This is the author’s version of a work that was submitted/accepted for pub-
lication in the following source:
Finlayson, Graham, King, Neil A., & Blundell, John E. (2007) Liking vs.
wanting food : importance for human appetite control and weight regula-
tion. Neuroscience & Biobehavioral Reviews, 31(7), pp. 987-1002.
This file was downloaded from: http://eprints.qut.edu.au/33693/
c ? Copyright c ? 2007 Elsevier Ltd. All rights reserved.
this is the author’s version of a work that was accepted for publication
in Neuroscience & Biobehavioral Reviews. Changes resulting from the
publishing process, such as peer review, editing, corrections, structural
formatting, and other quality control mechanisms may not be reflected in
this document. Changes may have been made to this work since it was
submitted for publication. A definitive version was subsequently published
in Neuroscience & Biobehavioral Reviews, [VOL 31, ISSUE , (2007)] DOI
Notice: Changes introduced as a result of publishing processes such as
copy-editing and formatting may not be reflected in this document. For a
definitive version of this work, please refer to the published source:
Liking vs. wanting food: Importance for human appetite control and weight
Graham Finlayson*1, Neil King1, John E. Blundell1
1Biopsychology Group, Institute of Psychological Sciences, University of Leeds, Leeds
LS2 9JT, UK
*E-mail address: email@example.com (Graham Finlayson)
Telephone: +44 (0) 113 343 5753
Fax: +44 (0) 113 343 6674
According to the French National Nutrition-Health Program (2001-2005) it is
essential that an individual’s food choice remains a ‘free act’ and that eating is
recognised as a moment of pure pleasure. In order to fully appreciate these
premeditated episodes of sensuality, it is not sufficient to focus only on the hedonic
sensations arising from events in the mouth. The joy of eating can be as much to do
with the preparation and effort that one invests in their chosen food and of course the
expectancy and anticipation that intensify then peak in the final moments before
ingestion. In summary, the pleasure of food can be seen as an interaction of liking
and wanting, and experiencing one without the other – although pleasurable in
isolation – stops short of full reward.
Advances in neurobiology are helping to characterise the substrate mediating hedonic
processes of consumption, and they are precipitating the emergence of a new
conceptual approach to reward where affect and motivation (a.k.a. liking and wanting)
can be seen as the major force in guiding human eating behaviour. This concept is
especially important for the study of ingestive behaviour in the modern world, where
food is plentiful, cheap, energy-dense, and enticing, and physical activity is being
reduced to a luxury afforded by environment and lifestyle. We have reached an age
where weight control has been turned upside down from an instinctual, highly
regulated system, to a process requiring considerable cognitive effort (Peters et al.
2002). Furthermore, where overweight and obesity have taken hold, losing weight
and defending that loss (especially in an environment where our hedonic drives are
encouraged and exploited) can change from a daily struggle, to a losing battle (e.g.
Ikeda et al. 2005).
Parsing reward from a unitary process into distinguishable liking and wanting
components in neurobiological studies (Berridge & Robinson, 2003) has struck a
chord that is resonating across many disciplines and in different areas of research. In
the field of ingestive behaviour alone it has implications for characterizing eating
disorders and obesity, identifying pharmacological targets, the psychology of appetite
control, phenotypic profiling of resistance and susceptibility to weight gain, and
industrial product development. Despite the possibilities of a dual process modulation
of food reward, several issues remain to be addressed: How can these concepts be
operationalised for use in human appetite research? Can they be translated into
observable entities that reflect the neural mechanisms by which they may be
influenced? Do liking and wanting operate independently to produce functionally
significant changes in behaviour? Can liking and wanting be truly separated or will
an expression of one inevitably contain elements of the other? In this review, current
progress in applying processes of liking and wanting to the study of human appetite
and ingestive behaviour are examined and the importance of these concepts for human
appetite research are discussed.
Neurobiological study of liking and wanting
Our capacities in neuroscience can reveal – to some extent – the circuits responsible
for the reward we derive from food. The picture emerging is that food reward, rather
than being a unitary neurological entity, is represented functionally and structurally by
distinct components. One such distinction, between processes associated with
affective vs. motivational consequences of ingesting food, has received much recent
attention (Berridge, 1996). With principle focus on opioid neurotransmission in the
nucleus accumbens shell and the mesolimbic dopamine system, research has shown
that core processes of ‘liking’ and ‘wanting’ can be separately manipulated in rodent
models to produce patterns of behaviour that are either exclusively affective or
motivational in conjunction with a food stimulus (see
http://www.lsa.umich.edu/psych/research&labs/berridge/Publications.htm for a
comprehensive list of publications).
How is liking measured?
In its simplest description, ‘liking’ is most commonly operationalised as the change in
affect observed using a technique to analyse taste reactivity patterns in rats (Grill &
Norgren, 1978). This involves a careful assessment of hedonic and aversive
behavioural reactions thought to be universal affective expressions, some of which
can also be observed – allowing for speed and body size – in primates and human
infants (Berridge, 2000). Taste reactivity patterns are thought to provide a relatively
pure indication of affect because they can be isolated from the sensory properties of a
taste (see Berridge, 2000), they can be dissociated from the desire to eat (e.g. Berridge
& Valenstein, 1991; Berridge et al. 1989) and they often correspond to human
subjective ratings of palatability.
How is wanting measured?
If liking is operationalised as the pattern of reactive behaviours associated with
affective aspects of food reward, then wanting is commonly measured by changes in
the propensity to eat that are independent of shifts in liking. Therefore, any measure
that requires the subject to actively engage with its environment in pursuit of a known
food stimulus can be said to be contain at least an element of wanting. It is worth
emphasising at this juncture however that wanting (specifically termed by some as
incentive salience attribution) is not adequately captured by appetitive drive or the
non-specific desire for food in general. Wanting is hypothesised to be the
consequence of an active process of assigning value to perceptual or representational
events wherein sensory and cognitive inputs are transformed into desirable, attractive
entities (Berridge, 1996). Hence wanting is likely to be modulated by sensory and/or
cognitive influences which set it apart from other appetitive processes (e.g. needing).
Wanting implies a direction, not just a force. Therefore, obtaining a measure of
wanting that can be dissociated from a non-specific drive to eat (e.g. see Beagley &
Holley, 1977 as cited in Berridge, 1996) can become awkward when a further
separation from liking is required. In passing, the investigation of specific food
cravings may provide a useful platform from which to further clarify these processes
(e.g. Robinson & Berridge, 1993; Robinson & Berridge, 2000 Pelchat, 2002).
Nevertheless, the more satisfactory measures of wanting concern the anticipatory or
instrumental phase of reward seeking behaviour.
Implicit and explicit components
Through the separate examination of specific neural substrates in the brain,
behavioural measures of liking and wanting are thought to reflect ‘core’ processes that
can operate without conscious awareness. However, these implicit components
clearly have their explicit counterparts which express themselves subjectively in the
form of hedonic feelings from the ingestion of a specific food (conscious liking) and
the intent or desire to consume a specific food (conscious wanting) (Berridge &
Robinson, 2003). Therefore it is important that we know how core liking and wanting
(operating at implicit and explicit levels of conscious awareness) might relate to
Berridge (2004) used the term ‘niceness gloss’ (the pleasantness added by the brain to
the sensory qualities of sweetness) in reference to the explicit component of liking,
with implicit liking and its associated brain structures forming an unconscious core.
In this review, implicit liking was presented as capable of producing objective
reactions without subjective awareness of their underlying cause. For instance, the
subliminal presentation of picture stimuli (positive, neutral or negative facial
expressions) caused no change in self reported mood ratings, but the emotional
valence of the stimuli was positively associated with pleasantness rating, consumption
and rated value of a beverage (Winkielman et al. 2005).
As explained above, the implicit component of wanting is linked to the attribution of
‘incentive salience’ to external stimuli. Explicit wanting (truer to the colloquial
understanding of the word), has been described as the conscious desire for a
cognitively represented outcome (Berridge, 2004). Interestingly, implicit wanting has
also been shown capable of producing objective reactions independent of any
subjective awareness (Berridge, 2004). For instance, it has been demonstrated that
animals can show ‘irrational wanting’ for food rewards. In one study, activation of
the dopamine system caused a reward cue (conditioned stimulus) for sucrose to
become a hyper incentive, temporarily outstripping the reward value of the sucrose
presented alone or with an irrelevant cue (Wyvell & Berridge, 2000). Such findings
support the hypothesis that implicit wanting can be a more compelling influence on
behaviour than explicit (conscious) wanting (Berridge, 2003).
In summary, liking and wanting are viewed as core processes with implicit and
explicit features. Explicit liking (acutely perceived hedonic reaction) can be
associated with explicit wanting (subjective desire for a perceived goal) and implicit
wanting (incentive salience attribution). Implicit liking (unconscious affect) can also
influence implicit wanting to influence ingestive behaviour without the subjective
awareness of either process. Lastly, wanting can be irrational when implicit wanting
for a reward is greater than explicit wanting, and not proportional to the experience or
expectation of liking (e.g. Wyvell & Berridge, 2000).
