The psychobiology of comfort eating

442
Review article
The psychobiology of comfort eating: implications for
neuropharmacological interventions
E. Leigh Gibson
Comfort eating, that is eating induced by negative affect,
has been a core theme of explanations for overeating and
obesity. Psychobiological explanations and processes
underlying comfort eating are examined, as well as its
prevalence in clinical and nonclinical populations, to
consider who may be susceptible, whether certain foods
are comforting, and what the implications for treatment
may be. Comfort eating may occur in a substantial minority,
particularly in women and the obese. Human and animal
theories and models of emotional or stress-induced eating
show some convergence, and may incorporate genetic
predispositions such as impulsivity and reward sensitivity,
associated with dopamine dysregulation underlying
incentive salience. Comfort eaters show vulnerability to
depression, emotional dysregulation and a need to escape
negative affect and rumination. During negative affect, they
preferentially consume sweet, fatty, energy-dense food,
which may confer protection against stress, evidenced by
suppression of the hypothalamic–pituitary–adrenal axis
response, although activation of the hypothalamic–
pituitary–adrenal axis may itself drive appetite for
these palatable foods, and the risk of weight gain is
increased. Benefits to mood may be transient, but
perhaps sufficient to encourage repeated attempts to
prolong mood improvement or distract from negative
rumination. Cognitive behavioural treatments may be
useful, but reliable drug therapy awaits further
pharmacogenomic developments. Behavioural
c 2012 Wolters Kluwer Health |
Pharmacology 23:442–460 Lippincott Williams & Wilkins.
Introduction
vascular system through unhealthy food choices, or, by
contrast, whether comfort eating might actually provide
some benefits to health and well-being. Evidence for a role
for particular neurotransmitter and neurohormonal systems
is considered and thus whether neuropharmacological
interventions may be feasible or even advisable.
This review considers the nature of comfort eating, and
describes psychobiological explanations for it, in addition to
discussing evidence for who might be susceptible and why,
and what sustains comfort eating in those individuals.
Although it is recognized that comfort eating may overlap
with binge eating, this review is not specifically concerned
with binge eating, which requires clear evidence of
excessive and frequent intake; rather, comfort eating is
considered a phenomenon that may be common in
nonclinical populations, and may not necessarily involve
episodes of excessive consumption. The relevance of
comfort eating to public health is gauged by estimating
its prevalence, particularly in nonclinical populations, from
data available in the literature. Major theories are discussed
that set comfort eating in the context of normal and
abnormal eating behaviour and help to understand both
how it may arise, including its relation to personality
characteristics, as well as the implications for therapy. A key
aspect of theories of comfort eating is understanding
whether certain foods or food groups may be chosen by
comfort eaters and why: theories are examined that
attempt to explain why negative affect and stress may
elicit eating in individuals with particular traits. The review
also addresses whether comfort eating may be harmful to
health, such as by increasing the risk of developing obesity
or other disordered eating, or by damaging the cardioc 2012 Wolters Kluwer Health | Lippincott Williams & Wilkins
0955-8810 Behavioural Pharmacology 2012, 23:442–460
Keywords: depression, dopamine D2 receptors, emotional eating,
glucocorticoid hormones, hypothalamic–pituitary–adrenal axis, mood,
obesity, reward sensitivity, stress
Department of Psychology, Whitelands College, University of Roehampton,
London, UK
Correspondence to Edward Leigh Gibson, BSc, PhD, Department of Psychology,
Whitelands College, University of Roehampton, Holybourne Avenue, London
SW15 4JD, UK
E-mail: [email protected]
Received 15 May 2012 Accepted as revised 3 July 2012
Defining ‘comfort eating’
In the academic literature, the normal understanding of
comfort eating is eating to relieve negative emotions or
affect (typically depression, anxiety or anger), which is
synonymous with ‘emotional eating’ (at least where the
definition is restricted to negative emotions), the latter
being the more commonly used term in such literature
(Ganley, 1989). Another term with similar meaning is
‘stress eating’, that is eating induced by stress; indeed,
consideration of ‘who eats more when stressed and why’ is
at the heart of understanding comfort eating. Even so, this
implies that individuals who eat more when stressed do so
because of some relief that eating provides, and clearly this
needs to be evidenced. This review will focus on these
issues, and the evidence for psychobiological mechanisms.
An implicit assumption of stress-induced eating is that
the stress results in the sort of negative affect that can
elicit eating, and thus is not severe enough to suppress
appetite. Whereas one can operationally define ‘comfort
DOI: 10.1097/FBP.0b013e328357bd4e
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Psychobiology of comfort eating Gibson 443
eaters’ as those who respond to stress by eating more (at
least of some foods), in practice, for ethical reasons, direct
evidence for this is mainly restricted to comparative
studies, usually in rats (Hagan et al., 2002; Wardle and
Gibson, 2002). In human studies, the tendency to
‘comfort eat’, or be an emotional eater, is typically
measured by self-report questionnaires: this can be as
simple as asking whether one eats more, less or the same
when stressed (Oliver and Wardle, 1999), but more
commonly is based on scales created from several items
referring to eating more when upset, etc. There are
several such instruments in common use, such as the
Dutch Eating Behaviour Questionnaire (DEBQ; Van
Strien et al., 1986), the revised Three Factor Eating
Questionnaire using 18-items (derived from an administered 21-items), versions 1 (TFEQ-R18; Karlsson et al.,
2000) and 2 (TFEQ-R18v2; Cappelleri et al., 2009), and
the Emotional Eating Scale (EES; Arnow et al., 1995), the
latter being developed principally for clinical populations,
such as binge eaters. Recently, Nolan et al. (2010)
validated the Emotional Appetite Questionnaire
(EMAQ; Geliebter and Aversa, 2003), which includes
assessment of the role of positive emotions and emotional
situations; however, eating in response to positive
emotions is beyond the definition of comfort eating. It
is beyond the scope of this review to discuss these
instruments in detail, but aspects of the scales will be
considered where important for interpretation of findings.
Although the usual inference is that ‘comfort foods’ are
those foods that, when eaten, can somehow provide relief
from emotional upset, it should be noted that there is a
lay concept of comfort foods that has slightly different
connotations: this is the idea that comfort foods are foods
highly familiar from one’s childhood, also known as
‘nursery foods’ or perhaps ‘home (or mother’s) cooking’.
These foods are often associated with special occasions,
and thus encoded in, and capable of recalling, happy
memories. The psychological impact of such foods may
be more related to security, in the sense of memories of
secure attachment (or a longing for it) or at least social
bonding, rather than necessarily relieving negative affect.
This implies that, for some individuals, comfort eating
may be related to facets of their ‘attachment style’, which
also has implications for neurohormonal influences
(Marazziti and Catena Dell’osso, 2008). Indeed, there is
recent evidence that ‘anxious attachment’ is associated
with disinhibited eating and BMI (Wilkinson et al., 2010),
although an emotional eating component may not be
directly involved (Phillips et al., 2012). In fact, familiarity
itself might be a facet of foods that is actively sought
during anxiety. As this notion of comfort ‘nursery’ or
‘home’ foods is not specifically excluded from definitions
of comfort food, we cannot be certain to what extent it
contributes to comfort eating. Moreover, Wansink et al.
(2003) have pointed out that sex differences in comfort
foods, such as men preferring hot, savoury meal-related
foods, whereas women preferring sweet snack foods,
might reflect differences in culinary upbringing, with
women more likely to be the food provider. These and
other cultural differences might underlie the observation
that comfort eating in men is more likely to be motivated
by positive emotions, whereas women invariably comfort
eat in response to negative emotions (Dubé et al., 2005).
Prevalence of comfort eating
To gauge the importance of comfort eating, particularly if
treatment is to be considered, some estimate of prevalence
is required. This section will summarize the indicators of
prevalence available in the literature: nonexperimental
survey studies are considered with moderate to large
sample sizes, primarily of noneating disordered populations,
in which adult versions of standardized emotional eating
questionnaires have been used (twin studies are reported in
the last section). Smaller ‘naturalistic’ studies – typically in
student populations – have addressed how eating is affected
by stress in nonclinical populations, and show that between
one-third and two-thirds of those surveyed report eating
less when stressed, whereas about a quarter to a half report
eating more when stressed, that is comfort eating; a small
minority report no change (reviewed by Gibson, 2006).
However, these surveys did not use standardized scales for
measuring emotional eating; thus, estimates of prevalence
are difficult to compare between study samples. Nevertheless, these findings are in notable contrast to the widely
accepted view that the majority of clinical populations of
disordered eating patients, obese or otherwise, show some
form of emotional or comfort eating or at least affect-driven
disturbance of eating (Wardle, 1987; Ganley, 1989; Vögele
and Gibson, 2010).
