Food vs money? Effects of hunger on mood and behavioral reactivity to reward in anorexia nervosa

Abstract

Background

Previous studies using neuroimaging and behavioral measures reported altered reward processing in anorexia nervosa (AN). In addition, anhedonia states are frequently reported in AN, potentially due to the physiological stress produced by the permanent starvation. We investigated the effect of fasting and satiety on mood and reaction times to monetary rewards in AN patients and healthy controls.

Methods

Twenty-four participants with acute AN (BMI 14.4 (11.9–15.5) Kg/m²) and 17 age and gender matched healthy, normal weight subjects (HW) (BMI 21.8 (18.9–24.9) Kg/m²) performed a reward task (the wheel of fortune) involving uncertain (50/50 probability of winning high and low rewards), safe and risky (30/70 and 10/90 probabilities) categories in fasted (after an 8-h fasting period) and fed (after intake of a standardized meal) states. Data analysis was done with linear mixed models.

Results

AN reacted slower than HW when maximum uncertainty (50/50) was involved. Positive mood in response to winning was higher when fasting especially for HW, while negative mood in response to not winning was higher in the fed state for both groups. Still, HW were more reactive than AN to not winning a highly predictable monetary reward (10/90 safe).

Conclusion

The data on the reaction times indicate an impaired motor response to uncertainty in AN. Mood reactivity to winning a monetary reward does not seem to be impaired in AN, however, our results suggest that negative mood in response to not winning is less adaptive in AN. Implications to clinical psychotherapy are discussed.

Introduction

Anorexia Nervosa (AN) is an eating disorder, mainly affecting women, with a prevalence of approximately 1% (Mohler-Kuo, Schnyder, Dermota, Wei, & Milos, 2016). AN is the eating disorder with the highest mortality rate (Fichter & Quadflieg, 2016), and its pathophysiology remains mostly unclear. It is important to elucidate the mechanisms involved in the acquisition and maintenance of the disorder, especially those related to what is responsible for triggering disease-maintained behaviors.

Previous studies reported altered reward processing in AN at neural level, showing reduced striatal activation to natural and monetary rewards as well as reduced dopamine (DA) function (Barbato, Fichele, Senatore, Casiello, & Muscettola, 2006; Bergen et al., 2005; Davis & Woodside, 2002; Ehrlich et al., 2014; Piazza et al., 1993; Wagner et al., 2007). As it is well established that DA is involved in the processing of reward, a DA deficiency has been hypothesized in the etiology of AN (Berridge & Robinson, 2016; Martin-Soelch et al., 2011; Schultz, 2000). Furthermore, animal studies showed that DA-deficient rats ingest less food than the least necessary for survival (also one of the characteristic behaviors for AN) (Szczypka et al., 1999). Anhedonia, a lessened ability to experience pleasure from rewards, also greater among patients with AN (Davis & Woodside, 2002), has also been connected to poor DA functioning (Willner, Daquila, Coventry, & Brain, 1995).

Tapper (2005) suggests the importance of controlling hunger for research involving feeding and eating behaviors. With regard to that, a recent study showed that fasting increases the reinforcing potential of food rewards, in particular highly caloric ones in normal weight controls (Goldstone et al., 2009). This effect has not been tested in participants with current AN. However, several studies indicated that food reward processing is different in AN patients. For instance, Stoner, Fedoroff, Andersen, and Rolls (1996) reported that preferences for high-fat food is stronger in controls than in AN, both in the acute phase and after weight gain. Nonetheless, hunger was not controlled in their study. Many studies with AN are performed with women remitted from AN in order to avoid the confounding effects of malnutrition (Wagner et al., 2007; Wierenga et al., 2015), since research with food rewards showed similarities between acute and remitted patients (Stoner et al., 1996). However, that might not be completely true when monetary rewards are used. One recent study, for instance, has shown that responses to monetary reward were normalized after weight gain in patients remitted from the disorder (Steinglass, Decker, Figner, Casey, & Walsh, 2014), therefore, studying patients in the acute phase may be essential for understanding specific responses to non-food rewards in the disorder.

Investigating monetary reward in AN patients is particularly relevant, because, on one hand, several studies indicated changes in the neural processing of monetary reward in AN patients; and, on the other hand, fasting resp. hunger was evidenced to influence responses to monetary reward (Briers, Pandelaere, Dewitte, & Warlop, 2006). For instance, behavioral results of a functional Magnetic Resonance Imaging (fMRI) study showed a diminished sensibility to the feedback associated with monetary reward in recovered AN patients that could be associated with increased cognitive control when dealing with monetary rewards (Wagner et al., 2007). A study using a gambling task revealed that individuals with AN showed poorer decision-making performances to reward compared to controls, but not remitted AN (Tchanturia et al., 2007). More specifically, patients with AN insisted on choosing cards from decks which allowed them to win larger amounts of money (but more often losing them too), and did not shift, different than did controls and remitted patients, to decks that represented safer conditions, lower rewards, but the possibility of actually winning money (Tchanturia et al., 2007). Taken together, the findings suggest an impairment in the processing of reward, which might also lead to altered perceptions of hedonic mood in severe AN. Regarding the effect of hunger on responses to money, one previous study reported that hunger increased the rewarding value of money in students, who were less willing to give it up while fasting (Briers et al., 2006). Yet, a recent study in women remitted from AN showed that hunger does not modulate responses to monetary reward in this group compared to healthy women (Wierenga et al., 2015), but, up to this date, no study has investigated its effect in acutely ill AN patients.

In summary, altered responses to monetary rewards were observed in patients with AN. Considering the occurrence of fasting in AN and its influence on responses to reward, it is relevant to investigate the effect of the feeding state on these responses in participants with AN. Therefore, the aim of this study was to investigate the effect of fasted and fed states on mood and behavioral reactivity to monetary rewards in acutely ill patients with AN. We hypothesized that AN would show more negative affect and less mood reactivity to winning and losing a monetary reward than healthy controls during fasting, and this group difference would be less strong during the fed state. Also, according to the results obtained by Tchanturia and colleagues (2007), we expected patients with AN to show less mood reactivity than controls to losing higher amounts of reward, as well as the group difference to be stronger during fasting.