From core processes to constructs in human appetite
If core processes of liking and wanting can independently modulate food reward both
implicitly and explicitly, it is important to understand how they can be approached
and rendered suitable for the study of ingestive behaviour in humans. Even to date,
the role of food reward in human appetite is mostly treated as a single entity
embodied by a ‘palatability’ or ‘pleasantness’ factor and its effects on appetite control
(Yeomans, 1998). The logical view is that liking and wanting co-vary in a natural
two-way sequence. Therefore in behavioural terms we assume that a change in liking
– e.g. through manipulating palatability – will lead to proportional adjustments in
wanting – e.g. amount of the commodity consumed (Bobroff & Kissileff, 1986;
Yeomans et al. 1997), and likewise changes in wanting – e.g. indirectly, through
manipulating hunger and satiety – will effect changes in liking – e.g. hedonic
response (Cabanac, 1979; Carr & Wolinsky, 1993). A dissociation of liking and
wanting in the brain suggests that this self-evident liking↔wanting sequence may not
always hold true, and this could throw new light on the way human ingestive
behaviour is understood.
Translating liking and wanting into constructs amenable to the study of human
ingestive behaviour has a number of obstacles, and most concern the accuracy and
accessibility of our subjective consciousness. Firstly, our feelings may not do justice
to the complexity of their underlying processes. For example, is the hedonic strength
of a response a consequence of one underlying dimension (e.g. pleasantness) or a
convergence of two (e.g. pleasantness and unpleasantness) or three (e.g. pleasantness,
unpleasantness, and salience) or numerous other dimensions? Affective taste
reactivity in rats and infants can simultaneously assess hedonic (i.e. positive) and
aversive (i.e. negative) aspects of affect (Berridge, 2000), but asking humans to
introspectively provide the same information may be too contradictory and confusing.
In human appetite research, it has been suggested that distinct sets of underlying
processes can be interpreted as a single more general variable which is only then
partitioned cognitively into the required domains (Booth, 1987). The distorting
potential of cognitions on affect is cautioned by Berridge (Berridge, 1996) where the
accuracy of actively reconstructing emotional events is vulnerable to excessive
cognitive processing. Hence – paradoxically – the more someone is encouraged to
consider their feelings, the less reliable their responses may become. However,
introspectionist constructions of underlying processes can refer to unconscious
experiences as well as those in conscious awareness (e.g. hunger, craving, fullness,
thirst), but it must be remembered that this method can only reveal a causal
relationship between the construct and the behaviour for the elements of the
experience that are explicit and subjective (Booth 1987a; Booth 1987b; Booth & Blair
1988). An adequate resolution of this problem requires more than the examination of
these constructs and their behavioural correlates. In the case of liking and wanting,
the utility of introspective constructions can only be judged in the context of their
interaction with objective mental/behavioural outputs and the ability to distinguish
these from other environmental contingencies (e.g. Booth 1990). Alternatively,
measures that are able to better reflect the unconscious components of these
constructs or are less reliant on the cognitions of the subject, or are specifically
designed to test for the dissociability of distinct underlying processes, could greatly
facilitate a dual-process approach to the study of reward in human appetite research.
Reward and appetite control: homeostatic and hedonic interplay
A key issue in the study of appetite control is the relationship between reward and
homeostatic drives arising from biological needs (Yeomans et al. 2004). Historically,
hedonic processes have been viewed as a function of nutritional need-state. In a state
of depletion, the hedonic response (experienced palatability) to energy providing
foods is enhanced and when replete, the hedonic effect of these foods is reduced
(Cabanac, 1989). This view is compatible with the link between energy density and
palatability (Drewnowski, 1998) and also that the consumption of fats and sugars –
rich sources of energy – may be under neuro-regulatory control (Levine et al. 2003).
However, the idea of reward as a consequence of the fulfilment of nutritional need is
not broad enough to explain non-homeostatic ingestive behaviour (non-compensated
patterns of over or under consumption) and it is perhaps more useful to try and
distinguish the neural substrates of homeostatic and hedonic systems and to assign
them separate identities (Blundell & Finlayson 2004).
Homeostasis and hedonics: Separate identities…
The homeostatic substrate comprises a network of neuropeptides and biogenic
aminergic neurotransmitters which link peripheral and central components. This
system has been well characterised (e.g. Schwartz et al. 2000) and involves insulin,
leptin, NPY, AgRP, MSH, CART, GLP-1, orexins, ghrelins, PYY, and other peptides
along with serotonin pathways and other aminergic systems. A biological substrate
mediating the reward processes of consumption is also being characterised and
ostensibly involves glutamate, benzodiazepines, endocannabinoids, opioids and
dopamine pathways (e.g. Saper et al. 2002; Flier, 2004). The implication of distinct
neural substrates for homeostatic and hedonic systems is that processes of reward can
operate free from biological need, and the extent to which this occurs can be
investigated. For example, pharmacological evidence suggests that these circuits are
somewhat separate. In obese subjects, administration of the serotonin drug D-
fenfluramine suppressed the sensation of hunger but had no effect on the appreciation
of the pleasantness of food (Blundell & Hill, 1987). Conversely, an opioid antagonist
reduced the rated pleasantness of palatable foods but had no effect on hunger
(Yeomans & Gray 2002). This double-dissociation concept indicates that hedonic
aspects of reward are associated with a specific biological substrate that can be
pharmacologically dissected from the substrate-mediating hunger (Rogers & Blundell,
1991). This is supported by evidence from animal studies. In one study (Giraudo et
al. 1999), saline (control), NPY or an opioid agonist (DAMGO) was injected into the
paraventricular nucleus of rats. The rats could freely consume from standard chow
and 10% sucrose solution. After injection of NPY, food intake was increased relative
to saline, and the rats were found to consume approximately half their calories from
the chow and half from the sucrose solution. Injection of DAMGO also stimulated
intake, but in this condition 85% of calories came from the sucrose. Therefore, NPY
and opiates may represent a demarcation between energy-driven versus reward-driven
feeding. However, it is still possible for a functional interaction to occur when the
manipulation is made through the natural commodity (food) rather than through more
selective pharmaceutical manipulations.
…or inseparable entities?
Advances in our understanding of the molecular and neural mechanisms behind
appetite regulation are revealing how the reward system can interact with homeostatic
mechanisms. As mentioned above, cannabinoid receptors and their endogenous
ligands (e.g. anandamide) are implicated in the reward system. Peripheral and central
administration of anandamide increased appetite in rodents, and this seemed to be
related to alterations in incentive value (desire) for palatable foods (Kirkham &
Williams, 2001). However, the cannabinoid system has been shown to interact with
homeostatic processes in a number of ways (Stanley et al. 2005): Leptin signalling
becomes defective when hypothalamic endocannabinoid levels are high (Di Marzo et
al. 2001); activation of CB1 receptors prevent the melanocortin system from altering
food intake (Verty et al. 2004); furthermore, CB1 receptors can be found on
adipocytes where they may directly increase lipogenesis (Cota et al. 2003). Opioid
neurotransmission also forms part of the biological substrate mediating reward
processes of consumption. For example, endogenous opioids are associated with the
reinforcing effect of food (especially when palatable) (Welch et al. 1996; Yamamoto
et al. 2000). However, there is evidence to show that in a fasted state, the reinforcing
effect of food can be reinstated in enkephalin and β-endorphin knock-out mice
(Hayward et al. 2002). Therefore, homeostatic processes may interact with hedonic
signalling to override selective reward deficit. Erlanson-Albertsson (2005)
summarised how ingestion of palatable food can offset normal (homeostatic) appetite
regulation. In the brain, research shows that energy deficit is registered in the
hypothalamus leading to the release of hunger signals and the activation of their
receptors. Consumption of ‘standard’ food generates information on its energy
content and taste in the brain stem. This information is transmitted to the
hypothalamus leading to the release or upregulation of various satiety peptides,
causing consumption to cease. However, a different scenario is apparent when the
reward system is activated by highly palatable food. With ingestion of palatable food,
taste sensing is different than with standard food; information is transmitted to the
reward circuit, leading to the release or upregulation of reward mediators like
dopamine, endocannabinoids, and opiates. The reward circuit has connections with
appetite-controlling neurones in the hypothalamus that can increase the expression of
hunger peptides such as NPY and orexins, while blunting the signalling of satiety
peptides like insulin, leptin and cholecystokinin. Therefore when food is highly
palatable, the drive to eat is maintained, with continued eating now mediated by
reward rather than biological need (see figure 1). Hence although homeostatic and
hedonic systems can be given separate identities (Blundell & Finlayson 2004), they
are also – to an extent – inseparable, with neural cross-talk permitting functional
interactions which may influence the organisation of ingestive behaviour. From this
standpoint, the interaction of homeostatic and non-homeostatic pathways in the neuro-
regulatory control of feeding may be given more importance than the two systems
studied in isolation. From behavioural and anatomical observations, Berthoud (2006)
suggested that projections from the hypothalamus to the nucleus accumbens may
modulate the motivation to feed via metabolic signals. Furthermore, direct and
indirect projections from the accumbens to the hypothalamus may explain the ability
for mesolimbic processes – activated by relevant environmental cues and incentives –
to essentially hijack the homeostatic regulatory circuits and drive up energy intake.