Even for studies using questionnaires to measure emotional
eating, it is difficult to determine prevalence reliably,
because overall means, and correlations, are typically
reported, rather than the use of behaviourally validated
cut-off scores or other distribution indicators. Moreover, the
distribution of scale scores is likely to vary depending on,
for example, whether the sample studied is a mainly
normal-weight student population or an older population
including a substantial proportion of overweight and obese
participants. Nevertheless, there have now been sufficient
moderate to large cross-sectional surveys, including some of
nationally representative samples, to make a summary of
prevalence, estimated from the mean scores of emotional
eating questionnaires worthwhile: that is, with quite large
normally distributed samples, one can draw conclusions
about the distribution of scores on the scale from the
means. These studies are summarized in Table 1: to allow
means to be compared across different instruments, all
means are given as percentage scores (i.e. scale ranges were
converted into 0–100, if not already reported as such). It is
clear from Table 1 that almost all of the means are less than
50%, many substantially so, except for obese groups
(Karlsson et al., 2000) and the bulimic group from Wardle
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
444 Behavioural Pharmacology 2012, Vol 23 No 5&6
Table 1
Comparison of mean % emotional eating scores from survey populations
Reference (Country)
Design
Emotional
eating measure
Van Strien et al. (1986)
(the Netherlands)
Survey, cross-sectional
DEBQ
Wardle (1987) (UK)
Survey, cross-sectional
DEBQ
Wardle et al. (1992) (UK)
Survey, cross-sectional
DEBQ
Lluch et al. (2000) (France) Survey, cross-sectional
DEBQ
Karlsson et al. (2000)
(Sweden)
De Lauzon et al. (2004)
(France)
Survey, cross-sectional
TFEQ-R18
Survey, cross-section of
cohort
TFEQ-R18
Snoek et al. (2007) (the
Netherlands)
Survey, cross-sectional
DEBQ
O’Connor et al. (2008)
(UK)
Survey, cross-sectional
DEBQ
Elfhag et al. (2008)
(Sweden)
Survey, cross-sectional
DEBQ
Cappelleri et al. (2009)
(USA and Canada)
Clinical trial (baseline) and
web-based survey
TFEQ-R18v2
Anschutz et al. (2009) (the Survey, cross-sectional
Netherlands)
Van Strien et al. (2009)
Survey, cross-sectional
(the Netherlands)
DEBQ
DEBQ
Nguyen-Rodriguez et al.
(2009) (USA)
Survey, cross-sectional
DEBQ
Anglé et al. (2009)
(Finland)
Clinical trial (baseline)
TFEQ-R18
Nolan et al. (2010) (USA)
Survey, cross-sectional
Konttinen et al. (2010)
(Finland)
Survey, cross-section of
cohort
DEBQ
EMAQ-NE
EMAQ-NS
DEBQ
EMAQ-NE
EMAQ-NS
TFEQ-R18
Sample (N; adult unless specified)
602 women
449 men (including 131 overweight/
obese)
102 female students
86 male students
107 ‘weight watchers’
61 bulimic
33 anorectic
439 girls (11–18 year olds)
402 boys (11–18 year olds)
381 women
379 men
290 girls (11–18 year olds)
270 boys (11–18 year olds)
2603 obese women
1774 obese men
284 women (30–67 year olds)
236 men (30–67 year olds)
163 girls/young women (14–27 year
olds)
171 boys/young men (14–27 year olds)
4581 girls (11–16 year olds)
4430 boys (11–16 year olds)
(including 18.1% overweight, 1.7%
obese)
229 women
193 men
(including 55 obese)
1795 women
1471 men
731 girls (11–12 year olds)
710 boys (11–12 year olds)
Trial (1660 obese; 68 nonobese)
1422 women
306 men
Web survey (817 obese; 458 nonobese)
503 women
772 men
475 female students (including 28
underweight, 6 obese)
845 womena
525 mena (including 717 normal weight,
625 overweight/obese)
372 girls (11–15 year olds)
132 boys (11–15 year olds)
2997 women (17–20 year olds)
181 underweight
2074 normal weight
469 overweight
154 obese
171 female students
59 male students (including 17 obese)
2035 women
1679 men (including 740 obese)
Result (mean
% score)
Findings on EE by sex and BMI
26.5
18.0
F > Mc
41.3
31.0
51.3
70.8
31.3
27.8
23.5
32.3
18.0
31.5
23.3
63.0b
–b
43.0
22.0
46.0
26.0
F>M
Bulimics > overweight > normal
weight
F>M
Unrelated to BMI
F > M (adult and child)
BMI positively related in adult F
only
F>M
Unrelated to BMI
F > M (adult and child/youth)
24.3
21.3
F>M
BMI negatively related
35.3b
–b
Not reported
22.5
15.0
15.0
15.0
46.7
30.0
34.3
26.7
Not reported
41.5
BMI positively related
31.5
40.3
F>M
BMI positively related
21.8
20.0
F=M
20.4% F, 16.5% M are above
EE cut-offd
BMI positively related from
underweight to obese
22.5
30.3
37.6
46.6
39.3
47.0
39.1
27.0
44.9
36.3
38.2
23.1
F>M
BMI positively related only for
web survey
EMAQ, but not DEBQ,
positively related to BMI
F>M
BMI positively related
DEBQ, Dutch Eating Behaviour Questionnaire; EE, emotional eating; EMAQ, Emotional Appetite Questionnaire (NE, negative emotions; NS, negative situations);
F, female; M, male; TFEQ-R18, revised 18-item version 1 of Three Factor Eating Questionnaire; TFEQ-R18v2, revised 18-item version 2 of Three Factor Eating
Questionnaire.
a
Means not reported separately for men and women.
b
Derived from the distribution of scores for the entire sample (men and women).
c
Significance of the mean difference between sexes was not reported.
d
From the DEBQ manual (Van Strien, 2005, cited in Nguyen-Rodriguez et al., 2009).
(1987). In the majority of studies, it is clear that the central
tendency is endorsing that emotional eating is occurring
less than ‘sometimes’. However, it is also apparent that the
majority of studies find emotional eating to be more likely
in women than men: the exception is no sex difference in
the American study of adolescent school children (NguyenRodriguez et al., 2009); in general, emotional eating is less
likely in adolescents than in adults (Table 1). This might
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Psychobiology of comfort eating Gibson 445
also explain why there is a lack of relationship (or even an
inverse relationship, Snoek et al., 2007) between emotional
eating and BMI in these young samples, although a lower
range of BMI may also contribute: similarly, the constrained
BMI range may underlie there being no relation between
emotional eating and BMI among obese Swedes (Karlsson
et al., 2000). By contrast, in most of the adult samples, BMI
is consistently positively related to emotional eating
(Table 1).
Another question that can be addressed from these surveys
is whether there is any trend over time for the prevalence
of emotional eating. Usefully, there have now been two
studies of adult Dutch samples representative of the
general population in the Netherlands that measured
emotional eating (DEBQ) about 25 years apart (Van Strien
et al., 1986, 2009). Intriguingly, for the more recent study
published in 2009, average emotional eating was substantially higher (mean = 36%) than in the first study (data
gathered in 1983; mean = 23%). Even so, the majority of
both populations were low to moderate emotional eaters;
yet, the ratings have clearly increased among the recent
Dutch sample – as has obesity, in line with national trends.
The two other scales of the DEBQ (Restrained and
External Eating) also increased, but to lesser extents. Van
Strien et al. (2009) speculated that this could reflect a
disconnection from normal eating contexts and consequences brought on by the increasing dominance of screen
media time among children and adolescents, promoting
‘mindless eating’. However, it also seems plausible that it
could reflect an increased self-awareness of eating behaviour and attitudes, engendered by increasing media
coverage of diet-related health concerns.
Thus, it seems probable that individuals who report a
consistent profile of comfort eating are in the minority of
the general population, even if quite a large one. Indeed,
as far as the response to stress is concerned, suppression
of appetite, as opposed to comfort eating, tends to be
the more commonly reported outcome, depending on the
degree and the nature of the stress, as well as the
individual’s ability to cope with stress (Gibson, 2006);
these issues are discussed later.
This tendency for emotional eating to be distributed as low
to moderate in these surveys has implications for experimental tests of the predictive power of the questionnaires.
Given that these questionnaires use response formats that
assess the frequency of occurrence of a situation/behaviour
(DEBQ; EMAQ) or the likelihood of a situation/behaviour
being true (TFEQ), median splitting of the scale scores
may not represent functional comfort eaters versus
noncomfort eaters, depending on how close the median is
to the midpoint of the scale.
A recent study specifically addressed this issue of the
predictive validity of the emotional eating measurement:
to test the predictive value of scores on the DEBQ scale for
emotional eating Van Strien et al. (2012) recently measured
snack intake, in female students, after negative or
neutral mood induction (film, between participants, data
from Anschutz et al., 2008), or after stress (Trier Social
Stress test; within participants) or no stress. One strength
of the study is its ability to define low and high emotional
eating groups using cut-off scores of the 20th and 80th
percentiles of norms for that population. In both experiments, emotional eating was a significant moderator of the
relationship between distress and snack intake, as expected, that is only high emotional eaters ate more when
distressed, whereas low emotional eaters ate less; however,
this interaction could not be found when relying only on
a median split for the entire sample (Study 1, n = 124).
Thus, the effect relies on having sufficiently extreme
samples of both low and high emotional eaters: the lack of
such variability in emotional eating may explain previous
failures to show this effect. It also suggests that the DEBQ
only robustly predicts overeating induced by negative affect
for individuals scoring near the highest quintile.
