tasks, and assessed whether AB mediates the negative mood-eating relationship. Participants were allocated to negative or neutral mood conditions. No evidence of AB was found with the dot probe, but greater levels of emotional eating were associated with slower responding. In the Stroop task, all participants displayed an orientation bias towards food. Emotional eating and drive for thinness (DFT) scores were significantly positively correlated with food word colour-naming times but only amongst participants in a negative mood. However, those with high levels of external eating showed greater AB towards food when in a neutral mood. Highly emotional eaters in a negative mood showed a greater desire to eat than those in a neutral mood but did not increase in food intake. Furthermore, those with a high DFT (in a negative mood) showed no evidence of increased desire to eat or food intake. AB was not significantly related to subjective appetite or food intake. Therefore, AB does not seem to mediate the negative-mood eating relationship. The present thesis provides important suggestions for modifications of Stroop and dot probe tasks targeting orientation and disengagement. A modified Stroop has been more sensitive at detecting food AB than the dot probe. Implications of biased attention processing are discussed in relation to the development of harmful eating behaviours, and the present findings have important implications for developing programmes to prevent eating disorders amongst at-risk females (e.g. through AT or training at-risk females how to effectively cope with negative mood).

Approximately one-third of U.S. adults are at increased risk for life-threatening diseases such as atherosclerosis and type 2 diabetes mellitus. Such individuals are considered healthy without any diagnosed cardiometabolic conditions but may have a constellation of cardiometabolic complications that include obesity, glucose intolerance, hyperinsulinemia, dyslipidemia, hypertension, insulin resistance, and hypertriglyceridemia. When most of these preclinical conditions comorbidly occur, the condition has been referred to as metabolic syndrome (MetS). MetS is considered to reflect one or more early pathophysiological processes in cardiometabolic disease; however, the extent to which these complications and their underlying pathophysiology interact with behavioral factors such as stress, diet, and physical activity have not been clearly established. For example, diet consisting of high total caloric intake and high fat composition is posited to contribute to obesity and other cardiometabolic risk factors, but research is inconsistent regarding the effect of psychological distress (i.e., anxiety, stress, depression, anger) on dietary intake and whether dietary intake mediates a relationship between distress and preclinical cardiometabolic disease risk. One factor that has been suggested to play a role in the distress – dietary intake relationship is eating style. Research on eating styles has identified four main types that may be related to distress and dietary intake: restrained, disinhibited, emotional, and external eating. Restrained eaters consciously restrict food intake to control body weight and body shape. Disinhibited eating refers to overeating that occurs following failure of restraint. Emotional eaters consume foods to reduce and alleviate negative emotions, such as anxiety. External eating occurs in response to immediate food-related external stimuli, regardless of internal physiological cues of hunger. Current evidence suggests each of these eating styles moderates the relationship between distress and dietary intake. There is also some research to suggest a relationship between eating styles and weight gain, body mass index (BMI), and development of obesity. However, no study has examined the interrelationships among psychological distress, eating style, and central obesity, and whether these relationships differ according to gender. Moreover, the extent to which distress and eating style may be associated with cardiometabolic risk beyond obesity is unknown. Thus, the main aim of the present study was to test a model of mediation and moderation to evaluate how psychological distress, eating styles, dietary intake, and gender are associated with measures of cardiometabolic risk in healthy individuals (Figure 1). Four hundred sixty-four participants contributed data from two different studies: Obesity, Metabolic Syndrome, and Meal-Related Glycemia (SUGAR) and Markers Assessing Risk for Cardiovascular Health (MARCH). All participants were aged 18-55 years, had no major systemic disease, were not using medications having a cardiovascular, carbohydrate, endocrine, or psychiatric effect, and had no history of substance or alcohol abuse or dependence. The study employed a structural equation modeling (SEM) approach to assess the following aims: 1) to develop composite, latent factors to reflect psychological distress, eating style, and dietary intake using confirmatory factor analysis (CFA) and to develop a hybrid model of cardiometabolic risk; and 2) to simultaneously test the interrelationships among factors in a comprehensive model so that the strength of direct and indirect effects can be evaluated while statistically controlling for the other factors and covariates in the model. Latent factor models of psychological distress and eating style fit the data and were statistically acceptable, and a hybrid model of cardiometabolic risk fit the data and its CFA components were acceptable. A latent factor model of dietary intake would have likely fit the data and been statistically acceptable given the high intercorrelations among dietary variables, but no such factor was created because dietary variables failed to confirm the hypothesized associations with other model components (e.g., waist girth, eating styles); thus, these measures were excluded from further SEM analyses. Final model results showed that psychological distress was positively related to restrained, emotional, and external eating styles, but only restrained eating was directly associated with greater waist girth. Distress was not directly related to cardiometabolic risk, but an indirect effect was found in which higher levels of distress led to greater waist girth via higher levels of restrained eating. Waist girth, in turn, served as a significant mediator between restrained eating and worse insulin sensitivity, higher blood pressure, diminished glucose tolerance, and greater dyslipidemia. These effects were significant when controlling for age, gender, education, and physical activity, and when analyzed in a comprehensive SEM model simultaneously including distress, eating style, and cardiometabolic risk variables. Of note, results suggest the possibility for a reversed effect such that waist girth leads to restrained eating. Findings also suggest that emotional eating may lead to distress. In contrast, the relationship between distress and the other two eating styles, restrained and external eating, appeared unidirectional such that distress leads to restrained and external eating but not the reverse. Future studies using longitudinal data are needed to better understand these relationships in regards to causality. Data from the MARCH subsample was excluded from the above final modeling analyses because eating style data were only available for the SUGAR subsample. Thus, the role of gender in how distress, eating styles, and cardiometabolic risk are interrelated could not be examined due to the small number of women in the SUGAR study (n = 38). It remains unknown whether the significant effect of distress on each of the eating styles found in the current study was driven primarily by men, women, or both equally. Similarly, the sample size would not permit the evaluation of whether gender moderated the effect of restrained eating on central obesity. Given that women in the current study reported more restrained, emotional, and external eating than men, future studies with larger samples should follow-up by assessing for potential moderating effects of gender. The present findings suggest that decreasing restrained eating style may lead to less central fat deposition and hence reduced cardiometabolic risk. Such “non-diet” interventions show potential for improved cardiometabolic health, but more research is needed. Particularly needed are studies examining prevention and intervention outcomes based on type of restrained eating – flexible versus rigid – to better understand how these different subtypes operate and how they can be altered effectively to improve health.