Further research is necessary to identify the pathways that mediate such interactions;
however some progress has been made (see Berthoud 2004).
Liking, wanting and ingestive behaviour: A re-examination of selected studies
Considering that most studies have repeatedly shown that palatability – a factor
influencing the reward value of food – has an effect on intake, this would support the
notion that reward plays a role in the process of satiation (de Graaf, 1999). However,
it is uncertain how processes of liking and wanting might independently modulate the
effect of reward on appetite to influence ingestive behaviour. With a dual process
perspective on reward, it becomes possible to re-examine some of the previous
research investigating its impact on human appetite.
Intake and meal size
Manipulating the palatability of test foods is a common way to investigate how
reward processes can influence ingestive behaviour. Most research using such an
approach has found increased palatability to have a stimulatory effect on intake and
meal size, leading some authors to propose a quantitative relationship between
subjective shifts in palatability and corresponding adjustment of food intake in grams
(e.g. Bobroff & Kissileff, 1986). However, differences in the palatability of foods do
not provide a comprehensive indication of their reward value. In most cases a
palatability manipulation (brought about by a change in the physical or sensory
properties of a food) can only account for differences in liking for the food. Less can
be assumed about the effects of the same manipulation on the desire or motivation to
consume the food. It is interesting therefore, that a small number studies do not find
an association between palatability and food intake. For example, supplements of
monosodium L-glutamate enhanced the palatability of a soup preload, but had no
effect on the consumption of a test meal delivered 2 or 30 minutes later (Rogers &
Blundell, 1990). Similarly, manipulation of the palatability of fat or carbohydrate
based meals did not significantly decrease intake despite reduced hunger and
enhanced satiety (Warwick et al. 1993). In one study, manipulated palatability of the
test foods correlated to overall intake in only half the subjects (Lucas & Bellisle,
1987). Some free-living studies also indicate that the role of liking in determining
meal size may not be as crucial as many laboratory studies have suggested. For
instance, in obese and lean subjects, no correlation was found between liking for
preferred taste (sweet foods) and total intake of foods in that taste category, and there
was very little variability between mean hedonic ratings for each taste category
(sweet, savoury/salty, bitter, sour) in self-selected foods (Cox et al. 1999). Other
research on free-living subjects in France and North America found a strong
association between palatability and meal size, but demonstrated that over 70% of all
recorded meals were rated at above neutral palatability (deCastro et al. 2000a; de
Castro et al. 2000b). These studies suggest that the subjective pleasantness of meals
is likely to play a role in food choice, but may be less important in accounting for the
variability in amount consumed. Indeed, it is intuitive that in real life, where people
have freedom of choice, palatability will be a relatively consistent factor in all meals.
Therefore, in isolation liking for foods may not tell the whole story about reward
driven food intake.
For example, in one quasi-experimental study, three different types of test meal
(conventional four course, sandwich, and semi-liquid) at two levels of palatability
(high or low) were compared after covert video recording (Guy-Grand et al. 1994).
The palatability manipulation was only found to stimulate intake in the conventional
meal, yet it produced comparable differences in subjective liking for all the meals.
However, in the sandwich meal, bite rate was significantly greater in the lower
palatability condition, and in the semi-liquid meal, pause duration was less. These
behaviours seem at odds with the lower ratings of pleasantness for this level of
palatability. Indeed, if bite rate and pause duration are taken as indicators of the
subjects’ overall motivation toward their food, then it appears that wanting was to
some extent higher, despite the meals being less liked. Interestingly, studies that have
employed some measure of eating rate in their methodologies have commonly found
the palatability-dependent effects of rate on intake were confined to the initial stages
of the eating bout (Bellisle et al. 1984; Bobroff & Kissileff, 1986; Spiegel et al. 1989;
Yeomans, 1996) – presumably when motivation to eat is at a peak. Such measures
may help to discern and separately track processes of liking and wanting within a
Hunger and satiety
Previously in the literature, it has been disputed whether explicit liking may have an
influence on hunger and satiety (see de Graaf, 1999). One early study (Hill et al.
1984) found an enhancing effect of an equi-caloric but preferred meal on rated
hunger. Even the sight of the preferred food increased hunger suggesting that the
expectancy or anticipation of food (wanting) was stimulating appetite before liking
for the food could be confirmed. Other research has found a hunger-dependent effect
of palatability on intake (Spiegel et al. 1989). When lean subjects were deprived of
breakfast, there was no difference in intake when offered either low or high
palatability foods at lunch. When breakfast was reinstated, intake of the high
palatability food was greater. One proposal is that liking is an important factor in
food intake within an acceptable threshold of hunger; once this threshold is crossed,
the palatability of available foods is then secondary to perceived energy requirements.
Therefore, the mediating effect of liking on food intake may be overridden by potent
homeostatic signals. Conversely, in a study where disguised high or low energy
preloads were administered before tasty or bland test meals (Yeomans et al. 2001),
subjects failed to compensate for the preloads when subsequent food was more
palatable. It was argued that tasty foods prevented short-term responses to satiety
cues. Thus, homeostatic signals may also be overridden by liking. However the
system does appear to operate asymmetrically; although enhanced liking for foods can
augment hunger (and therefore food intake), the presence of strong satiety developing
over the course of a meal does not always downregulate a food’s palatability (e.g. see
Yeomans & Symes, 1999; Looy et al. 1992).
Ratings of subjective liking
Nearly all human studies interested in reward processes in appetite include some
subjective measure of liking. Nevertheless, several studies have revealed that such
introspective ratings may be of limited use. The most common tests involving
subjective ratings of liking are brief exposure tests and fixed quantity tests. Although
easy to conduct, these tests are considerably different to free-living eating situations
where much larger quantities of food are eaten and in combination with other foods
and drinks. Bellisle et al. (1984), Lucas & Bellisle (1987), Monneuse et al. (1991),
Perez et al. (1994) and Zandstra et al. (1999) have all found brief exposure tests to
give a biased estimation of the optimal palatability of a food. Zandstra et al. (1999)
found consumption to correlate best with ad libitum tests and worst with taste-and-spit
tests. This finding gains support from several other studies who find liking ratings
alone insufficient to predict subsequent intake (Bellisle & Le Magnen, 1980; 1981;
Helleman & Tuorila, 1991). This point has been explored elsewhere (Mook & Votaw
1991): By asking college students to rank or choose from a list of reasons for
terminating a meal, the authors found hedonic factors to be of little perceived
importance. They cautioned that studies employing subjective ratings may enhance
the perceived importance of the hedonic properties of test foods that otherwise might
not be considered. For instance, one study investigated repeated within meal
interruptions (in order to make subjective appetite and hedonic ratings) and their
effect on intake (Yeomans et al. 1997). Surprisingly, when meals were interrupted at
every 50g interval of intake, overall consumption was significantly greater compared
to one continuous bout of feeding. It is possible that by repeatedly eliciting a
conscious, subjective evaluation of liking for available foods, the incentive salience of
the food may also have been inadvertently enhanced.
Dual-component contributions to the study of reward in human appetite
With a dual-process model of reward, retrospective evaluation of the literature may
help to throw light on the relationship between food reward, appetite and ingestive
behaviour. More recently, researchers are beginning to consider liking and wanting
interpretations of their own study findings and some are specifically tuning their
methodologies to allow for separations of motivation and affective responding to
food. These studies provide examples of how liking and wanting may be
operationally defined in human appetite research.
In one study (Zandstra et al. 2000), three levels of palatable food were consumed as
open sandwiches each day for three weeks (5 days repeated exposure per level of
palatability). As well as measuring ad libitum food intake and subjective appetite on
each day, change in rated pleasantness of the food and change in rated desire-to-eat
were assessed by subtracting post-meal ratings from ratings taken after the first bite.
The study found that intake on the first two days reflected the manipulated palatability
of the foods, with proportional differences between low, medium and high levels. On
the third day of repeated exposure, intake between the medium and high levels was
equivalent, and on the fifth day, intake did not vary between any of the conditions.
Similarly, ratings of fullness suggested that satiety following the low palatability food
was increasing with time. On examination of ratings of pleasantness and desire-to-
eat, an interesting dissociation was observed. After a considerable reduction in
desire-to-eat on the first day of exposure for the low palatability food relative to the
other levels, this decline dissipated after a few days and ratings for all foods became
equivalent. In contrast, the change in rated pleasantness for each level of palatability
was stable over time. Therefore, the convergence of intake and satiety for three levels
of palatability brought about over five days of repeated exposure was associated with
changes in desire to eat (explicit wanting) and not rated pleasantness (explicit liking)
for the foods. More recently, these explicit processes have also been dissociated by
manipulating expectations by labelling or withholding information about fresh versus
processed food (Zandstra et al. 2004). When consumed blind, foods were desired and
liked similarly. However, when labelled as fresh or processed, desire for the fresh
food was greater, with no change in liking. The findings of both studies may justify
the importance of differentiating questions of liking and desire (explicit wanting) for
foods. The authors suggested that without the opportunity to rate liking and wanting
aspects of the foods separately, subjects’ ratings may have contained elements of
both. In other words, they thought the explicit components of each underlying
process may have been interpreted as single subjective feeling.