Theories of eating styles relevant to comfort
eating
To understand current notions of comfort eating and its
aetiology, it is helpful to consider briefly some relevant
theoretical descriptions of eating styles and their
proposed relationships with risks for overeating. Where
possible, the effects of stress or negative affect on these
eating styles are discussed.
Psychosomatic and externality theories
At a time when obesity was rare, and typically treated
from a psychiatric or an analytic perspective, two theories
of obesity and overeating emerged, with some similarities.
Psychosomatic theory essentially proposed that the major
cause of obesity was indeed hyperphagia because of
emotional eating, as a means to reduce anxiety (Kaplan
et al., 1957). Subsequently, Bruch (1961) suggested that
a possible cause of emotional eating was a confusion
between emotions, particularly negative affect, and
sensations related to hunger, and that this might be
acquired during childhood perhaps because of particular
parenting strategies. This then begs the question why
some individuals might be reinforced in their drive to eat
more by emotional calming than by satiation of hunger.
This issue of individual variation in susceptibility is a
recurring theme of this review.
The apparent insensitivity of the obese to physiological
cues underlying hunger, for which there was early
experimental evidence (Stunkard, 1959), is also a key
facet of the ‘externality’ theory of obesity, in
which Schachter (1968) proposed that overeating, in at
least some obese individuals, reflected an over-responsiveness to external cues related to food, and reviewed
evidence in support of his hypothesis (e.g. eating in
relation to sensory qualities, portion size, availability, time
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
446 Behavioural Pharmacology 2012, Vol 23 No 5&6
of day), contrasting with a tendency to ignore or be
insensitive to internal cues related to physiological need.
Indeed, Schachter (1968) was explicitly following up these
ideas from Bruch (1961) and Stunkard (1959). In terms of
comfort eating, Schachter (1968) reported the results of an
experiment (very much of its time) in which fear was
induced by the threat of painful electric shocks: this
suppressed snack intake in a ‘taste test’ in healthy controls
but not in obese individuals, which could be interpreted as
early experimental evidence for comfort eating. This effect
mirrored that found when comparing the amount eaten in
fed and fasted states, in the same study; that is, again
intake in the obese did not vary, implying insensitivity to
physiological need; disappointingly, Schachter (1968) did
not report the three-way ‘obesity by deprivation by fear’
interaction, which might have been informative for the
theories considered here, that is by showing whether there
was any dissociation for obese individuals of the influence
of hunger versus fear on snack intake. Externality theory is
echoed in the measures of ‘external eating’ aspects of
disinhibited eating, and the renewed interest in reward
sensitivity, discussed below.
Restrained and disinhibited eating
Current theories of emotional eating emerged from a
literature examining the influence of negative affect
including stress on overeating; the two key concepts in
this literature are ‘restrained’ and ‘disinhibited’ eating (Van
Strien et al., 1986; Wardle, 1987). Restrained eating, or
dietary restraint, is the tendency to restrain (restrict) food
(or energy) intake to maintain or reduce body weight,
sometimes also characterized as ‘eating less than you want
to’ (Van Strien et al., 2009). Although this concept overlaps
with ‘dieting’, it does not necessarily imply successful
weight maintenance; on the contrary, the original theory
was developed to explain why ‘dieters’ are often unsuccessful in this (Herman and Mack, 1975), with the Restraint
Scale consisting of items measuring weight fluctuation,
dieting, bingeing and cognitions and emotions concerning
amounts eaten. Thus, eating was said to be ‘restrained’
within certain limits, but events and situations, including
emotional challenges, could conspire to induce eating
beyond those ‘boundaries’ – an effect known as ‘disinhibition’, or ‘counter regulation’, where the event was previous
food consumption. Following on from Schachter’s (1968)
results showing a lack of suppression of eating by fear in
obese individuals, a substantial number of experimental
studies have tested the hypothesis that this effect is
predicted by high restraint, whether obese or not (Greeno
and Wing, 1994; Wardle and Gibson, 2002), that is
restrained eaters do not suppress eating when stressed.
The results are generally in support of this prediction,
although, in line with comfort eating being associated with
overeating, sometimes the finding was that restrained
eaters and/or obese ate more under stress. Refinements to
the hypothesis have included whether or not the stressor is
ego threatening, and a possible role for cognitive load,
rather than stress per se, in disrupting the cognitively driven
restraint on eating (Ward and Mann, 2000; Wallis and
Hetherington, 2004). Indeed, Wallis and Hetherington
(2004) showed that eating chocolate was increased by both
ego-threatening and cognitively demanding stressful tasks
in restrained eaters, but only by the ego-threatening stress
in emotional eaters (defined by DEBQ).
Given the nature of those Restraint Scale items, it is not
surprising that obese individuals often score more highly on
restrained eating than nonobese individuals (Ruderman,
1983). However, subsequent theories have proposed that
normal-weight successful dieters may regularly restrain
their intake by using cognitive strategies, but avoiding
susceptibility to external or emotional stimuli: thus, scales
were developed that explicitly sought to separate out pure
cognitive restraint strategies from tendencies to be induced
to overeat or at least to break the diet. These tendencies
were measured together by the Disinhibition scale of the
TFEQ (or Eating Inventory; Stunkard and Messick, 1985),
but were separated into External and emotional eating
scales in the DEBQ (Van Strien et al., 1986). Weinstein et al.
(1997) reported student survey data (52 women, 49 men)
that supported disinhibition (TFEQ) as a better predictor
of comfort eating in women than restraint, although neither
was predictive in men (possibly because of lower scores and
less variance). In that study, comfort, or stress-induced,
eating was defined by two questions on how stress altered
the amount they would eat, one asking about a specific
stressful day and the other about stress in general: stress
eaters were categorized as those who responded that they
ate either more than usual or ‘binged’ when stressed (and
were as likely to be men as women in that sample). By
comparison, Westenhoefer et al. (1994) showed that overeating could be better predicted by the combination of
high scores on both the restraint and the disinhibition
scales of the TFEQ (and either of those could be reflected
as high scores on the Restraint Scale), at least in nonobese
young women: it was argued that such individuals may be
particularly vulnerable to loss of control overeating, because
of using ‘all or nothing’ rigid rules that are too easily
threatened (Westenhoefer et al., 1994). Using data from
version 2 of the revised 18-item TFEQ (TFEQ-R18v2;
emotional eating scale increased to six items) in a large
clinical trial of obese patients (N = 1741) and also a
nonclinical web-based survey (N = 1275), Cappelleri et al.
(2009) found a small positive correlation between cognitive
restraint and BMI in the nonobese sample, but a negative
correlation in the obese sample: one might infer a tipping
point such that restraint is lost beyond a certain BMI, and it
may be relevant that in both samples the group with the
highest restraint score had significantly more variability in
BMI than the lower restraint score groups. Nevertheless,
many other studies that included overweight and/or obese
participants have reported positive correlations between
dietary restraint and BMI (Ruderman, 1983; Van Strien
et al., 2009), so that it remains possible that being
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Psychobiology of comfort eating Gibson 447
overweight and at least moderately obese may cause
restrained eating rather than vice versa – this causality
question has not been resolved (Van Strien and Koenders,
2012).
Of particular relevance to comfort eating, Haynes et al.
(2003) showed, using experimentally induced stress (task
failure), that stress-induced eating, from a buffet lunch,
was seen reliably in women scoring high on both the
restraint and the disinhibition scales of the TFEQ, and to
a lesser extent in those scoring low on both scales,
whereas high disinhibition combined with low restraint
was associated with eating the most under no stress but
less under stress. These latter women were presumably
easily disinhibited to eat more by the buffet meal;
however, it would appear that a history of restrained
eating is necessary to permit disinhibition under stress.
Although appetite ratings did not differ by restraint or
disinhibition, it is possible that a tendency to undereat
when not stressed might encourage comfort eating when
stressed, in easily disinhibited women, as the comfort
foods would likely be harder to resist if any negative
energy balance were present – a prediction supported by
animal models of ‘comfort eating’ discussed below. It may
be relevant that mood data suggested that high
disinhibited eaters may also be more affected by stress.
Similar findings were reported in a more recent study
examining the effects of positive, negative and neutral
mood manipulations (using films) on snack intake in
women (Yeomans and Coughlan, 2009). That is, women
who were both disinhibited and restrained eaters ate more
after the negative mood manipulation. Interestingly,
disinhibited women who were also low in restraint ate
most after the positive mood film; any relation to adiposity
was not addressed in that study, but in a questionnaire
survey of emotional eating, underweight respondents
claimed to eat more during positive mood states (Geliebter
and Aversa, 2003). Furthermore, negative mood induction
was recently shown to increase the urge to eat only in ‘high
disinhibition’ women who had been exposed to (but not
actually tasted) desired foods (Loxton et al., 2011). These
and other studies (Lowe and Kral, 2006) suggest that
restrained eating per se is not a very reliable predictor of
eating behaviour during negative affect.
Using a progressive ratio reinforcement task to measure
the relative reinforcing value of energy-dense snack foods
(vs. that for reading interesting media) in 273 obese and
nonobese adults, Epstein et al. (2012) showed recently
that disinhibition alone predicted snack energy intake
and BMI, but it also enhanced the association of food
reinforcement with both intake of such snacks and BMI.