This study examined the effect of an 8-week, progressive exercise intervention on neural responses, specifically N2 amplitude as a measure of inhibitory control, to pictures of food. Healthy women ages 18-44 years were randomized to a morning (AM) exercise group or evening (PM) exercise group. The AM group did moderate-to-vigorous intensity exercise on 4 days per week between 6:30 and 9:30 a.m. while the PM group had the identical volume of exercise between 6:30 and 9:30 p.m. Neural responses, eating behaviors, cardiovascular fitness outcomes, and body weight/composition were measured at baseline and after the 8-week intervention. The N2 amplitude in response to pictures of high- and low-calorie foods was assessed using electroencephalography during a go/no-go task. Dietary restraint, emotional eating, and external eating were assessed using the Dutch Eating Behavior Questionnaire. VO2peak, HRmax, and time to completion were measured during a maximal treadmill test. Body weight was measured on a digital scale, and body composition was measured using dual-energy x-ray absorptiometry. There was not a significant task (go, no-go) × group (AM, PM) × period (baseline, 8 weeks) interaction (F = 0.18; p = 0.677), but there was a main effect of exercise over 8 weeks (F = 6.26; p = 0.017) with increased N2 amplitude following the intervention. There was not a significant interaction as a function of picture type (high-calorie, low-calorie), task, group, and period (F = 0.52; p = 0.478). Changes in body weight and neural outcomes were not significantly associated with changes in eating behaviors for either group (ps < 0.05). There was a significant group × period interaction for body weight (F = 4.90; p = 0.032). Body weight increased by 0.79 ± 1.16 kg in the AM group and decreased by 0.21 ± 1.46 kg in the PM group (effect size = 0.77; CI = 0.15-1.35). There was not a significant group × period interaction for body fat percentage, total body fat or fat-free mass (ps < 0.05). When examining the main effect of exercise on cardiovascular fitness outcomes, VO2peak was not different (F = 1.80; p = 0.187), time-to-completion on treadmill increased (F = 6.51; p = 0.014), and HRmax during the treadmill test was significantly lower (F = 5.49; p = 0.025). This study suggests that 8 weeks of exercise training may increase the inhibitory response to pictures of both high- and low-calorie foods. However, time of day of exercise did not influence this response. Eight weeks of exercise training did not change self-reported dietary restraint, external eating, or emotional eating, and there was no correlation between these eating behaviors and inhibitory control. However, evening exercise was more beneficial for body weight than morning exercise. Given the novelty of this study and its results, additional studies on the influence of time of day of exercise on weight management are needed.

Indiana University-Purdue University Indianapolis (IUPUI) / Obesity is a substantial threat to the health of over a third of adults in the United States. Some evidence suggests that food attentional bias, or the tendency to automatically direct attention toward food-related stimuli in the environment, may contribute to the development of obesity in susceptible individuals. This study hypothesized that (1) food attentional bias would be positively associated with adiposity, (2) food attentional bias would be positively associated with energy intake and external eating, and (3) energy intake and external eating would partially mediate the association between food attentional bias and adiposity. Data were collected from a sample of 120 undergraduate students. Three measures of food attentional bias were obtained: reaction time bias obtained from a visual dot-probe task and direction bias and duration bias obtained from eye tracking. Adiposity indices of body mass index (kg/m2) and body fat percent were measured using standard medical devices. Data were obtained for two mediators: 1) energy intake was assessed by web-based automated 24-hour dietary recall and 2) external eating was assessed using the External Eating Subscale of the Dutch Eating Behavior Questionnaire. Separate linear regression models examining the association between each measure of food attentional bias with each measure of adiposity (adjusted for age, sex, race/ethnicity, and subjective hunger) indicated no associations. Similarly, linear regression analyses revealed no associations between measures of food attentional bias and energy intake or external eating. Models testing for statistical mediation demonstrated that energy intake and external eating were not significant mediators. However, mediation analyses demonstrated a significant overall effect and direct effect between direction bias and BMI in a reduced sample used to test for energy intake as a mediator, suggesting the presence of an association which may not have been detected in the larger sample due to methodological issues, measurement error, or type I error. Despite the overall null results, these findings, in conjunction with previous studies on food attentional biases and adiposity, highlight the need for future investigations examining prospective associations between food attentional bias and adiposity.