Elsevier

Physiology & Behavior

Volume 90, Issue 1, 30 January 2007, Pages 36-42
Physiology & Behavior

Is it possible to dissociate ‘liking’ and ‘wanting’ for foods in humans? A novel experimental procedure

https://doi.org/10.1016/j.physbeh.2006.08.020Get rights and content

Abstract

Berridge's model (e.g. [Berridge KC. Food reward: Brain substrates of wanting and liking. Neurosci Biobehav Rev 1996;20:1–25.; Berridge KC, Robinson T E. Parsing reward. Trends Neurosci 2003;26:507–513.; Berridge KC. Motivation concepts in behavioral neuroscience. Physiol Behav 2004;81:179–209]) outlines the brain substrates thought to mediate food reward with distinct ‘liking’ (hedonic/affective) and ‘wanting’ (incentive salience/motivation) components. Understanding the dual aspects of food reward could throw light on food choice, appetite control and overconsumption. The present study reports the development of a procedure to measure these processes in humans. A computer-based paradigm was used to assess ‘liking’ (through pleasantness ratings) and ‘wanting’ (through forced-choice photographic procedure) for foods that varied in fat (high or low) and taste (savoury or sweet). 60 participants completed the program when hungry and after an ad libitum meal. Findings indicate a state (hungry–satiated)-dependent, partial dissociation between ‘liking’ and ‘wanting’ for generic food categories. In the hungry state, participants ‘wanted’ high-fat savoury > low-fat savoury with no corresponding difference in ‘liking’, and ‘liked’ high-fat sweet > low-fat sweet but did not differ in ‘wanting’ for these foods. In the satiated state, participants ‘liked’, but did not ‘want’, high-fat savoury > low-fat savoury, and ‘wanted’ but did not ‘like’ low-fat sweet > high-fat sweet. More differences in ‘liking’ and ‘wanting’ were observed when hungry than when satiated. This procedure provides the first step in proof of concept that ‘liking’ and ‘wanting’ can be dissociated in humans and can be further developed for foods varying along different dimensions. Other experimental procedures may also be devised to separate ‘liking’ and ‘wanting’.

Introduction

Neuropsychological research has done much to help our understanding of the brain substrates mediating food reward. For example, research at the University of Michigan led by Kent Berridge (e.g. [1], [2], [3]) has led to a model proposing that there are two distinct components involved. The first hedonic/affective component (termed ‘liking’) is the result of a central process incorporating not only sensory properties but also the individual's physiological state and associative history. The second incentive salience/motivation component (termed ‘wanting’) refers to an underlying implicit and objective drive process and can be seen as the directed impulse or demand for a targeted food stimulus. In operational terms this reflects the neural process that mediates a change in behaviour from active seeking of an object to ignoring it. The distinction of two neural substrates for ‘liking’ and ‘wanting’ components of food reward has fundamental implications for ingestive behaviour in humans. If food reward is determined by a dual process, the relative contribution of each component must be identified before changes in consumption can be fully understood. Unfortunately, direct study of these components in humans (i.e. by asking them in terms of liking and wanting to say how they feel) may not always be accurate or valid. In the case of ‘wanting’, problems are encountered when people fail to dissociate affective aspects from motivational aspects of the process (e.g. “It’s pleasant so I want it”). Indeed, the attribution of ‘wanting’ to an external stimulus transforms its basic sensory elements into incentives that are desired and alluring, but critically this, on its own, does not equate to pleasure. Furthermore, people assume that ‘wanting' is a process they are consciously aware of. However, the subjective perception of such feelings could be argued to be the product of an ‘active reconstruction’ by cognitive mechanisms, and this could lead to significant inaccuracies in the tracking of this process. This is particularly important, if incentive salience is to be isolated from other, more cognitive, forms of wanting. Introspective measures of ‘liking’ can also lead to potential inaccuracies. In human studies, these measures most commonly take the form of numerical scales with discrete labels at the beginning and end or at each point, or visual analogue scales where judgements are marked along a line, anchored by statements at each end. Such measures can be misleading as even subtle differences in the question or statement presented can affect responses. For example Rogers and Blundell [4] reported differences in the change in rated pleasantness across a meal depending on whether subjects rated the pleasantness of the taste of food or the pleasantness of eating that food. Indeed, the most striking, differences in interpretation are noted when the stable, intrinsic palatability of a food is confounded with the aggregate response elicited by ingesting the food, which may fluctuate [4], [5], [6] and could be attributed to changes in either palatability or motivation. ‘Liking’ is essentially an affective reaction reflecting the acute hedonic impact of a stimulus. For this reason, ultimately an introspective measure of ‘liking’ should be capable of tracking changes in the potency of the underlying neural reaction to the immediate reward of an external stimulus. A further problem arises when similar forms of measurement for ‘liking’ and ‘wanting’ are taken sequentially. If a person is tempted to perceive them as the same question, they might adjust their responses to be consistent and avoid dissonance. Therefore, an integrated measure of these processes must be constructed so as to treat the processes differently, and discourage the contamination of one conscious judgement by the other.

The ‘liking’/‘wanting’ model indicates that there may be a functional significance to the involvement of two systems of food reward in humans. If this is the case, then it is critical that valid ways to measure ‘liking’ and ‘wanting’ are established. It is logical for these measures to be applied simultaneously, and for the outputs to be compared. Additionally, the procedures should prevent subjects from discerning the nature of the enquiry, and therefore prevent the elicitation of heavily cognitively mediated responses. The present report describes the development of an experimental procedure to separate ‘liking’ and ‘wanting’ for foods in humans. The main objectives were to develop a paradigm that: a) is theoretically capable of measuring the concepts of ‘wanting’ (incentive salience) and ‘liking’ (hedonic response), b) uses dissimilar methodology to measure each process, c) is sufficiently sensitive to detect changes in and dissociations between these processes, and d) can be simply administered and efficiently computed.

Section snippets

Subjects

Subjects were recruited from the staff-student population of the University of Leeds. 30 males and 30 females aged 18–30 were selected from over 500 volunteers who responded to an e-mail requesting non-vegetarian subjects, willing to eat to fullness and complete a computerised questionnaire. These subjects were the first to meet the inclusion and exclusion criteria. Of the 60 participants tested, data from 7 were not included in the final analysis. One participant withdrew due to nausea and six

Subjective state measures and meal intervention

Males consumed on average 1511 kcal (S.D. = 481.6) compared to 1022 kcal (S.D. = 338.8) in females [t(51) = 4.28, p < 0.01]. As expected, the test meal caused a significant decrease in hunger (mean ± S.E.M.: 62.6 ± 18.7 mm) and increase in fullness (− 67.5 ± 17.6 mm) [t(52) = 24.35, p < 0.01 and t(52) =  27.87, p < 0.01, respectively]. Hedonic ratings of the test meal on the first mouthful and at the end of the meal revealed significant differences for both the ‘pleasantness/palatability’ (11.4 ± 2.0 mm) of the food [t

Discussion

The aim of this study was to test a novel experimental procedure to measure the dual components of ‘liking’ and ‘wanting’ for generic food categories in human subjects. We have based the procedure on two key dimensions associated with food choice; the fat content and taste of foods. These categories have often been associated with problems of controlling intake and with overconsumption (e.g. [8]). ‘Liking and ‘wanting’ have been assessed through separate techniques, but using the same target

Acknowledgements

This research was supported by Medical Research Council (MRC) Case award G78/8223 in conjunction with NRC, Lausanne.

Cited by (257)

View all citing articles on Scopus
View full text