There was also a weaker moderating effect of restrained
eating (TFEQ), in that restrained eaters who also found
snack food more reinforcing were more likely to have a
higher BMI. Together with the findings discussed earlier,
this would fit with restraint both resulting in enhanced
food reinforcement through eating during negative
energy balance and allowing overeating when the
cognitive control is broken by overpowering temptation.
These eating styles of restraint and disinhibition have
begun to be considered in the context of broader
personality characteristics such as impulsivity, (poor)
delay discounting (ability to wait for a larger delayed
reward) and reward sensitivity (Davis and Fox, 2008;
Yeomans et al., 2008; Appelhans et al., 2011; Volkow et al.,
2011). However, whereas disinhibition may include an
emotional as well as an external eating component, when
these constructs were considered separately, and their
relationships with depression scores were examined by
structural equation modelling, in a sample of 549 mainly
obese or overweight women, depression was directly
predictive of emotional but not external eating (Ouwens
et al., 2009). Moreover, the final model showed that
impulsivity was predictive of both emotional and external
eating, but mainly contributed to emotional, but not
external, eating by interacting with a measure of
sensitivity to one’s feelings. These findings may help to
explain the results from an experimental manipulation of
negative affect (task failure) in a younger nonobese
student sample, whereby the ability of negative affect to
increase self-reported emotional eating tendencies was
slightly enhanced by impulsivity (Bekker et al., 2004).
These intermediate traits between depression and
emotional eating are also in line with understanding
comfort eating as a component of broader emotional
dysregulation (Tice et al., 2001).
External or uncontrolled eating may combine facets of
impulsivity, weak delay discounting and reward sensitivity: it is therefore quite surprising that, in a nationally
representative sample of 1342 Dutch adults, restrained
and emotional eating both moderated the relationship
between reported overeating tendencies and being overweight, whereas external eating did not (Van Strien et al.,
2009). Restrained and emotional eating were both higher
in overweight than in normal-weight participants, but
external eating was not. Moreover, when the effects of
restrained and emotional eating were partialled out in a
regression analysis, external eating actually became
associated with a lower risk of being overweight.
In summary, this section has shown how a need to
measure emotional eating as a separate construct
emerged from earlier theories, and that emotional eating
scores are often stronger predictors of overeating than
other eating styles such as restraint or disinhibition.
Emotional eating as a risk factor for obesity is considered
further in the final section.
Are some foods more comforting than others?
Intuitively, foods eaten to provide relief from negative
affect may have particular qualities, such as sensory or
nutritional attributes, that allow them to have an
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
448 Behavioural Pharmacology 2012, Vol 23 No 5&6
emotional impact on the eater. Thus, in contrast to the
effects of restrained eating, for which the total energy
intake may be the key goal, the prediction for emotional
eating is that any increase in intake should be for
particular food groups that could induce a reinforcing
shift in emotional state. This section briefly reviews
animal and human evidence to support this hypothesis. It
should be acknowledged that the animal work typically
involves the use of apparently quite strong stressors, such
as physical restraint in rats, whereas comfort eating in
human beings occurs to negative affect where the level of
stress (if present) is evidently not severe enough to
suppress appetite, at least in susceptible individuals.
Nevertheless, there is often a striking similarity in the
effects on eating behaviour and food choice between
studies in the different species, and the purpose of this
section is to summarize evidence that comfort eating is
associated with consumption of particular foods. Furthermore, in humans, stimulation of the hypothalamic–
pituitary–adrenal (HPA) axis and the release of cortisol
can occur in response to relatively mild eating-related
events including protein-rich meals (Gibson et al., 1999)
and food anticipation or withdrawal (Ott et al., 2011), not
just severe stressors. The extent of HPA axis activity can
also be predicted by quite subtle variations in affective
predisposition (Steptoe et al., 2007).
Evidence from animal studies
Results of earlier work using animal models to study the
effects of stress on food choice were rather inconsistent,
perhaps because of the varied approaches used (Wardle and
Gibson, 2002). However, an interesting observation was
that stress could reduce rats’ tolerance of unpalatable
adulteration of food, for example with bitter quinine,
implying one mechanism by which comfort eating might
focus on palatable rather than on bland or unpalatable
foods. By comparison, more recent models suggest that
stress-induced ‘binge eating’ can be found in rats, provided
that they experience both stress and food restriction, and
that they have a choice of highly palatable food (e.g.
chocolate) as well as bland laboratory chow (Hagan et al.,
2002; Boggiano et al., 2005) (cf. the interaction of restraint
and disinhibition in women, above).
Compelling evidence also began to emerge from animal
models that eating certain foods might counteract some
physiological effects of stress, for example an energy-dense
diet can prevent the loss of appetite caused by inescapable
stress in rats (Dess et al., 1998). Indeed, there is now
substantial support for suppression of the HPA axis
activation in rats allowed to eat sweet and/or fatty foods
(Buwalda et al., 2001; Dallman et al., 2003; Pecoraro et al.,
2006). A particularly powerful finding was that when
adrenalectomized rats voluntarily consumed sucrose (but
not saccharin), the metabolic deficits that normally follow
this procedure, in part because of the lack of corticosterone,
seemed to be prevented (Bell et al., 2000); in other words,
sucrose intake appeared to substitute for the missing
glucocorticoid stress hormone. Another related finding of
potential relevance to comfort eating was that rats allowed
a free choice of solid fat (e.g. beef, pork or vegetable
shortening) or chow had reduced HPA axis responses to
stress (physical restraint), including both lower adrenocorticotrophic hormone (ACTH) and corticosterone responses, compared with rats fed a 50% solid fat/chow mix
(similar in ratio and amount to that eaten by the choice
group) (La Fleur et al., 2005). Thus, choice rather than
energy per se may be important, or conversely, having one’s
preferred food forcibly adulterated may prevent the stress
reduction that free consumption of solid fat allows. That
study also suggested that the adipocyte hormone leptin was
not a critical signal for dampening down the HPA axis
responses (ACTH, corticosterone) as the levels increased
equally for both diet conditions.
In a more recent study, in which rats were fed for 7 days
with sucrose or solid fat or both (separately) before
restraint stress, all diets inhibited the HPA axis responses
to stress, and were particularly effective in reducing the
initial peak in ACTH release (Foster et al., 2009).
However, the least effective condition was the choice of
both sucrose and solid fat, despite this group eating the
most energy and depositing the most fat. Foster et al.
(2009) also showed that eating sucrose compared with
solid fat can exert separate effects on brain limbic
pathways and corticotrophin releasing hormone synthesis
and distribution of activation. In addition, after the
stressor, rats ate some sucrose (where available), but not
much solid fat, indicating that sucrose may be more
reinforcing to eat after stress than fat, and possibly that
the rats had learned this, although consuming sweet foods
when stressed may also be driven innately (Booth et al.,
2010). Another study showed that 2-h restricted daily (or
3 days/week) access to a sugar/solid fat mix, which
induced binge-type consumption, was more effective in
suppressing HPA axis activation to restraint stress than
unrestricted access to this mix, and also maintained
poststress intake of the mix (Kinzig et al., 2008).
It is important to note that these studies were primarily
investigating how energy-rich diets eaten for at least a
week before the stress, and associated body fat gained,
were able to suppress the neuroendocrine response to
that stress. However, most of the studies were not set up
to test the hypothesis that rats (or human beings) may
learn to like foods that alleviate stress. Nevertheless, a
recent study in rhesus monkeys examining the effects of
psychosocial stress from subordination reported that,
after an acute stressor, energy eaten specifically from a
high-fat and sugar diet was positively predicted by the
increase in cortisol caused by the stressor (Michopoulos
et al., 2012). Furthermore, there was evidence that these
stressed monkeys acquired a habit of greater energy
intake, even when access to the high-fat and sugar diet
was withdrawn. It has been pointed out that it makes
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Psychobiology of comfort eating Gibson 449
adaptive sense for an animal to eat energy-dense foods
when their safety, and therefore the time available to
choose and eat food, may be threatened (Pecoraro et al.,
2006). Indeed, Pecoraro et al. (2004) showed that chronic
stress caused rats to select more sugar and fat instead of
chow, that is a more energy-dense diet, even though they
did not eat more energy overall. This shift in diet in turn
was associated with inhibition of the HPA axis. There is
increasing evidence from animal studies that activation of
the HPA axis and the impact of the released glucocorticoid
hormones lead to amplification of the incentive salience of
cues to reward and perhaps also to punishment (Pecoraro
et al., 2006). For example, injection of corticotrophin
releasing hormone into the nucleus accumbens, a key site
processing incentive salience, enhanced rats’ lever pressing
for sucrose in response to a predictive cue for that reward
(Pecina et al., 2006). Finally, perhaps echoing the observation in humans that comfort eaters are in the minority and
reflect individual variation, it should be noted that Pecoraro
et al. (2006) discovered that some of the observations in rats
described above could not be reproduced consistently in
other breeding lines of rats, even within the Sprague–
Dawley strain.