Another study compared lean and obese subjects on their liking for an array of food
products (sandwiches or snacks) in hungry and satiated states (Snoek et al. 2004).
Subjects tested and rated a selection of foods before consuming one food item to
fullness as an ad libitum meal. Ratings of appetite that were specific (e.g. “appetite
for a meal”, “appetite for something sweet”, “appetite for a snack” etc.) as well as
non-specific (e.g. hunger, fullness) were recorded immediately before and after a
sandwich or snack ‘meal’. Results showed that obese and lean subjects did not rate
the pleasantness of any of the food items differently after consuming the meal.
However, differences were apparent in some of the appetite ratings; specifically
“appetite for a snack” and “appetite for a meal” following consumption of either
snacks or sandwiches, and also “appetite for something savoury” after snack
consumption. The authors speculated that obese-lean differences referred to
differences in wanting and not liking, on the basis of higher perceived energy needs of
obese subjects. The authors were careful however, to point out that their measures
“reflect conscious, rationalised liking and wanting” (p.830). They suggested that to
study the differences in obese and lean humans will require improved more objective
techniques to measure liking and wanting, for example through neuroimaging
techniques, or indirect methods that do not rely on subjective ratings.
One neuroimaging study (Small et al. 2001) used consecutive PET scans to track
changes in brain activity as subjects consumed chocolate to beyond satiety. In this
design, subjects ate one piece of chocolate at a time, followed by a rating of how
pleasant or unpleasant they found it (explicit liking), and how much they would like
or not like to have another piece of chocolate (explicit wanting). Each feeding period
– with a scan followed by a rest period – was determined by the amount of chocolate
it took to produce a two unit decrease in rated liking. Thus, as subjects ate chocolate
from a state where it was highly motivating and pleasurable, to a state where
motivation to stop eating was high and chocolate consumption unpleasant, it was
possible to observe the differential recruitment of brain regions as the reward value of
the chocolate decreased. The study found that subjects ratings of wanting decreased
faster and to a greater extent than ratings of liking, suggesting that some separation of
these constructs was occurring. In order to see if these processes could be identified
physically in the brain, scan order was controlled for and regional cerebral blood flow
was regressed against subjective ratings of liking and wanting. It was found that
activity in a region of the retrosplenial cortex correlated to a greater extent to ratings
of wanting than ratings of liking. The authors suggested that this region may form
part of the substrate for a dissociation between liking and wanting. Indeed, studies
that incorporate objective measures of brain activity in conjunction with sensory and
motivational challenges may help to further characterise separable processes of
reward. As these techniques become more sophisticated it is likely that more will be
understood about the connection between implicit and explicit components of reward
and their influence on ingestive behaviour in more complex scenarios.
One way to overcome the difficulties of extracting reliable subjective responses may
be to find alternative indirect measures of motivation or affect. For example, Saelens
& 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 amounts of tasty snack food or
allotments of time that could be spent playing an enjoyable computer game. The
reinforcing value of the tasty 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 subjective ratings of liking for snack
food items – including the most preferred item used in the progressive ratio task – did
not differ; however, the obese subjects were found to work harder for food relative to
playing the game, and this corresponded to the amount of calories consumed. These
findings confirmed earlier research associating obesity with willingness to work for
food in the presence of food cues (Johnson, 1974). These findings may also
demonstrate important differences in reward-driven behaviour in these populations.
Although liking for rewards may be involved in establishing their reinforcing value, it
is possible for wanting to become a more significant factor in influencing ingestive
behaviour (e.g. Robinson & Berridge, 1993). Furthermore, the lack of differences in
subjective hunger ratings suggests that the differences in reward-driven behaviour
were somewhat independent of homeostatic processes.
Using a similar paradigm, the effects of food deprivation on hedonics (liking) and
reinforcing value (wanting) of foods was assessed (Epstein et al. 2003). Subjects
were assigned to satiated (i.e. fed) or hungry (i.e. unfed) conditions and the subjective
pleasantness of a range of drinks as well as the reinforcement value of a preferred
snack item were measured before and after a satiating meal or equivalent time
abstaining from food. There were no differences in temporal profiles of pleasantness
between the two conditions. In contrast, the deprived subjects responded more for
food on the progressive ratio task compared to fed subjects. The authors concluded
that food deprivation influenced the motivation to eat, but not the hedonics of a range
of drinks. These findings are consistent with a dual-component view of reward.
Wanting was shown to increase under deprived conditions while liking did not
change. In passing, liking was found to decrease for the most pleasant drink in the fed
state. Although non-significant, this finding may reveal certain conditions (e.g.
satiated) where liking can decrease independently from wanting.
In our laboratory, some progress has been made in developing a methodology to
separately assess and observe dissociation between liking and wanting (Finlayson et
al. In Press). We designed a novel computer-based procedure to identify liking and
wanting for the same target food stimuli through separate tasks consisting of a number
of response trials. These separate procedures prevented cross-contamination in the
evaluation of liking and wanting. The presentation of trials for both tasks was
integrated and fully randomised into a single executable program. The stimuli were a
number of photographic food items selected to vary in fat content and taste (two key
dimensions associated with loss of appetite control and overconsumption). In this
procedure, liking is measured explicitly using visual analogue scales combined with
the prompt “how pleasant would it be to experience a mouthful of this food now?”
this question was designed to carefully direct attention to the (imagined) hedonic
experience in the mouth at the point of ingestion, rather than a more global rating of a
foods pleasantness. In contrast, the wanting task was designed to provide an implicit
measure derived from a forced choice task whereby stimuli are presented in pairs and
responded to according to which food is most wanted at the present moment. The
speed with which one stimulus is chosen rather than its alternative provides a
quantifiable measure (reaction time) related to relative wanting for that food item.
Reaction time of each decision can convey information about the degree (on a
continuous, interval unit of measurement) to which a chosen stimuli is wanted relative
to an alternative. Furthermore, mean reaction time for each food category can give an
indication of whether motivation is increasing or decreasing independent of the other
categories – in essence an implicit measure of wanting. The forced choice paradigm
carries with it several advantages as an implicit measure of wanting. The visual
nature of the food stimuli means that the dimensions of the categories can be simply
adapted to compare motivation for innumerable combinations of food properties or
specifically selected to complement a particular experimental intervention. Perhaps
most crucially, the forced choice paradigm allows the concurrent measurement of
both implicit and explicit elements of food reward by using dissimilar methodology to
measure each process. Using the computer task, one recent study has revealed
evidence to suggest that the sensory specific satiety phenomenon brought about by
consumption of a standardised, uniform test meal may implicate different roles for
liking and wanting in human appetite (Finlayson et al. In Preparation). We
demonstrated that liking for foods in the task was reduced more for stimuli sharing
properties similar to the test meal (e.g. taste), while wanting had increased for stimuli
with contrasting properties. If a wanting component to sensory specific satiety exists,
it may be activated or enhanced for foods with contrasting sensory properties. We
suggest that the outputs of this computer-based procedure resemble and may
correspond to separate processes of liking and wanting.
The studies reviewed above provide different examples of how cross-contamination
between explicit measures of liking and wanting can be circumvented. Repeated
exposure ratings within a bout of eating (Small et al. 2001) or over a number of days
(Zandstra et al. 2001) can track separations in changes of liking or desire for a target
food. Alternatively, distinguishing general appetite from appetite for a specific food
category (Snoek et al. 2004) may isolate wanting from hunger. Finally, instrumental
measures of motivation (e.g. Saelens & Epstein, 1996; Finlayson et al. 2007) for
target food stimuli may provide operational definitions for the concept of implicit
wanting (incentive salience).
Further conceptual development
Parsing reward into separable components provides a parsimonious theoretical
framework within which to study ingestive behaviour. However, liking and wanting
are not well understood constructs in human appetite. Berridge (1996) advocated a
minimal definition of these processes, opting to focus instead on their role and
function through expanding empirical study. However, as more investigations start to
draw on dual component theories to tackle their research questions, and interpret their
findings, there is a more pressing need to expand and develop theoretical
conceptualisation of these constructs. Four recent theoretical models that – through
the diversity of their approach – may help to enhance a description of liking and
wanting are discussed below.
