The role of implicit wanting in relation to explicit liking and wanting for food: Implications for appetite control

Abstract

Eating is an action open to awareness by the individual; however, it cannot be claimed that processes that control the expression of eating habits are necessarily explicit. This distinction between implicit and explicit processes may enhance understanding of the expression of food reward (particularly the concepts of liking and wanting [Berridge, K. C., & Robinson, T. E. (2003). Parsing reward. Trends in Neurosciences, 26, 507–513] and its importance for human appetite control [Finlayson, G. S., King, N. A., & Blundell, J. E. (2007b). Liking vs. wanting food: Importance for human appetite control and weight regulation. Neuroscience and Biobehavioural Reviews, in press]. The present study investigated the effect of meal-induced satiation on implicit and explicit processes of liking (L) and wanting (W) by developing a computer-based procedure to measure L and W in hungry and satiated states. Explicit measures were derived from analogue ratings whilst an implicit W measure was derived from reaction time in a forced-choice procedure, which also identified food preferences. Seventy subjects (21.8±0.9 years, BMI: 22.2±0.5 kg/m²) completed the procedure before and immediately following consumption of a savoury test meal. Satiation caused explicit ratings of L and W to decrease in all food categories (p<0.01); but with a more marked decrease for savoury foods compared with sweet foods (p<0.01). Implicit W was increased for sweet categories (p<0.01), but not for savoury. Implicit and explicit measures of L and W independently correlated with preference for sweet foods. This study provides support that implicit and explicit processes of food reward can be simultaneously measured and dissociated using a test meal. Adjustments in hunger were linked to changes in explicit L and W in a manner consistent with sensory specific satiety, while a relationship between hunger and implicit W was absent. We suggest that implicit W is not systematically downregulated by the physiological consequences of food consumption in the same way as hunger and therefore may be largely independent of homoeostatic processes influencing intake.

Introduction

Traditionally, homoeostatic and hedonic influences over eating have been treated separately both through experimentation and theory development. However, recent considerations of homoeostatic and non-homoeostatic determinants of eating (Berthoud (2004), Berthoud (2006); Blundell & Finlayson, 2004) imply that such systems interact in the overall expression of appetite. Since eating is a behavioural act carried out through the voluntary skeletal musculature, it is often believed that eating is invariably under voluntary conscious control. Despite the fact that eating is an action open to awareness by the individual, it cannot be claimed that processes that control the expression of eating habits are necessarily explicit. Furthermore, introspective accounts of explicit processes are often subject to misinterpretation and cross-contamination (Berridge, 1996; Booth, 1987). An individual may have awareness of the act of eating per se whilst remaining unaware of the processes that determine the expression of appetite and the pattern of eating. It is, of course, obvious that a person cannot be aware of the (implicit) changes in neuropeptides, hormones or other physiological processes that help to determine the initiation, termination and topography of eating. Moreover, in considering just the psychological mechanisms that influence eating, we assume that both implicit and explicit processes will operate. Verplanken (Verplanken, 2006; Verplanken & Aarts, 1999) argued that the formation of habits should be constructed as a psychological process with automatic (implicit) features that operate somewhat independently to the explicit experience of a repeated behaviour. Using a 12-item measure derived from the conception of habit as an implicit process (self-report habit index; Verplanken & Orbell, 2003), Verplanken (2006) demonstrated that the consumption of unhealthy snack food during a week was better accounted for by their measure of habit than by behavioural frequency of snack consumption during the previous week.

Recently, implicit and explicit processes of reward and their involvement in the regulation of feeding behaviour have been examined in more detail. Through the examination of specific neural substrates in the brain, Berridge and colleagues argue that distinct processes of affective “liking” and motivational “wanting” for food can influence behaviour without explicit awareness of their underlying cause, i.e. changes in hedonic feelings (explicit liking) or the intent or desire to consume a specific food (explicit wanting) (Berridge & Robinson, 2003; Finlayson, King, & Blundell, 2007b). To test the influence of implicit processes on measurable aspects of feeding behaviour, Winkielman, Berridge, and Wilbarger (2005) subliminally presented subjects with emotionally valenced picture stimuli (16 ms exposure to positive or negative facial expressions) which were subsequently masked with neutral faces (400 ms exposure). Interestingly, this implicit affective priming technique had no effect on the subject's explicit liking ratings recorded online, but interacted with thirst levels such that positive facial expressions increased the serving size and consumption of a beverage in thirsty subjects, while negative facial expressions inhibited drinking, relative to neutral primes.

In a recent review, Mela (2006) explored the liking/wanting distinction of food reward in human appetite and suggested that implicit measures of wanting (e.g. behavioural tasks, physiological correlates, etc.) are more valuable assays of obese/lean differences in food reward and food preferences than explicit liking or wanting responses. For example, Saelens and Epstein (1996) assessed the reinforcing value of food using a slot-machine-like progressive ratio computer task. In this paradigm, subjects’ commitment to the task was rewarded with points that could be exchanged for palatable snacks or allotments of time that could be spent playing a fun computer game. The reinforcing value of the palatable food was calibrated as the willingness to work for amounts of the food relative to the time playing the game. In a study comparing obese and lean subjects, the authors found that ratings of explicit liking for snack food items—including the most preferred item used in the progressive ratio task—did not differ between the groups; however, the obese subjects were found to work harder (i.e. greater implicit wanting) for food relative to playing the game, and this corresponded to the amount subsequently consumed. These findings may also demonstrate important differences in implicit and explicit processes as mediating variables in food consumption. Although explicit liking for foods as they are ingested may be involved in establishing their reinforcing value, it is possible for implicit processes to play a more significant role in maintaining consumption (e.g. Berridge, 1996; Robinson & Berridge, 1993) and this could facilitate over-consumption in people at risk of weight gain.

To further examine the significance of dual liking and wanting components of food reward for appetite control, we previously reported the development of a novel computer-based procedure to allow the separate and concurrent assessments of explicit and implicit liking and wanting for the same target stimuli while preventing cross-contamination of these measures (Finlayson, King, & Blundell, 2007a). This was achieved by using a “forced choice” behavioural measure of implicit wanting in addition to explicit subjective measures of liking and wanting for photographic food stimuli varying in fat content and taste. The present study reports a key modification to our previous operation of implicit wanting in order to improve its validity. In line with other laboratories successful use of reaction time as an indicator of implicit processes (for example, the Implicit Association Test; Greenwald, McGhee, & Schwartz, 1998), implicit wanting was operationalised as the reaction time of each pair trial decision. Thus, the speed with which one stimulus is chosen in preference to its alternative provides a quantifiable measure which may relate to the implicit wanting for that food item. We were interested to explore whether this new measure could convey information about the degree (on a continuous, scaled unit of measurement) to which the chosen stimuli is wanted relative to its alternative. Furthermore, mean reaction time for each food category (distinguished according to fat content and taste) can give an indication of whether motivation is generally increasing or decreasing independent of the other food categories.

Therefore, the aim of this study was to extend our previous exploratory investigation into the effect of manipulating hunger state—by meal-induced satiation—on novel laboratory measures of implicit and explicit liking and wanting processes. By measuring explicit liking (subjective feelings of orosensory pleasure), explicit wanting (subjective feelings of intent or desire), implicit wanting (motivational expression of reward attribution) and relative preference (forced-choice) for a range of food stimuli in hungry and satiated states, changes in these processes were assessed and compared.