Evidence in human beings
Several human studies have also linked the HPA axis
response to stress to changes in food choice and intake:
these will now be reviewed, together with other human
studies linking comfort eating with changes in food
choice. Although the experimental studies largely address
the impact of stress on eating and food choice, rather than
specifically emotional eating, such studies are thus
comparable with the animal studies reviewed above;
furthermore, observational studies that include measures
of emotional eating are also considered here. Research
into comfort eating, or stress-induced eating, and food
choice (or dietary composition) in human beings has
involved surveys, observational (or quasi-experimental),
experimental laboratory, or epidemiological studies. The
observational (or naturalistic) studies typically make use
of known stressful episodes in participants’ lives, such as
exams for students, or very busy work periods, or else ask
participants to keep a diary of mood and stressful events.
For example, during high workload, female office workers
reported eating more fat and energy compared with a
normal workload period (McCann et al., 1990). Similarly,
in department store workers, busy periods were associated with an increased intake of fat, sugar and energy,
but only in restrained eaters (Wardle et al., 2000).
Exam periods among adolescent or young adult students
have been associated with higher energy intake (Michaud
et al., 1990) or, when negative mood is also present, with
a less healthy diet (Weidner et al., 1996). Similarly, a
national cross-sectional survey of Turkish students found
that stress was strongly linked to less likelihood of eating
fruit and vegetables (Unusan, 2006). However, not all
studies have detected changes in diet during exam
periods (O’Donnell et al., 1987; Pollard et al., 1995);
moreover, students have financial constraints (themselves
stressful) on the healthiness of their diets, and also
frequently experience stress at other times of the year.
In the UK study of the DEBQ measures in adolescents
(Wardle et al., 1992), school children completed 24-h
dietary recalls; whereas restraint was associated with less
energy intake and external eating with more, there was no
consistent relationship with emotional eating for either
energy or macronutrient intake. However, this could in
part be because of the relatively low levels of emotional
eating in that sample, as well as the likelihood that
emotional eating may not have been elicited in many of
those 24-h periods, that is the impact of emotional eating
on what or how much is eaten may be more ephemeral
than that for either restraint or externality.
In a sample of 47 female American college students asked
to record their food intake and mood just before each
eating episode, over 12 days, eating in the presence of
negative mood was only associated with snack intake, not
meal consumption, and greater snack intake under these
conditions was only observed for overweight students
(Lowe and Fisher, 1983).
In a diary study comparing high and low stress weeks
(hours worked), adult male and female nurses and teachers
(N = 69) reported eating more ‘fast food’ during high stress
weeks (Steptoe et al., 1998). Moreover, comfort eaters
(defined by the mood scale of the Food Choice Questionnaire) ate more sweet foods (excluding chocolate) than
noncomfort eaters, irrespective of workload. The men, but
not women, reported greater intake of red meat during
stress; however, the male sample was quite small (N = 24),
and thus requires caution in interpretation. Fruit consumption was not affected by stress, although consumption of
vegetables was not reported.
In a larger study of adult workers (N = 422) recording
daily stressors and diet, stressful periods were associated
with eating more high fat/sugar snacks and less main
meals and vegetables (O’Connor et al., 2008). Moreover,
the associations for stress and high-fat and sugar snacks
were moderated specifically by emotional eating
(DEBQ), supporting the prediction that high emotional
eaters eat more such snacks when stressed. This same
group also examined a possible role for the HPA axis in
linking stress (daily hassles) to snacking (defined as
between-meal food intake) by measuring cortisol response to a stressor involving public speaking and a
mental subtraction task in 50 women (Newman et al.,
2007). On the basis of this cortisol stress response,
women were grouped either as high or as low reactors for
analysis. The intensity and frequency of daily hassles
both predicted snack intake, and this relationship was
stronger for the high reactors. However, the relations
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
450
Behavioural Pharmacology 2012, Vol 23 No 5&6
between positive and negative mood and snacking (less
and more, respectively) were only significant for the low
reactors. This interesting finding suggests there could be a
dissociation between cortisol-driven snacking and mooddriven snacking; in other words, it questions whether high
stress reactors are in fact primarily comfort eaters. It should
be noted that the ‘high reactors’ were defined on their
response to an acute stressor only. It is also possible that
such ‘high reactors’ are not in fact dysfunctional; on the
contrary, an unresponsive HPA axis has been associated
with atypical depression (Gold and Chrousos, 2002), which
is the form associated with hyperphagia, as distinct from
the loss of appetite associated with more typical melancholic depression, and a recent report showed that a history
of chronic stress in women was associated with a lower
cortisol response to an acute stressor (while at the same
time they felt more stressed), and that these low-reactor
but higher stressed women reported higher levels of
emotional eating (Tomiyama et al., 2011). In addition, the
evidence that these women tended to have more
abdominal adiposity, which in turn was correlated with
lower cortisol, is in line with Dallman’s ‘chronic stress
response network’ theory (Dallman et al., 2003) that eating
an energy-dense diet and accumulating more central fat can
suppress the responsiveness of the HPA axis, possibly
mediated by an increased sensitivity of the axis to negative
feedback inhibition from the circulating hormone. Nevertheless, it is at odds with earlier research showing that
abdominal adiposity in women was associated with higher
cortisol release after stress (Moyer et al., 1994; Epel et al.,
2000), although this finding was not replicated in men
(Epel et al., 1999). A key difference might be the level of
chronic stress experienced in the highly stressed women in
the study of Tomiyama et al. (2011). In addition, daily
cortisol secretion, indexed by urinary free cortisol, is
positively correlated with waist circumference and energy
intake in women, independent of BMI (Vicennati et al.,
2011): this appears to conflict with the finding, in nonobese
young women, that free fatty acid infusion inhibits the
HPA axis (Lanfranco et al., 2004), although in obese
individuals, insulin resistance as well as stress can lead to an
increase in free fatty acid levels in association with higher
cortisol levels (Bjorntorp, 1999). Furthermore, elevated
glucocorticoid release is known to enhance motivation for
appetitive behaviours (Pecoraro et al., 2006), and thus, it is
difficult to dissociate cause and effect in such findings.
In another laboratory stress study, in 59 women, that
similarly divided participants into high and low stress
reactors on the basis of cortisol response, and incidentally
offered snacks to eat, high stress reactors were seen to eat
more of sweet snack foods, and more energy when
stressed (Epel et al., 2001). By comparison, when the
effects of a laboratory stressor were examined on food
choice over lunch, in 68 men and women, there was no
effect of stress on intake or choice overall, but emotional
eaters specifically ate more sweet fatty foods when
stressed (Oliver et al., 2000). By contrast, cognitive
restraint (DEBQ) did not significantly interact with
stress and eating. In an experimental study in women
using anticipated public speaking as a stressor and a
computer-based task to assess the relative reinforcing
value of snacks versus fruit and vegetables, binge eating
tendency (associated with comfort eating) interacted
with self-reported stress sensitivity (Goldfield et al.,
2008): stress-reactive binge eaters found snacks more
reinforcing when stressed, whereas the opposite was the
case for nonbinge eaters. Stress made no difference to
food value in those less sensitive to the stress.
Sometimes, an effect of stress on food choice can be easy to
demonstrate: Zellner et al. (2006) simply gave 34 young
female students either solvable (control) or unsolvable
(stressed group) anagrams and allowed them to eat freely
from servings of snacks, that is crisps, peanuts, chocolate
M&Ms and grapes. The stressed group ate more M&Ms
and fewer grapes than the unstressed group; the other
foods were less popular and unaffected by stress condition.
However, eating attitudes were not measured in this
sample, and thus we cannot know the contribution of
emotional eating tendencies. Zellner et al. (2007) repeated
this experiment in men, but found quite the opposite
result, in that it was the unstressed group who ate more of
the unhealthy snacks (crisps and M&Ms). Again, we do not
know whether any of the men were in fact emotional
eaters, although it seems unlikely that many were: Oliver
et al. (2000) found that emotional eating predicted food
choice changes induced by stress in both men and women.
However, there are many reports that stress suppresses
food intake in individuals low in restraint, as men often are
(Greeno and Wing, 1994; O’Connor et al., 2008). By
contrast, there is additional evidence that some men may
indeed be susceptible to comfort eating: using food
frequency and questionnaire data, we reported recently
that men who were both poor at stress coping and suffering
from stress at work reported eating more sweet fatty foods
than other men (Gibson et al., 2008). Two recent
experimental studies of stress and eating in female
students (Van Strien et al., 2012) make it clear that to
show reliably that emotional eaters are susceptible to
increased energy intake during stress, it is necessary to have
a wide distribution of emotional eating tendencies in the
sample (see above). In agreement with other findings,
where foods varying in energy density and sensory properties have been available, stressed emotional eaters specifically ate more of the sweet fatty ‘butter cake’ versus the
healthier less energy-dense foods.
There is also survey-based and epidemiological evidence
to suggest that stress and/or negative mood may be
associated with changes in food choice, not just intake.