Homeostatic-hedonic model of hunger
In a discursive paper on the controversy over dietary restriction, Lowe and Levine
(Lowe & Levine, 2005) distinguish physiological or homeostatic hunger (resulting
from nutritional or energy need) from psychological or hedonic hunger (reward-
driven eating when not hungry), which operates “beyond the need to counteract
physiological signals of energy depletion” (p.798). Although reward-driven hunger is
thought to be an important factor in passive overconsumption and obesity (Blundell,
2002; Blundell & Gillett, 2001); the authors argue that both motivations are
intuitively adaptive when viewed in terms of human evolution in a climate of food
scarcity. Just as physiological hunger can strongly motivate food seeking behaviour
in response to declining energy stores, hunger that promotes eating in the absence of
an energy deficit is protective against future energy crises. Therefore, food seeking
behaviour can be motivated by the presence or availability of food (especially when
energy dense or palatable) as well as by genuine homeostatic need. A further
implication is that restriction of wanted foods – even in the absence of an energy
deficit – can cause difficulties in suppressing further food seeking behaviours, similar
to a denial of food that is genuinely required. For example, “200kcal of unflavoured
hot cereal might result in short term satiation whereas 200kcal of chocolate cake
might not satisfy hedonic needs” (p.799). As the authors point out, the boundary
between these two processes is unclear; the point at which hunger becomes reward-
driven rather than necessity-based is subject to a host of inter and intra-individual
factors. The model also provides an interesting perspective on the relationship
between liking and wanting. The authors note that liking is largely a learned
phenomenon, and food preferences that are learned generally retain their motivational
properties independent of homeostatic needs. Hence, palatability seems to be
maintained by oro-sensory factors of reward rather than the energetic delivery of the
ingested nutrients and separate to their perceived energy value. It is argued that food
intake behaviour is more ambiguous than other reward-driven behaviours because the
need to consume food is innate. Therefore, need and reward based want are likely to
interact. However, this assertion can be taken further: feeding behaviour is not alone
in being innate (e.g. mating, parent-child attachment), but it is a habitual behaviour
that is uniquely vital to continued survival. Therefore, the motivating processes
behind feeding are likely to be different to non-essential reward driven behaviours.
The authors emphasise the need to differentiate these behaviours (needing/wanting vs.
liking/wanting) in the study of reward. However, this issue was potentially resolved
in a recent symposium on non-homeostatic behaviour (Corwin & Hajnal, 2005). The
speakers developed a collection of defining principles from which feeding could be
included in an operational definition for all non-homeostatic behaviour. They
proposed that over- or under-eating in response to a homeostatic drive for energy is
normal and will vary within a certain regulatory range (termed ‘compensated non-
homeostatic eating’). When over- or under-eating becomes repetitive or excessive
however, this is termed ‘maladaptive non-homeostatic eating’. This distinction may
represent a framework in which the reward mechanisms central to all non-homeostatic
behaviours can be understood.
Motivating operations in human appetite
Advances in the field of behaviourism have led to an update in the theoretical
principles that can be applied to the study of human appetite. In an interesting review
of the literature (Tapper, 2005) the concept of ‘motivating operations’ (MO) was
discussed. The paper examines the extent to which MOs can account for different
aspects of ingestive behaviour and discusses the conceptual overlap between this and
liking/wanting components of reward. An MO is a motivating event, operation or
stimulus that temporarily affects an organism on two levels. Firstly there is a value-
altering effect: how reinforcing or punishing a relevant stimulus is. Secondly there is
a behaviour-altering effect: whether a relevant behaviour is evoked or abated
(Michael, 1993). In most circumstances, value-altering and behaviour-altering MOs
occur simultaneously and independently. A straight-forward example of an MO is
food deprivation. This physical state acts to increase both the reinforcing
effectiveness of eating (i.e. enhances the experienced pleasure, Cabanac 1989) and
evokes food seeking behaviour (i.e. increases the motivation to consume). Similarly,
a state of excessive-satiety would qualify as an MO; simultaneously reducing the
reinforcing effectiveness of eating and the frequency of behaviours that lead to eating.
Tapper points out that through behavioural analysis, it is possible to observe the
independence of an MO’s separate effects. In food deprivation for example, the
frequency and intensity of food seeking behaviour increases prior to any physical
contact with food. Hence the actual reinforcing value of the food is still unknown and
cannot factor into the observed behaviour. Once oral contact is made with the food,
these seeking behaviours are sometimes observed to intensify (e.g. the ‘appetiser
effect’; see Yeomans, 1996). This temporary increase in responding can be seen as
the impact of the value-altering MO on ingestive behaviour. The author suggests that
the strength of an MO’s value-altering effects may not always be proportional to its
behaviour-altering effects. Moreover, an MO could have value-altering effects in the
absence of behaviour-altering effects. These possibilities are reminiscent of neural
dissociations of liking and wanting. For example, dopamine depletion causes
available food incentives to be ignored – behaviour-altering effect – but does not
reduce the hedonic reaction to food – no value-altering effect (Berridge et al. 1989).
Another feature of the MO model is that MOs can influence the reward value of
neutral stimuli (termed discriminative stimuli, SD) that have since become associated
with the availability of a reinforcer. To adapt the author’s example, through repeated
association with the availability of food, a photograph of a set meal on a restaurant
menu (SD) may elicit some of the behaviours previously linked only to the food itself
(e.g. salivation, stomach rumbling). MOs work to influence the strength of SDs,
therefore, food deprivation could increase responsivity to the food photograph. From
this simple scenario it is possible to envisage a situation where the presence of an MO
could raise the power of an SD above that of the availability of the reinforcer alone.
Consequently, in a state of energy need, the food photograph may elicit a more
intense response than the actual availability of the food and disproportionate to the
reinforcing value of consuming the food. In this way, the MO model could provide a
methodological platform from which ‘irrational wanting’ (Wyvell and Berridge,
2000) might be simulated in humans. The MO model may also help to explain some
non-homeostatic ingestive behaviour. According to the approach, it is possible for
environmental factors to take on motive-like properties through repeated association
with an MO. For example, stimuli paired with hunger (an MO associated with
homeostatic feeding) could eventually acquire similar value and behaviour altering
effects even in the absence of the MO.
The tonic/phasic model of DA system regulation
In a paper written to inform research on craving, Grace (2000) presented a model in
which aspects of drug craving can be explained by distinguishing tonic and phasic
responses in the dopamine system (see figure 2). Phasic DA release refers to the
action potential (‘spike’) dependent release of DA into the synaptic cleft. Phasic DA
release is necessary for the behaviourally relevant actions of DA system activation
including reward signals (e.g. Schultz, 1997; Schultz, 1998). The phasic response
causes a flood of DA to be released into the synapse which under normal conditions is
quickly and efficiently dealt with by re-uptake mechanisms before it can diffuse into
extra-cellular space. Tonic DA release on the other hand refers to the escape of very
small concentrations of DA into extra-cellular space caused by sustained increases in
DA neuron firing or presynaptic stimulation of DA terminals by glutamate. Although
too low to stimulate post-synaptic targets, these levels of DA are sufficient to be
detected by DA terminal auto receptors that regulate DA release from the terminal.
These extra-cellular levels of DA are highly regulated by feedback systems. Through
their effect on the DA auto receptors, increases in tonic DA can cause the inhibition of
phasic DA release (Grace 1991; Grace 1995).
Administration of potent sources of reward (e.g. psychostimulants and alcohol) differ
in their mechanism of DA transmission, but what they have in common is that
repeated administration causes the overflow of DA into the synapse which can result
in the escape of extra-cellular DA. Since increased levels of tonic DA can diminish
phasic release, stimuli that would normally be rewarding will instead produce a
blunted reinforcing signal. This imbalance between phasic and tonic systems is
detected by the organism which attempts to restore the equilibrium. The tonic/phasic
model has interesting implications for the conceptualisation of liking and wanting.
