Eat your troubles away: Electrocortical and experiential correlates of food image processing are related to emotional eating style and emotional state

Highlights

• The interaction of emotions and eating is relevant for overeating and obesity.

• We studied subjective and neural responses to food images in neutral/negative emotion state.

• Craving ratings increased in negative mood in high emotional eaters (HEE).

• HEE showed enhanced occipital ERPs regardless of emotional state.

• Attending to food under negative mood may serve as emotion regulation strategy in HEE.

Abstract

Emotional eating, a trait-like style of food intake in response to negative emotion states, represents an important aspect of overeating and eating related psychopathology. The mechanisms of emotional eating both on experiential and neuronal levels are not well delineated. We recorded event related potentials (ERPs) while individuals with high or low emotional eating style (HEE, n = 25; LEE, n = 20) viewed and rated pictures of high-caloric food during neutral state vs. negative idiosyncratic emotion induction. Craving ratings increased in HEE and decreased in LEE during negative relative to neutral states. ERPs to food pictures showed an enhanced late positive potential (LPP) over parieto-occipital regions for HEE compared to LEE. Emotional state modulated food picture evoked ERPs over right frontal regions in HEE only. This suggests that appetitive food processing is susceptible to both concurrent emotion and habitual eating style which is of relevance for overeating in healthy and abnormal eating.

Introduction

In today's industrialized societies, energy dense food is available almost any time and anywhere. It also seems that hedonic factors have become an important, if not predominant factor in food intake: stress, negative moods, and emotions can determine eating even in the absence of homeostatic hunger and energy need. Whereas some individuals decrease their food intake when stressed, others increase it (e.g., Wardle, Chida, Gibson, Whitaker, & Steptoe, 2011), raising the question about the individual difference variables that put someone at risk for weight gain.

Since stress is a very broad and all-encompassing concept, the present study focuses on the modulating effect of negative emotions and moods on appetitive responses. The original concept of emotional eating derives from early obesity theories (Bruch, 1969, Bruch, 1973) but has since been applied to a wide range of normal and abnormal eating. Emotional eating can be seen as learned emotion regulation strategy through which negative moods increase the motivation to eat, and actual eating reduced these negative moods. In the terms of classical learning theory, the first process would be due to classical conditioning (increased craving in negative mood) and the second one would be due to operant conditioning (eating reinforced by reduction in negative moods, e.g., Booth, 1994). A concurring account explains emotional eating via the disinhibiting effect of emotions on chronic dietary restraint (Ruderman, 1985) since emotions deplete self-control resources that are then missing in the control of eating (Baumeister, Zell, & Tice, 2007).

Since its original postulation, the concept of emotional eating has received considerable empirical support from behavioral, ambulatory, and psychometric studies (Macht and Simons, 2011, Munsch et al., 2012, van Strien et al., 2012). For example, van Strien et al. (2012) used the emotionality subscale of the Dutch Eating Behavior Questionnaire, DEBQ (Van Strien, 1986) to classify participants into high emotional eaters (HEE) and low emotional eaters (LEE). Negative mood was induced by means of a sad movie (Study 1) and social stressor task (Study 2). As expected, HEE participants tended to increase food intake whereas LEE participants tended to decrease food intake under the negative mood conditions in both studies (similar results reported by Oliver, Wardle, & Gibson, 2000).

However, the effects obtained in that study were relatively subtle, even in participants with extreme scores on the DEBQ (top and bottom 20% in large screening samples). This might be due to the strong cognitive control many individuals exert over their eating behavior, particularly in a laboratory context. More covert (and less regulated) aspects of food cue processing might be uncovered by assessing central nervous system activity. Using functional magnetic resonance imaging, Bohon, Stice, and Spoor (2009) showed that emotional eaters responded with stronger responses in the anterior cingulate cortex to milkshake anticipation and receipt under negative mood whereas non-emotional eaters showed a decrease in reward related areas. Furthermore, in bulimia nervosa patients, negative mood was correlated with regions implicated in reward processing (Bohon & Stice, 2012). To our knowledge, event related potentials (ERPs) have not yet been used to examine the interplay of negative emotions and appetitive food processing. ERPs are particularly promising in this context since they are not only modulated by emotional stimuli but also show sensitivity for food density (Toepel, Knebel, Hudry, le Coutre, & Murray, 2009), food deprivation (Stockburger, Weike, Hamm, & Schupp, 2008; LPP; Stockburger, Schmalzle, Flaisch, Bublatzky, & Schupp, 2009b) and disordered eating (Blechert, Feige, Joos, Zeeck, & Tuschen-Caffier, 2011; Nikendei et al., 2012). Furthermore, ERPs are associated with distinct cognitive processing stages that can speak to underlying mechanisms of emotion-motivation interactions. Thus, the present study examined ERPs and subjective ratings in response to appetizing food pictures as indices of appetitive responding under neutral and negative mood.

An important aspect of this kind of research is the efficient elicitation of negative emotion. How should negative emotional states be triggered in a powerful and yet ecologically valid way and food cue responses assessed simultaneously? Film viewing and social stress tasks make ERP measurements difficult. Furthermore, the individual triggers of emotion-induced eating vary greatly between individuals and a specific eating-relevant emotion has not yet been identified (Macht & Simons, 2011). Another line of emotion research has therefore relied on an idiosyncratic emotion elicitation technique. In this approach, participants are asked to report recent highly emotional events from their lives from which emotional or neutral scripts are generated and presented during the experimental task (e.g. Hilbert, Vogele, Tuschen-Caffier, & Hartmann, 2011). Neuroscientific studies have started to use this approach and successfully adapted it to the needs of event related, cue-reactivity designs by presenting cues and emotion elicitation scripts in an interleaved manner (Goldin, Manber-Ball, Werner, Heimberg, & Gross, 2009).

Drawing on these lines of research, the present study examined HEE and LEE participants on experiential and neuronal responding to high-calorie food images under script-driven neutral and idiosyncratic negative moods. We predicted increased food wanting (desire to eat) and food liking (palatability) in HEE and the opposite pattern in LEE (van Strien et al., 2012). Regarding ERPs, assuming that the late positive potential, LPP, reflects attention to stimuli that are salient due to motivational context (Stockburger et al., 2008, Stockburger et al., 2009b, Blechert et al., 2010) or eating habits/pathology (Blechert et al., 2011, Nikendei et al., 2012) we expected more positive amplitudes on this component in HEE relative to LEE. The LPP is also sensitive to cognitive regulation during viewing of emotional images (Hajcak, MacNamara, & Olvet, 2010) and appetitive foods (Meule et al., 2013, Sarlo et al., 2013). Thus, if food was rendered more salient under negative emotion, potentially due to its relevance for emotion regulation, HEE (but not LEE) might evidence an emotional modulation of the food-cue elicited LPP. Because of the evolutionary survival relevance of efficient food stimulus processing, earlier ERP activity might also discriminate groups and/or emotion states and will be looked at in an exploratory analysis (cf., Toepel et al., 2009).