For example, in US adolescents, depressive symptoms
were associated with increased intake of sweet soft
drinks, as well as more disordered eating including meal
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Psychobiology of comfort eating Gibson 451
skipping (Fulkerson et al., 2004). Similarly, in 11–13-yearold children in London, those with greater self-reported
stress ate more fatty foods and snacks but less fruit and
vegetables, and tended to skip breakfast (Cartwright
et al., 2003). In Finnish adults, ‘stress-driven eaters’ ate
more high-fat foods and had higher BMI (Laitinen et al.,
2002). In a more recent Finnish survey of 1679 men and
2035 women, emotional eating (TFEQ-R18) was predictive of higher intake of sweet foods for both men and
women and for nonsweet fatty foods in men (Konttinen
et al., 2010). In addition, the positive correlation between
depression and sweet food intake was largely explained by
emotional eating. In a survey of 3266 Swedish parents and
1441 12-year-old children, emotional eating (albeit
averaging <2 on the DEBQ scale) was correlated with
greater frequency of eating confectionery and soft drinks
(except for boys); external eating was also similarly
related, whereas restrained eating was associated with
eating confectionery less often (Elfhag et al., 2008). In
520 French adults, emotional eating was associated with
more snacking on cakes, pastries, biscuits and oily fruit
such as olives, as well as sweet beverages, whereas eating
more energy-dense food was predicted by uncontrolled
(disinhibited) eating (De Lauzon et al., 2004). However,
in another French sample, of 1320 parents and adolescent
children, emotional eating (DEBQ) was unrelated to
macronutrient intake (from 3-day diet diaries) (Lluch
et al., 2000). Similarly, in a sample of 475 young Dutch
women, emotional eating was not related to macronutrient or energy intake (Anschutz et al., 2009), although
such nutrient-based data can obscure subtle changes in
food choice.
In an online survey of nearly 3000 Australian adults with
clinical depression, chocolate was craved by half the
respondents and believed by them to provide relief from
the negative affect (Parker and Crawford, 2007); these
chocolate cravers also had more stress-prone personalities. A similar pattern was observed in 931 men and
women from San Diego, California, who were free from
antidepressant medication: Rose et al. (2010) examined
the relationship between depression scores (Center for
Epidemiological Studies Depression Scale) and the rate
of chocolate consumption. Using three cut-points in the
distribution of depression scores, chocolate consumption
was shown to increase markedly with increasing severity
of depression: thus, the group scoring below the cut-point
for any depression reported eating 5.4 servings per month
on average, whereas those with scores indicating moderate depression ate 8.4 servings per month, and the
highest scoring group (likely major depression) ate a
mean of 11.8 servings per month. This trend was clearly
statistically significant, and could not be accounted for by
greater energy, fat or carbohydrate intake. The trend was
present for both men and women, although the effect was
somewhat stronger in women. Obviously, in these crosssectional data, causality cannot be determined. Further-
more, a prospective study in 3486 UK civil servants,
which allows some assessment of causality, or at least
sequence, examined the risk for depression (using the
same instrument) 5 years after the assessment of diet and
other confounders (Akbaraly et al., 2009): a healthy ‘whole
food’ diet was found to reduce the risk of depression by
26%, whereas a high-fat ‘processed food’ diet, including
sweet desserts, predicted 58% greater risk of depression 5
years later. Still, the lack of effect of energy and fat intake
on the depression-chocolate link makes this reversecausality explanation less likely for the findings of Rose
et al. (2010). A harder finding to interpret, but nevertheless intriguing, and related, is that per capita sugar
consumption across six nations was very highly correlated
(r = 0.95) with the annual rates of depression for those
countries (Westover and Marangell, 2002).
In smaller surveys, some quite consistent findings emerge.
Among 212 London students asked about how stress affects
their eating, including the intake of a variety of main food
groups, sweet fatty foods including chocolate were chosen
when stressed, even among those who claimed they usually
ate less overall when stressed. By contrast, fruit, vegetables,
meat and fish were clearly eaten less when stressed. This
effect was replicated in another London student population, in whom it was shown that both emotional and
uncontrolled eating were associated with self-reported
stress eating (Vögele and Gibson, 2010). Among US male
and female students, preferred comfort foods were sweet,
especially chocolate (Zellner et al., 2006); comfort (stress)
eating was predicted by higher scores on the Restraint
Scale, but as discussed earlier, this could reflect emotional
and uncontrolled eating components as well. In an online
survey of female US students, stress resulted in less
healthy food choices, and among those eating more when
stressed, sweet and/or high-fat foods were typically chosen
(Kandiah et al., 2006). In 60 obese Swedes, higher
neuroticism was associated with a preference for sweet
and fatty foods (Elfhag and Erlanson-Albertsson, 2006):
disinhibited eating (TFEQ) showed a similar but nonsignificant trend; emotional eating was not measured but
may have been subsumed under the disinhibition measure.
In two surveys of comfort food preferences in randomly
sampled North American adults, a slightly different
picture emerged (Wansink et al., 2003). In the first
survey asking about preferred comfort foods, with 411
complete responses out of 1000 sampled, potato crisps
(US ‘chips’) were the most frequently endorsed comfort
food; other fatty and sweet snack foods were popular,
including chocolate and ice cream. However, 40% of the
sample preferred meal-type foods as ‘comfort foods’
including pasta and pizza. Unfortunately, we do not know
how many of the participants would actually score highly
on an emotional eating questionnaire, and the interpretation of comfort foods here could include the nursery
food connotation discussed earlier. Indeed, the finding
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
452 Behavioural Pharmacology
2012, Vol 23 No 5&6
in the second survey that crisps were more likely to be
comfort foods for 18–34 year olds, whereas savoury mealtype foods were more likely comfort foods for the older
groups might support such an interpretation. The second
survey (602 women, 401 men) also found sex differences,
in particular that women were more likely to endorse
chocolate as a comfort food (and feel more guilty about
eating it), whereas men were more likely to endorse steak.
Can comfort foods improve mood and/or
relieve stress?
Although the question of whether comfort foods can
actually comfort the eater is clearly a key one, there is
relatively limited direct evidence available, and the
temporal characteristics of such an effect are even less
understood. Thus, both direct and indirect evidence is
discussed here. Mood and food can be related in several
ways, including influences of positive mood on sensory
appeal and appetite, as reviewed elsewhere (Gibson,
2006; Macht, 2008), but this review focuses on the effects
of foods on negative affect. Furthermore, Ganley (1989)
reviewed some early studies of the impact of eating on
negative affect; these suggested that eating was most likely
to reduce negative affect that was diffuse and difficult both
to explain or label and to control. This is a theme that is
reflected in the experimental manipulations of affect
described in the following studies. Another theme that
emerges from those earlier studies is that stress-induced
eating depended on the food available being palatable and
energy dense. This may be important to the extent that
comfort eating can be distinguished from binge eating:
although both may be elicited by some form of negative
affect, the very large intakes associated with binge eating
may suppress emotions such as anxiety and anger in part by
the huge demands on the digestive system shifting
autonomic balance in favour of parasympathetic outflow
for an extended time: thus, bingeing is typically followed by
lower energy levels, sleepiness, etc. (Ganley, 1989; Waters
et al., 2001). By comparison, comfort eating, which might
include relatively moderate snack food intake, may rely
more on subtle sensory or neurophysiological consequences
of particular foods. Nevertheless, they may both share an
important outcome that the foods eaten provide some
alleviation of negative affect, even if primarily by distraction
from the thoughts and threats that may be engendering the
negative affect (Heatherton and Baumeister, 1991; Waters
et al., 2001). Interestingly, in a study of emotional
antecedents of binge eating, bingeing was predicted best
by tense arousal (anxiety) and yet the binge itself did not
reduce that emotional state (Waters et al., 2001): the
inference is that bingeing may be reinforced more by escape
from negative rumination than by affect reduction. However, bingeing may have its own negative consequences on
mood that more moderate comfort eating may not.
Chocolate is probably the archetypal comfort food: survey
data suggest that chocolate is one of the most likely foods
to be eaten during stress or negative mood (Gibson,
2006), as well as being the most craved food (Rozin et al.,
1991; Gibson and Desmond, 1999). There are several
experimental studies that lend some support for the
popular idea that chocolate can lift mood. Chocolate,
particularly milk chocolate, is most likely popular as a
snack in part because of the energy content (Gibson and
Desmond, 1999), and the sensory properties predictive of
that (Michener and Rozin, 1994), but also because of the
psychoactive effects of the methylxanthine components,
caffeine and theobromine: thus, Smit et al. (2004)
showed, in a well-controlled study, that these compounds,
in levels normally found in a small bar of chocolate, can
improve mood as well as cognition, and led to increased
liking of a drink flavour paired with them (Smit and
Blackburn, 2005). Indeed, the beneficial effects of these
compounds in chocolate might underlie its long history of
use apparently for psychoactive reasons, even in forms
lacking any innate sensory appeal. It is possible that
chocolate, eaten on an empty stomach, could provide a
sufficient stimulus to insulin release, such that brain
serotonin synthesis is increased by increased uptake of
the precursor amino acid tryptophan across the blood–
brain barrier; yet, we are not aware of any attempt to test
this directly (Gibson, 2011). However, the sweetness of
some chocolate might be enough to alleviate stress
through stimulation of endogenous opioid release, as
sweet and perhaps fatty taste is known to calm distraught
human and rat infants, and in the latter at least, this
effect is blocked by opiate antagonists (Blass et al., 1989);
moreover, chocolate chip cookies can alleviate pain in
female students when other nonsweet foods are ineffective (Mercer and Holder, 1997).