Firstly, it is possible to envisage at the molecular level how a ‘natural’ dissociation
between these processes might occur. Chronic overstimulation of the reward
commodity may cause a gradual increase of tonic DA. This disequilibrium of tonic
and phasic systems could be responded to via psychological drivers such as craving (a
form of wanting). Consequently, food cues become more salient and food seeking
behaviours are initiated. However, due to inhibited phasic response, consumption of
the desired food does not produce the expected rewarding effects. In this way, liking
and wanting become separated by subtle yet significant changes in tonic DA
Determinants of food choice model
Mela (2001; 2006) developed a model to support the assertion that desire for food – a
function of innate and learned liking, internal need state, and environmental cues –
can be viewed as one of the major determinants of food choice in man. Liking is
constructed as the experience or anticipation of pleasure either innately present or
acquired through associative conditioning. In this way, liking is seen as an essential
component contributing to desire. Desire is also influenced by environmental cues
such as situation and appropriateness which may increase or decrease motivation to
eat – possibly through a change in anticipated liking. In this way, it is suggested that
external stimuli can become integrated into a system of cues that trigger motivation
for specific foods under certain conditions. The third factor thought to influence
desire is internal need state. This factor refers to psychological and/or physiological
needs such as hunger, thirst and specific food cravings which may also reinforce the
development of liking. The author uses this model to support the notion that food
liking (as an isolated factor) is not a crucial process in weight gain and obesity, as
liking is often a rather stable characteristic within an individual and less influenced by
weight status (e.g. Cox et al. 1999; Snoek et al. 2004). The implication is that food
intake is driven in part by the stimulation or suppression of desire that incorporates,
yet remains distinguishable from liking. The presence of environmental cues,
strength of internal need state and degree of liking, may all modulate the strength of
desire which in turn determines what (and possibly how much) is consumed next (see
The four models discussed here provide very different perspectives on processes of
liking and wanting in reward. In the homeostatic-hedonic model, wanting is viewed
as an adaptive response to the presence of rewarding food that is closely related to
physical hunger stemming from nutritional need. The motivating operations account
distinguishes value-altering effects from behaviour-altering effects to account for the
variable impact of reward cues on food seeking behaviour and the possible
dissociation of explicit and implicit wanting. The tonic/phasic model describes a
possible mechanism by which wanting can be enhanced yet subsequent reward
diminished through repeated stimulation of the reward commodity. Lastly, the
determinants of food choice model conceptualises liking as a component of wanting,
the strength of which can determine preference. All four models are compatible with
reward driven (non-homeostatic) consumption, and permit a degree of independence
between processes of liking and wanting. Such theoretical accounts may help to
inform our understanding of how separate components of reward should be
Implication for weight gain and obesity
Thanks in part to a better understanding of the interaction of homeostatic and hedonic
processes of appetite and the phenomenon of non-homeostatic consumption, reward is
growing to be viewed as an significant risk factor in weight gain leading to obesity
(Nasser, 2001; Yeomans et al. 2004; Blundell & Finlayson, 2004; Erlanson-
Albertsson, 2005). But what is the evidence that reward may play a role in the
aetiology of obesity? Some studies have implicated individual variability in
sensitivity to reward (STR) – a psychobiological trait linked to the mesolimbic
dopamine pathway – in the development of obesity. Davis et al (Davis et al. 2004)
used the Physical Anhedonia scale (Chapman, 1976) to measure the capacity to
experience reward in normal, overweight and obese patients. The authors found an
inverted U relationship, with overweight subjects scoring higher (more anhedonic)
than obese and normal weight. Similarly, Franken & Muris (2005) demonstrated that
STR (Torrubia et al. 2001) is associated with food cravings and BMI. More recently,
the STR trait has been shown to correlate strongly with activation in relevant regions
of the brain in response to appetising foods relative to bland, aversive and non-food
stimuli (Beaver et al. 2006). Taken together, these findings suggest that a high STR
may characterise those individuals who are at risk of weight gain due to exaggerated
responding to rewarding food cues. Interestingly, because Davis et al. found no
differences between normal weight and patients who had become obese, STR may
represent a risk factor for weight gain (leading to obesity) without necessarily
characterising an obese person. Therefore, STR could be a factor in enhancing
susceptibility to weight gain, rather than a definitive trait of obese people. There is
evidence to support this proposition. For example, PET scans of normal-weight
healthy subjects eating a favourite meal revealed an increase in dopamine release that
correlated with the degree of experienced pleasure. Another study demonstrated that
the availability of the dopamine D2 receptor was decreased in very obese subjects in
proportion to their BMI (Wang et al. 2001). These studies map onto the idea that
there is an optimal level or inverted U relationship between the capacity to experience
reward and dopamine activation (Volkow et al. 1999). Subjects administered the
dopamine agonist methylphenidate found it either pleasant or unpleasant depending
on their DRD2 receptor levels. Those subjects reporting pleasant effects had
significantly lower levels of the dopamine receptor than those who found it
unpleasant. It can therefore be construed that potent stimuli (including palatable
foods) elicit a dampened positive hedonic response through the stimulation of
dopamine activity in people of an anhedonic (low STR) predisposition. One way to
overcome this might simply be to consume more of the rewarding commodity.
However, there is no evidence to date that tests the notion that particularly hedonic
individuals respond aversively to highly palatable foods. It has been suggested that
dopamine activity relating to excessive food consumption might only involve
activation of brain reward circuitry within normal limits, with other psychological
factors exerting stronger effects (Robinson and Berridge, 1998).
Recent investigation of behavioural phenotypes characterised by habitual diet
suggests that the hedonic response to palatable food can influence appetite control via
effects on both food choice and energy intake. Groups of overweight and lean young
males matched for age and the habitual high consumption of fat (high-fat phenotypes)
were compared (Blundell et al. 2005; Le Noury et al. 2002). Although both groups
were eating a diet which theoretically favours a positive energy balance, the
overweight phenotypes consumed greater amounts of the high-fat foods in a test meal
and reported greater feelings of pleasantness, satisfaction, and tastiness for the foods
consumed. One interpretation of these data is that, for at least this group of
overweight people (susceptible to weight gain), they habitually self-select (high-
fat/palatable) foods with a high probability of generating a positive energy balance
(on the basis of their energy density), consume these foods in greater amounts, and
derive greater pleasure from eating these foods. This outcome also demonstrates that
certain high-fat phenotypes – susceptible to weight gain – have a disposition to
perceive foods as being more pleasant than their lean counterparts (who consume the
same habitual high-fat diet). Interestingly, Salbe et al. (2004) found a heightened
hedonic response to sweet and creamy solutions to be associated with subsequent
weight gain in a sample of Pima Indians (a population highly prone to obesity).
Furthermore, Drewnowski and Schwartz (1990) found higher liking for dietary fat in
obese individuals compared to lean. Given this capacity to obtain a high level of
pleasure from foods (and eating), it is not surprising that many obese people show a
tendency to self-select high-palatability foods.
There is also evidence to suggest that obese people may differ in their motivation to
eat. As mentioned previously, Saelens and Epstein (1996) found obese subjects invest
more effort for liked foods on a progressive ratio task, with similar findings when a
normal weight sample of subjects were food-deprived compared to fed (Epstein et al.
2003). Nasser et al. (2005) have also reported the propensity to binge eat to be
associated with persistence of motivation to eat in a post-fed state. Furthermore,
some neuroimaging studies have indicated greater neural activation at specific reward
sites in the brain of weight gaining subjects after consumption of a meal compared to
lean controls (Gautier et al. 2000; 2001). These findings were interpreted as weaker
post-meal satiety signals in overweight subjects. Lastly, Tetley et al. (2006) have
reported greater reactivity in overweight subjects (measured by self-estimated
prospective consumption) when exposed to the sight and smell of a palatable food
compared to lean controls. Cue reactivity was correlated to habitual portion size and
TFEQ disinhibition. Taken together these studies provide some evidence to indicate
that susceptible individuals may be characterised by a diminished ability to resist the
foods they want, possibly due to greater responsiveness and motivation toward cues
associated with tasty foods.
Indeed, individual differences in reward may lead to overconsumption relative to
homeostasis via a number of routes. In terms of processes of liking, it is possible that
some susceptible individuals may experience an exaggerated hedonic response to
palatable foods, in that foods are enjoyed more and therefore eaten in greater amounts
for longer periods of time. In addition to this, some people may have a diminished
ability to experience pleasure from food and therefore greater consumption of
palatable food is promoted to satisfy an optimum level of stimulation. Processes of
wanting may also leave some individuals vulnerable to weight gain through increased
motivation (or reactivity) towards cues signalling the availability of food. Further to
this, a reduced capacity to resist the motivation to eat when replete may also promote
non-homeostatic eating. Finally, some individuals may simply habitually choose
highly-palatable, energy-dense foods that promote overconsumption and lead to a
positive energy balance. This latter possibility can be viewed as (unhealthy/weight
gain promoting) preference and is a behavioural outcome likely to contain elements of
both liking and wanting.
Prompted by findings on the neural structure of food reward in the brain, it is possible
to take a fresh look at the role of reward in human appetite and weight regulation.
Research shows that hedonic processes interact with the homeostatic system of energy
regulation, and that this can influence the organisation of ingestive behaviour, but less
is understood about how liking and wanting components of reward might work
together or separately to modulate appetite. An important consideration is how these
concepts can be operationalised for use in human appetite research. Most previous
research has approached these processes on an explicit (subjective) level, although
there have been some more recent attempts to explore implicit processes through
behavioural measures and brain imaging techniques. There is some interesting
overlap between a number of recent theoretical models and the dual process
modulation of food reward, and these may help to flesh out the framework from
which these processes can be better understood. Processes of liking and wanting may
have independent roles in characterising those susceptible to weight gain and obesity.
Further research into the dissociability of these processes would help to assess the
relative importance of these conceptual components of reward in appetite control and
It should not be forgotten that – strictly speaking – liking and wanting should be seen
to have the logical status of theoretical constructs. Our preferred view is that liking
and wanting should be viewed as intervening variables that help us to understand the
role of hedonics in appetite control. Their existence should not be taken to mean that
these processes are structurally embodied in a neural substrate. Rather, that different
neurochemical pathways can separately influence the events that can be measured
objectively (in animals and humans) and which imply the existence of processes here
referred to as liking and wanting. It is in this sense that we have used the terms in this
This research was supported by Medical Research Council (MRC) Case award
G78/8223 in conjunction with NRC, Lausanne.
Beagley WK & Holley TL 1977 Hypothalamic stimulation facilitates contralateral
visual control of a learned response. Science, 196, 321-323.
Beaver JD, Lawrence AD, van Ditzhuijzen J, Davis MH, Woods A & Calder AJ 2006
Individual differences in reward drive predict neural responses to images of food. J
Neurosci, 26, 5160-5166.
Bellisle F & Le Magnen J 1980 The analysis of human feeding patterns: the edogram.
