Alcohol use disorder (AUD) and bulimia nervosa (BN) are two psychiatric disorders that are characterized by binge behavior where large quantities of alcohol (i.e., binge drinking [BD]) or food (i.e., binge eating [BE]) are consumed within a short period of time. Furthermore, both disorders are highly prevalent, impactful, and challenging to treat. Therefore, new and improved treatments are needed for both disorders, but in order to develop them, a better understanding of what triggers binge behavior is required.
It is thought that stress and negative affect (NA) are important triggers for BD and BE. Indeed, studies in a laboratory context have shown that inducing stress and NA can lead to increased alcohol and food consumption in patients with AUD or an eating disorder respectively. Furthermore, studies conducted in daily life report that NA increases in the hours before a BE episode and that NA is higher before a BE episode than before a regular meal. However, these findings raise the question how stress and NA cause patients to experience a higher desire to binge drink or bing eat, as well as lose control, making them more likely to display binge behavior. Previous authors have hypothesized that there are several important factors such as craving, negative urgency (i.e., the tendency to act rashly under elevated stress or NA), and disturbances in reward processing including delay discounting (i.e., preferring more immediate rewards). Therefore, this thesis explores the role of these factors in how stress and NA lead to binge behavior.
First, an experience sampling method (ESM) study was performed where 76 controls, 53 patients with AUD, 51 patients with BN, and 19 patients with AUD and BN reported on their mood, behavior, and context in daily life. When it comes to BE, we found that NA was related to subsequent BE in patients with BN through rash action and craving, highlighting the importance of negative urgency and craving in the relation between NA and BE. Contrastingly, we also observed that NA was associated with subsequent not eating, indicating that NA can have competing effects on eating behaviors in patients with BN. When it comes to alcohol use, we saw that NA was non-linearly related to craving, alcohol use, and BD in patients with AUD, emphasizing that both lower and higher levels of NA can lead to alcohol consumption in patients. However, these findings raise the question whether factors such as NA, rash action, and craving can actually predict binge behavior in daily life. Therefore, we used machine learning to build person-specific and pooled prediction models for BE, alcohol use, and BD in patients with AUD and/or BN. We found that pooled models performed better at predicting BE, alcohol use, and BD, but that predictors from person-specific models might be more useful clinically. Importantly, craving and time of day were the most important predictors for all behaviors, while there were differences in how affect and social context were related to BE, alcohol use, and BD.
Second, the role of the neurobiological reward system in BE was explored by conducting a systematic review of previously published studies. We found that individuals who binge eat display a lower striatal dopamine release in rest, a change in the volume of the striatum, frontal cortex and insula as well as a lower fronto-striatal connectivity. Furthermore, there was a higher activity of the brain reward system when anticipating or receiving food, and individuals who binge eat relied more on previous experiences when making decisions and displayed more habitual behavior. These results show that individuals who binge eat display structural and functional changes in the neurobiological reward system, which could play a vital role in the onset and maintenance of BE episodes.
Third, an MRI study was performed with 50 controls, 27 patients with AUD, and 25 patients with BN. In this study, the effect of stress on alcohol and food delay discounting was investigated to see whether stress makes patients with AUD or BN prefer more immediately available alcohol or food respectively. We found that stress increased delay discounting of alcohol in patients with AUD, but not in controls, and that this was related to a lower activity of the right supplementary area. In contrast, we observed that stress increased delay discounting of food in controls, but not in patients with BN, and that this was related to a lower activity of the anterior cingulate cortex. These results suggest that acute stress could indeed make patients with AUD prefer more immediately available alcohol, while this relation might not be as straightforward in patients with BN.
Fourth, a PET/MR study was conducted in 12 controls investigating the relation between stress-induced dopamine release in the ventromedial prefrontal cortex (vmPFC), fronto-striatal functional connectivity and negative urgency in daily life. Here, stress decreased functional connectivity between the vmPFC and dorsal striatum, but increased connectivity with the contralateral ventral stiatum. However, individuals with a higher connectivity between the vmPFC and dorsal striatum showed more negative urgency in daily life. Furthermore, individuals with a higher stress-induced DA release had a higher change in fronto-striatal connectivity and displayed more daily life negative urgency. These results highlight how stress can impact dopamine signaling and fronto-striatal connectivity and how this can lead to rash action.
Taken together, this thesis shows that there is a complex relation between stress and NA on the one hand and binge behavior on the other hand. Specifically, the results indicate that stress and NA might have competing effects on eating behaviors in patients with BN, while both lower and higher levels of stress and NA might be related to alcohol use in patients with AUD. Furthermore, the findings of this thesis highlight the importance of craving, negative urgency, reward processing, and delay discounting in this relation, and demonstrate the key role of dopamine transmission and fronto-striatal connectivity. Future studies should explore how stress and NA lead to binge behavior in more depth with more diverse samples, longitudinal designs, and by combining multiple modalities.