In female students in whom negative mood was induced
by an unsolvable task, those with binge eating or bulimic
tendencies (i.e. with at least moderate scores on the
Bulimia Test-Revised questionnaire) ate more chocolate
chip cookies during a bogus taste test and reported
improved mood, compared with controls (low scorers)
(Kisler and Corcoran, 1997). The authors pointed out
that their apparently ‘comfort eating’ group scored below
the usual clinical cut-off for bulimia nervosa (BN) on that
questionnaire, but clearly had greater bingeing tendencies than the control group: the point is that eating
chocolate chip cookies alleviated negative mood in this
susceptible group. Similarly, Macht and Mueller (2007)
showed that chocolate could reverse the negative mood
induced by a sad film, but more strongly in comfort
(emotional; DEBQ) eaters. This effect occurred within a
minute but lasted for just three minutes. The additional
benefit to comfort eaters could reflect either a disposition
or being more practised in using chocolate in that way.
Willner et al. (1998) examined the effects of negative
versus positive mood induction, using music, on motivation to eat sweet milk chocolate ‘buttons’, compared with
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Psychobiology of comfort eating Gibson 453
that for nonsweet carob buttons (a form of chocolate
substitute that participants reported disliking, perhaps in
part because of unfamiliarity), in a progressive ratio task.
This task involved the young women being rewarded with
chocolate for pressing the spacebar a required number of
times, which increased geometrically by doubling each
time. The main measure was the breakpoint, that is the
maximum number of presses after which the participant
no longer responded for 2 min. Negative mood substantially increased the breakpoint, and number of buttons
achieved, specifically for the chocolate reinforcer. It was
found that the concurrent craving for chocolate (measured by questionnaire), which was increased by negative
mood induction, correlated in this group with the
breakpoint for chocolate: as Willner et al. (1998) cite
other evidence that negative mood is associated with a
reduction in sensory pleasure from foods such as
chocolate, it may have been the increased craving, rather
than the sensory reinforcer value of the chocolate per se,
that was driving the responding. By comparison, despite
participants’ self-report about dislike for carob, following
the positive mood induction, both the breakpoints
and the number of buttons consumed did not differ
between the chocolate and the carob conditions. There
was also a slight indication that contented mood, which
declined after the negative mood music, improved on
completion of the task, whereas contentedness remained
stable throughout in the positive mood condition. Of
course, it is not known to what extent the women were
comfort eaters, but it would be expected that comfort
eaters would show increased chocolate craving during
negative mood, reflecting the pattern observed in this
study. Moreover, we discussed above evidence that
comfort eaters who are stressed work harder for energydense snacks relative to fruits and vegetables (Goldfield
et al., 2008), which contrasts with the reduction of this
relative reinforcing value of snacks by state anxiety,
although only in unrestrained eaters (Goldfield and Legg,
2006) (disinhibited eating was not measured). In other
words, feeling stressed or anxious is likely to distract from
wanting energy-dense food, unless one is a comfort eater,
when the opposite is the case, probably because the
comfort foods distract from the perceived stress
(Heatherton and Baumeister, 1991; see below).
Some indirect evidence that comfort foods may be
comforting comes from the observation that patients
with winter depression (Seasonal Affective Disorder)
typically eat more sweet (and perhaps starchy) foods
during winter (and less after light therapy) (Kräuchi et al.,
1997); moreover, seasonal affective disorder sufferers
score highly on the DEBQ emotional eating scale. These
facets are predictors of weight gain during the winter.
This is reminiscent of the increased frequency of chocolate
eating associated with greater depression scores in a
nonclinical sample discussed earlier (Rose et al., 2010);
however, one might therefore infer that chocolate is not
very effective in improving mood over the long term, and
this conclusion was reached by Parker et al. (2006) in a
review of chocolate and mood. It is also in agreement with
the conclusion of Willner et al. (1998) that negative affect
drives a desire for chocolate (in some individuals at least)
that may be independent of the reinforcing value of
chocolate. Instead, comfort eating may be one aspect of a
broader problem of emotional dysregulation; for example,
rumination of negative thoughts that is typical of
depression predicts binge eating tendencies as well as
other unhealthy behaviours, all of which may serve to
distract from the negative thoughts (Selby et al., 2008).
Indeed, the apparent transient improvement in mood after
eating comfort foods might simply follow from a break
from such rumination.
Overall, it is clear that comfort eating is likely to be
associated with greater intake of highly palatable, especially
sweet and/or fatty foods (with variation between sexes).
These foods may have a short-term effect in reducing
negative affect, but the habitual use of such foods to
improve mood may be more dependent on their ability to
distract from negative rumination, in susceptible individuals.
Is comfort eating harmful and can it be treated?
The most obvious concern for comfort eating and health
is that it will lead to overconsumption of energy and
ultimately to obesity. Indeed, emotional or comfort eating
is one of the earliest theoretical accounts of obesity
(Kaplan et al., 1957; Stunkard, 1959; Bruch, 1961);
however, those early accounts included aspects in their
theories that would now be regarded as related to
restrained or external eating (Schachter, 1968), which
are now more widely regarded as separable from
emotional eating (Van Strien et al., 2009).
Ganley (1989) reviewed three decades of mainly clinical
studies since those early theories were published, to
address the contribution of comfort eating to obesity.
Although clinical reports of episodic eating induced by
stress or negative affect (sometimes termed binge eating,
although unless excessive intake is confirmed, comfort or
emotional eating is more accurate) were commonplace for
obese patients, nevertheless, the actual proportion of
patients (undergoing weight loss programmes) showing
this eating behaviour was almost invariably no more than
half, that is often a minority of patients. This is in
agreement with the findings on prevalence discussed
earlier. However, where nonobese controls were studied,
evidence of emotional eating tended to be less than that
in obese individuals, and in surveys of obese individuals
not in weight loss programmes, reports of emotional
eating were much higher, and also higher than that in
control groups. Even so, there is a risk that at least some
of the association of negative affect and eating commonly
reported by obese patients could be an epiphenomenon,
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
454
Behavioural Pharmacology 2012, Vol 23 No 5&6
that is the obese overeat regularly and suffer negative
affect and stressful lives, but there may not necessarily be
a causal link. However, experimental studies have largely
found that obese individuals were more likely to show
comfort eating (of energy-dense foods) in response to
stress than nonobese individuals, and a meta-analysis of
relevant prospective studies concluded that stress made a
small but significant contribution towards weight gain
(Wardle et al., 2011). Moreover, most modern conceptualizations of the aetiology of obesity incorporate the
impact of stress on eating in some form (Butland et al.,
2007).
Earlier studies, of at least a moderate sample size, that
included measures of both emotional eating and obesity
typically reported higher emotional eating in obese or
overweight groups (Van Strien et al., 1985; Wardle, 1987)
(Table 1). However, since the 1980s, the prevalence of
obesity has increased at least three-fold in many
countries; thus, one might expect a similar increase in
emotional eating. There are insufficient data to address
this directly, but the studies carried out 25 years apart on
two representative samples of Dutch adults discussed
earlier did show that emotional eating had on average
increased significantly, as had obesity (Van Strien et al.,
1986, 2009). In the recent study, both low dietary
restraint and high emotional eating were predictive of
overeating and weight gain (cross-sectionally), whereas
external eating was not, and the authors conclude that
‘we may be well advised to seek an ‘‘emotional’’
explanation for the current obesity epidemic’ (Van Strien
et al., 2009, p. 386).
In their study of a French population, De Lauzon et al.
(2004) did not report relationships between BMI and
emotional eating; however, they found that middle-aged
men, but not women, who scored high on emotional
eating consumed significantly more energy. By contrast,
in the slimmer young adults and teenagers, emotional
eating was unrelated to energy intake, whereas cognitive
restraint in the women was predictive of less intake. In a
follow-up study in this population, emotional eating was
not related to adiposity measures (De Lauzon-Guillain
et al., 2006), unlike cognitive restraint, although it should
be borne in mind that the emotional eating scale of that
version of the TFEQ-R18 is limited to three items.
One quite impressive finding is from a study of 2997
young Finnish women (aged 17–20) (Anglé et al., 2009):
the sample included women classified by BMI as
underweight, normal weight, overweight and obese, and
emotional eating (TFEQ-R18) increased significantly
with each category of increasing BMI (Table 1).
In a recent large online survey of lifestyle in Dutch bank
employees (N = 3272), emotional eating was the strongest predictor, among lifestyle factors, of behaviours
related to weight gain, in both men and women
(Van Strien and Koenders, 2012), although the perceived
or the actual role of stress and emotions in these workers
might be somewhat inflated, as a zeitgeist of the current
economic situation.
In a prospective study of 4065 UK teenagers over 5 years,
perceived stress did not predict greater weight gain;
instead, indices of higher adiposity were associated
over all 5 years with higher stress, after adjusting for
likely confounding variables (Van Jaarsveld et al., 2009).
It seems probable that comfort eating often, although not
invariably, leads to weight gain: but even if obesity is not a
result, health could still be compromised by the increased
consumption of high-fat energy-rich foods. A short-term
prospective study looked at stress (comfort) eating in
medical students as a possible predictor of health
outcomes related to the metabolic syndrome. Epel et al.