Appetite, 1, 141-150.
Bellisle F, Lucas F, Amrani R & Le Magnen J 1984 Deprivation, palatability and the
micro-structure of meals in human subjects. Appetite, 5, 85-94.
Berridge KC & Robinson TE 2003 Parsing reward. Trends Neurosci, 26, 507-513.
Berridge KC & Valenstein ES 1991 What psychological process mediates feeding
evoked by electrical stimulation of the lateral hypothalamus? Behav Neurosci, 105, 3-
Berridge KC 2004 Pleasure, unconscious affect and irrational desire. In: Feelings and
Emotions: The Amsterdam Symposium, Eds, Manstead ASR, Frijda NH & Fischer
AH, Cambridge University Press, Cambridge, 43-62.
Berridge KC, Venier IL & Robinson TE 1989 Taste reactivity analysis of 6-
hydroxydopamine-induced aphagia: implications for arousal and anhedonia
hypothesis of dopamine function. Behav Neurosci, 103, 36-45.
Berridge KC 1996 Food reward: brain substrates of wanting and liking. Neurosci
Biobehav Rev, 20, 1-25.
Berridge KC 2000 Measuring hedonic impact in animals and infants: microstructure
of affective taste reactivity patterns. Neurosci Biobehav Rev, 24, 173-198.
Berthoud HR 2004 Neural control of appetite: cross-talk between homeostatic and
non-homeostatic systems. Appetite, 43, 315-317.
Berthoud HR 2006 Homeostatic and non-homeostatic pathways involved in the
control of food intake and energy balance. Obesity, 14, S197-S200.
Blundell JE & Finlayson GS 2004 Is susceptibility to weight gain characterised by
homeostatic or hedonic risk factors for overconsumption? Physiol Behav, 82, 21-25.
Blundell JE & Gillett A 2001 Control of food intake in the obese. Obes Res, 9, S263-
Blundell JE, & Hill AJ 1987 Nutrition, serotonin and appetite: case study in the
evolution of a scientific idea. Appettie, 8, 183-194.
Blundell JE, Stubbs RJ, Golding C, Croden F, Alam R, Whybrow S, LeNoury J &
Lawton CL 2005 Resistance and susceptibility to weight gain: individual variability in
response to a high-fat diet. Physiol Behav, 86, 614-622.
Blundell JE 2002 A psychobiological system approach to appetite and weight control.
In: Eating Disorders and Obesity, Ed, Brownell KD, Guildford, New York, 49.
Bobroff EM & Kissileff HR 1986 Effects of changes in palatability on food intake
and the cumulative food intake curve in man. Appetite, 7, 85-96.
Booth DA & Blair AJ 1988 Objective factors in the appeal of a brand during use by
the individual consumer. In: Food acceptability, Ed, Thomson DMH, Elsevier
Applied Science, London, 329-346.
Booth DA 1987a Cognitive experimental psychology of appetite. In: Eating habits,
Eds, Boakes RA, Burton MJ & Popplewell DA, Wiley, Chichester, 175-209.
Booth DA 1987b Objective measurement of determinants of food acceptance:
sensory, physiological and psychosocial. In: Food acceptance and nutrition, Eds,
Solms J, Booth DA, Pangbom RM & Raunhardt O, Academic Press, London, 1-27.
Booth DA 1990 How not to think about immediate dietary and postingestional
influences on appetites and satieties. Appetite, 14, 171-179.
Cabanac M 1979 Sensory pleasure. Q Rev Biol, 54, 1-29.
Cabanac M 1989 Maximization of pleasure, the answer to a conflict of motivations. C
R Acad Sci III, 309, 397-402.
Carr KD & Wolinsky TD 1993 Chronic food restriction and weight loss produce
opioid facilitation of perifornical hypothalamic self-stimulation. Brain Res, 607, 141-
Chapman LJ, Chapman JP & Raulin ML 1976 Scales for physical and social
anhedonia. J Abnorm Psychol, 85, 374-382.
Corwin RL & Hajnal A 2005 Too much of a good thing: neurobiology of non-
homeostatic eating and drug abuse. Physiol Behav, 86, 5-8.
Cota D, Marsicano G, Tschop M, Grubler Y, Flachskamm C, Schubert M, Auer D,
Yassouridis A, Thone-Reineke C, Ortmann S, Tomassoni F, Cervino C, Nisoli E,
Linthorst AC, Pasquali R, Lutz B, Stalla GK & Pagotto U 2003 The endogenous
cannabinoid system affects energy balance via central orexigenic drive and peripheral
lipogenesis. J Clin Invest, 112, 423-431.
Cox DN, Perry L, Moore PB, Vallis L & Mela DJ 1999 Sensory and hedonic
associations with macronutrient and energy intakes of lean and obese consumers. Int J
Obes, 23, 403-410.
Davis C, Strachan S, & Berkson M 2004 Sensitivity to reward: implications for
overeating and overweight. Appetite, 42, 131-138.
De Castro JM, Bellisle F & Dalix A-M 2000a Palatability and intake relationships in
free-living humans: characterization and independence of influence in the French.
Physiol Behav, 68, 271-277.
De Castro JM, Bellisle F, Dalix A-M & Pearcey SM 2000b Palatability and intake
relationships in free-living humans: characterization and independence of influence in
North Americans. Physiol Behav, 70, 343-350.
De Graaf C, De Jong LS & Lambers AC 1999 Platability affects satiation but not
satiety. Physiol Behav, 66, 681-688.
Di Marzo V, Goparaju SK, Wang L, Liu J, Batkai S, Jarai Z, Fezza F, Miura GI,
Palmiter RD, Sugiura T & Kunos G 2001 Leptin-regulated endocannabinoids are
involved in maintaining food intake. Nature, 410, 822-825.
Drewnowski A & Schwartz M 1990 Invisible fats: sensory assessment of sugar/fat
mixtures. Appetite, 14, 203-217.
Drewnowski A 1998 Energy density, palatability, and satiety: implications for weight
control. Nutr Rev, 56, 347-353.
Epstein LH, Truesdale R, Wojcik A, Paluch RA & Raynor HA (2003) Effects of
deprivation on hedonics and reinforcing value of food. PhysiolBehav, 78, 221-227.
Erlanson-Albertsson C 2005 How palatable food disrupts appetite regulation. Basic
Clin Pharmacol Toxicol, 97, 61-73.
Finlayson G, King N & Blundell J In Press Is it possible to dissociate ‘liking’ and
‘wanting’ for foods in humans: a novel experimental procedure. Physiol Behav.
Flier JS 2004 Obesity wars: molecular progress confronts and expanding epidemic.
Cell, 116, 337-350.
Franken IH & Muris P 2005 Individual differences in reward sensitivity are related to
food craving and relative body weight in healthy women. Appetite, 45, 198-201.
Gautier JF, DelParigi A, Chen K et al. 2000 Differential brain responses to satiation in
obese and lean men. Obes Res, 9, 676-684.
Gautier JF, DelParigi A, Salbe AD et al. 2001 Effect of satiation on brain activity in
obese and lean women. Diabetes, 49, 838-846.
Giraudo SQ, Grace MK, Billington CJ & Levine AS 1999 Differential effects of
neuropeptides Y and the mu-agonist DAMGO on ‘palatability’ vs. ‘energy’. Brain
Res, 10, 160-163.
Grace AA 2000 The tonic/phasic model of dopamine system regulation and its
implications for understanding alcohol and psychostimulant craving. Addiction, 95,
Grace AA 1991 Phasic versus tonic dopamine release and the modulation of
dopamine system responsivity: a hypothesis for the etiology of schizophrenia.
Neuroscience, 41, 1-24.
Grace AA 1995 The tonic/phasic model of dopamine system regulation: its relevance
for understanding how stimulant abuse can alter basal ganglia function. Drug Alcohol
Depend, 37, 111-129.
Grill HJ & Norgren R 1978 The taste reactivity test. I. Mimetic responses to gustatory
stimuli in neurologically normal rats. Brain Res, 143, 263-279.
Guy-Grand B, Lehnert V, Doassans M & Bellisle F 1994 Type of test-meal affects
palatability and eating style in humans. Appetite, 22, 125-134.
Hayward MD, Pintar JE & Low MJ 2002 Selective reward deficit in mice lacking
beta-endorphin and enkephalin. J Neurosci, 22, 8251-8258.
Hellemann U & Tuorila H 1991 Pleasantness ratings and consumption of open
sandwiches with varying NaCl and acid contents. Appetite. 17, 229-238.
Hill AJ, Magson LD & Blundell JE 1984 Hunger and palatability: tracking ratings of
subjective experience before, during and after the consumption of preferred and less
preferred food. Appetite, 5, 361-371.
Ikeda J, Amy NK, Ernsberger P, Gaesser GA, Berg FM, Clark CA, Parham ES &
Peters P 2005 The National Weight Control Registry: a critique. J Nutr Educ Behav,
Johnson J & Vickers Z 1993 Effects of flavour and macronutrient composition of
food servings on liking, hunger and subsequent intake. Appetite, 21, 25-39.
Kirkham TC & Williams G 2001 Endogenous cannabinoids and appetite. Nutr Res
Rev, 14, 65-86.
LeNoury JC, Lawton C & Blundell JE 2002 Food choice and hedonic responses:
differences between overweight and lean high fat phenotypes. Int J Obes, 26, S125.
Levine AS, Kotz CM, Gosnell BA 2003 Sugars and fats: the neurobiology of
preference. J Nutr, 133, S831-834.
Looy H, Callaghan S, Weingarten HP 1992 Hedonic response of sucrose likers and
dislikers to other gustatory stimuli. Physiol Behav, 52, 219-225.
Lowe MR & Levine AS 2005 Eating motives and the controversy over dieting: eating
less than needed versus less than wanted. Obes Res, 13, 797-806.
Lucas F & Bellisle F 1987 The measurement of food preferences in humans: do taste-
and-spit tests predict consumption? Physiol Behav, 39, 739-743.
Mela DJ 2001 Determinants of food choice: re;ationships with obesity and weight
control. Obes Res, 9, S249-255.
Mela DJ 2006 Eating for pleasure or just wanting to eat? Reconsidering sensory
hedonic responses as a driver of obesity. Appetite, 47, 10-17.
Michael J 1993b Establishing operations. Behav Anal, 16, 191-206.
Monneuse MO, Bellisle F & Louis-Sylverstre J 1991 Responses to an intense
sweetener in humans: immediate preference and delayed effects on intake. Physiology
& Behavior, 49, 325-330.
Mook DG & Votaw MC 1992 How important is hedonism? Reasons given by college
students for ending a meal. Appetite, 18, 69-75.
Nasser J 2001 Taste, food intake and obesity. Obes Rev, 2, 213-218.
Nasser JA, Geliebter A & Pi-Sunyer FX 2005 Persistence of food reinforcement after
a caloric preload in women is correlated with binge eating score and hunger in the
fasted state. Appetite, 44, 329.
Pelchat ML 2002 Of human bondage: food craving, obsession, compulsion, and
addiction. Physiol Behav, 76, 347-352.
Perez C, Dalix AM, Guy-Grand B & Bellisle F 1994 Human responses to five
concentrations of sucrose in a dairy product: Immediate and delayed palatability
effects. Appetite, 23, 165-178.
Peters JC, Wyatt HR, Donahoo WT & Hill JO 2002 From instinct to intellect: the
challenge of maintaining healthy weight in the modern world. Obes Rev, 3, 69-74.
Robinson TE & Berridge KC 2000 The psychology and neurobiology of addiction: an
incentive-sensitization view. Addiction, 95, S91-117.
Robinson TE & Berridge KC 2003 The neural basis of drug craving: an incentive-
sensitization theory of addiction. Brain Res Rev, 18, 247-291.
Rogers PJ & Blundell JE 1990 Umami and appetite: effects of monosodium glutamate
on hunger and food intake in human subjects. Physiol Behav, 48, 801-804.
Rogers PJ & Blundell JE 1991 Mechanisms of diet selection: the translation of needs
into behaviour. Proc Nutr Soc, 50, 65-70.
Saelens BE & Epstein LH 1996 Reinforcing value of food in obese and non-obese
women. Appetite, 27, 41-50.
Salbe AD, DelParigi A, Pratley RE, Drewnowski A & Tataranni PA 2004 Taste
preferences and body weight changes in an obesity-prone population. Am J CLin Nutr,
Saper CB, Chou CT & Elemquist JK 2002 The need to feed: homeostaticand hedonic
control of eating. Neuron, 36, 199-211.
Schultz W 1997 Dopamine neurons and their role in reward mechanisms. Curr Opin
Neurobiol, 7, 191-197.
Schultz W 1998 The phasic reward signal of dopamine neurons. Adv Pharmacol, 42,
Schwartz MW, Woods SC, Porte D, Seeley RJ & Baskin DG 2000 Central nervous
system control of food intake. Nature, 404, 661-671.
Small DM, Zatorre RJ, Dagher A, Evans AC & Jones-Gotman M 2001 Changes in
brain activity related to eating chocolate: from pleasure to aversion. Brain, 124, 1720-
Snoek HM, Huntjens L, Van Gemert LJ, De Graaf C & Weenen H 2004 Sensory-
specific satiety in obese and normal-weight women. Am J Clin Nutr, 80, 823-831.
Spiegel TA, Shrager E & Stellar E 1989 Responses of lean and obese subjects to
preloads, deprivation, and palatability. Appetite, 13, 45-69.
Stanley S, Wynne K, McGowan B & Bloom S 2005 Hormonal regulation of food
intake. Physiol Rev, 85, 1131-1158.
Tapper K 2005 Motivating operations in appetite research. Appetite, 45, 95-107.
Tetley AC, Brunstrom JM & Griffiths P (2006) Individual differences in food-cue
reactivity. Appetite, 47, 278.
Torrubia R, Avila C, Molto J, Caseras X 2001 The sensitivity to punishment and
sensitivity to reward questionnaire (SPSRQ) as a measure of Gray’s anxiety and
impulsivity dimensions. Pers Individ Dif. 31, 837-862.
Verty AN, Geller F, Dempfle A, Schauble N, Friedel S, Lichtner P, Fontela-Horro F,
Wudy S, Hagemann S, Gortner L, Huse K, Remschmidt H, Bettecken T, Meitinger T,
Schafer H, Hebebrand J & Hinney A 2004 Ghrelin receptor gene: identification of
several sequence variants in extremely obese children and adolescents, healthy
normal-weight and underweight students, and children with short normal stature. J
Clin Endocrinol Metab, 89, 157-162.
Volkow ND, Wang GJ, Fowler JS, Logan J, Gatley SJ, Gifford A et al. 1999
Prediction of reinforcing responses to psychostimulants in humans by brain dopamine
D2 receptor levels. Am J Psychiatr, 156, 1440-1443.
Wang GJ, Volkow ND, Logan J, Pappas NR, Wong CT, Zhu W et al. 2001 Brain
dopamine and obesity. Lancet, 357, 354-357.
Warwick ZS, Hall WG, Pappas TN & Schiffman SS 1993 Taste and smell sensations
enhance the satiating effect of both a high-carbohydrate and a high-fat meal in
humans. Physiol Behav, 53, 553-563.
Welch CC, Kim EM, Grace MK, Billington CJ & Levine AS 1996 Palatability-
induced hyperphagia increases hypothalamic dynorphin peptide and mRNA levels.
Brain Res, 721, 126-131.
Winkielman P, Berridge KC & Wilbarger JL 2005 Unconscious affective reactions to
masked happy versus angry faces influence consumption behavior and judgements of
value. Pers Soc Psychol Bull, 31, 121-135.
Wyvell CL & Berridge KC 2000 Intra-accumbens amphetamine increases the
conditioned incentive salience of sucrose reward: Enhancement of reward ‘wanting’
without enhanced ‘liking’ or response reinforcement. J Neurosci, 20, 8122-8130.
Yamamoto T, Sako N & Maeda S 2000 Effects of taste stimulation on beta-endorphin
levels in rat cerebrospinal fluid and plasma. Physiol Behav, 69, 345-350.
Yeomans MR & Gray RW 2002 Opioid peptides and the control of human ingestive
behaviour. Neurosci Biobehav Rev, 26, 713-728.
Yeomans MR & Symes T 1999 Individual differences in the use of pleasantness and
palatability ratings. Appetite, 32, 383-394.
Yeomans MR 1996 Palatability and the micro-structure of feeding in humans: the
appetizer effect. Appetite, 27, 119-133.
Yeomans MR 1998 Taste, palatability and the control of appetite. Proc Nutr Soc, 57,
37 Download full-text
Yeomans MR, Blundell JE & Leshem M 2004 Palatability: response to nutritional
need or need-free stimulation of appetite? Br J Nutr, 92, S3-14.
Yeomans MR, Gray RW, Mitchell CJ & True S 1997 Independent effects of
palatability and within-meal pauses on intake and appetite ratings in human
volunteers. Appetite, 29, 61-76.
Yeomans MR, Lee MD, Gray RW & French SJ 2001 Effects of test-meal palatability
on compensatory eating following disguised fat and carbohydrate preloads. Int J Obes
Related Metab Disord, 25, 1215-1224.
Yeomans MR, Tovey HM, Tinley EM & Haynes CJ 2004 Effects of manipulated
palatability on appetite depend on restraint and disinhibition scores from the Three-
Factor Eating Questionnaire. Int J Obes Relat Metab Disord, 28, 144-51.
Zandstra EH, De Graaf C, Mela DJ & Van Staveren WA 2000 Short- and long-term
effects of changes in pleasantness on food intake. Appetite, 34, 253-260.
Zandstra EH, De Graaf C, van Trijp HCM & van Staveren WA 1999 Laboratory
hedonic ratings as predictors of consumption. Food Quality and Preference, 10, 411-
Zandstra EH, Weegels MF, van Spronsen AA, Klerk M 2004 Scoring or Boring?
Predicting boredom through repeated in-home consumption. Food Qual Pref, 15, 549-