(2004) followed these students from a baseline measurement through two exam periods: stress eaters gained
more weight, and had higher nocturnal insulin, cortisol
levels and blood total/high-density lipoprotein-cholesterol
ratio; thus, comfort eating can predict poorer health
outcomes, even in a young educated sample. This effect
might be driven by poor diet, but if comfort eaters are
also more stress reactive (see above), the physiological
consequences could be contributing towards worse health
and greater adiposity (Steptoe and Wardle, 2005). In
older male caregivers of spouses with Alzheimer’s disease,
the chronic stress of caring was predictive of psychological distress, which predicted poorer health behaviour,
including a high-fat diet, and that in turn predicted
metabolic syndrome and coronary heart disease (Vitaliano
et al., 2002); the women in the study were largely
protected by hormone replacement therapy.
In students in the UK, both self-reported weight gain and
weight loss after the first year at university were predicted by stress levels, particularly in women (Serlachius
et al., 2007); a plausible explanation for the weight gain
could be comfort eating. In a study of stress and weight
change in 71 student nurses, with measures 12 weeks
apart, before and during exams, the independent
predictor was dietary restraint (measured by the Eating
Disorders Examination Questionnaire), which was associated with weight gain, after adjusting for other eating
disorder tendencies, including bingeing, and changes in
cortisol (Roberts et al., 2007). Emotional eating was not
explicitly measured, although it is possible that the
restraint measure captured some aspect of that behaviour; also, as discussed above, restrained eating can
interact with and moderate the impact of emotional
eating on diet (Yeomans and Coughlan, 2009). The
reduction in restraint might have permitted emotional
eating to enhance energy intake. Furthermore, Van Strien
and Van de Laar (2008) have shown prospectively that,
whereas dietary restraint and external eating can change
after 4 years, and predict changes in fat and energy
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Psychobiology of comfort eating Gibson 455
intake, respectively, emotional eating tendencies seem to
be more resistant to change.
In summary, the balance of evidence links emotional
eating both to overeating, particularly of energy-dense
foods, and to increased adiposity, with a likely increased
risk to cardiovascular health. However, there is some
variability that may be related to predisposing factors.
The next section considers examples of these in terms of
genetic and neurochemical associations.
Genetic and neuropharmacological aspects
This stability of emotional eating over the years has also
been reported in children (Ashcroft et al., 2008), among
other eating phenotypes, and suggests a trait-like quality
that likely has a strong genetic component. Indeed, a
study of Swedish male twins concluded that emotional
eating (TFEQ-R18) showed high heritability of 60%
(Tholin et al., 2005); however, other twin studies of both
sexes have reported somewhat less heritability, for
example 9–45% for UK and Finnish populations
(TFEQ-R18) (Keskitalo et al., 2008), and 25% for a
Korean population (DEBQ) (Sung et al., 2010). Interestingly, in the UK and Finnish twin samples, emotional
eating also showed quite a strong genetic link to liking for
sweet fatty foods (Keskitalo et al., 2008).
Although developmental stability of emotional eating could
also be attributable to parenting style, one study that
compared 214 families with obese or overweight parents
with 114 families with normal-weight parents did not find
any difference among mothers’ use of an emotional feeding
style, although the obese mothers did show less control of
their children’s eating (Wardle et al., 2002). Conversely, and
intriguingly, excessive parental control over a child’s
behaviour in general has been found to increase the risk
of emotional eating, but only if the child has some genetic
vulnerability related to their dopamine D2 receptor binding
or expression. Thus, Van Strien et al. (2010) genotyped 428
adolescents for alleles of the Taq1 polymorphism (A1A1,
A1A2 or A2A2), as well as measuring, at baseline and 4 years
later, emotional eating (DEBQ) and the adolescents’
ratings of their parents control over their children’s
behaviour. The Taq1 A1 allele may be associated with
reduced dopamine-binding affinity at the D2 receptor
(Thompson et al., 1997) or possibly reduced expression of
D2 receptors (Stice et al., 2008a); as the D2 receptor can be
either a postsynaptic receptor on nondopaminergic target
cells or a presynaptic autoreceptor on dopamine releasing
neurones (Le Foll et al., 2009), predictions for the behaviour
of the A1 variants are not clear cut, and postsynaptic D2
receptors may also be downregulated over time, perhaps as
a result of chronic overeating (Stice et al., 2008a; Johnson
and Kenny, 2010). For instance, if this polymorphism
compromises autoreceptor function, it would be associated
with greater release of dopamine, for example in the ventral
tegmental pathways, than in the case of the A2 homozygous
variant. Thus, in D2 receptor knockout mice, psycho-
stimulant drugs release more dopamine than in their wildtype littermates (reviewed by Le Foll et al., 2009);
moreover, reward sensitivity, for both cocaine and food, is
enhanced in mice lacking specifically the D2 autoreceptor,
and that therefore release more dopamine in response to
cocaine (Bello et al., 2011). In any event, Van Strien et al.
(2010) did not find direct associations between A1 allele
variants and either emotional eating or parental control,
although emotional eating was significantly positively
correlated with parental control at baseline. Instead, they
found a significant interaction, such that the A1 genotype
was associated with greater emotional eating after a 4-year
follow-up only if the adolescents also reported high parental
control, that is a gene–environment interaction. One could
speculate that the restrictive parenting interacted with the
poorly regulated dopamine output to exaggerate the
incentive salience of energy-rich foods on the rare occasions
they were accessed. Also, because analyses controlled for
baseline levels of emotional, external and restrained eating
and BMI, a reverse-causality explanation, that is that
parents were more controlling because of their child’s
overeating tendencies, is less likely.
These results also need to be considered in the context of
emerging findings on dopamine polymorphisms, the risk of
obesity and activation of reward and sensory pathways in
response to presentation of visual or gustatory food stimuli
using brain imaging, particularly functional MRI. For
example, Stice et al. (2008a) found that young women with
higher BMI showed less activation of reward pathways such
as the caudate nucleus in response to expecting or tasting
milkshake versus tasteless water, and furthermore, this
inverse relationship was only true for women with the
Taq1A A1 allele. In a second prospective study of weight
change in female adolescents, the presence of the A1 allele
again exaggerated the association at baseline between
adiposity and blunted caudate responses to milkshake. At
the 1-year follow-up, the increase in BMI was positively
related to caudate activation only in girls without the A1
allele (Stice et al., 2008a). These associations in motivational pathways contrast with the finding that, at the same
time, obese adolescent females show greater activation of
gustatory cortex to milkshake, compared with lean
adolescents (Stice et al., 2008b), implying a dissociation
between perceptual and motivational responding. Notwithstanding any functional differences in dopamine activation (assuming that the regional activations are closely
related to dopamine function), this might suggest cognitive
inhibition of motivational pathways such as might result from
greater restraint. However, a role for the distinct genotypes in
restraint would also be needed to support such a process.
In a similar design in female adolescents, but including
genotyping for dopamine D2 and D4 receptor polymorphisms, future weight gain was associated with weaker
responding to food images in brain regions mediating
incentive value for those with D2 Taq1A A1 and D4
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
456
Behavioural Pharmacology 2012, Vol 23 No 5&6
seven-repeat alleles, whereas for those without these
alleles, greater activation in those areas was predictive of
future weight gain (Stice et al., 2010). The authors
suggest that there may be two distinct neurochemical
predispositions for obesity, governed by particular dopamine polymorphisms.
Of particular relevance here, this group has also used the
milkshake anticipation and tasting paradigm with functional MRI to dissociate responses of emotional and
nonemotional eaters (DEBQ) following negative mood
induction (by music and negative situation recall) (Bohon
et al., 2009). For activation of brain cortical regions
concerned with affective state and anticipatory reward
(parahippocampal gyrus and anterior cingulate gyrus), there
was an interaction between induced affect and emotional
eating status: that is, emotional eaters showed relatively
greater activation in those regions, in response to chocolate
milkshake, when in the negative mood, but less activation
when in the neutral mood. By comparison, emotional eaters
also showed greater overall activation of reward areas,
including the caudate and pallidum, to the taste of the
milkshake, although not its anticipation. Moreover, key
regions such as the amygdala and the ventromedial
prefrontal cortex were not differentially activated in
emotional versus nonemotional eaters in this context,
although the study could have been underpowered to
detect such effects. Thus, the authors conclude that
emotional eaters do not differ in their emotional responses
to food, but rather show heightened reward induced by (at
least) chocolate milkshake, particularly when in a negative
mood. This could in part reflect a learned response cued by
negative mood; however, it is also possible that activation of
the HPA axis is driving the enhanced reward sensitivity, as
discussed earlier. This result is also in agreement with the
earlier conclusion that such foods may be particularly
effective at distracting emotional eaters from their negative
ruminations.
These complex findings make it difficult to consider
recommendations for pharmacotherapy for comfort eating, if it were warranted, even just for dopaminergic
agents, although they also suggest a future role for
pharmacogenomic approaches. For instance, one theoretical approach would be to use dopamine agonists
selective for the presynaptic D2 receptor specifically in
comfort eaters having the Taq1 A1 allele. Conversely, a
therapy targeted at promoting the expression of postsynaptic D2 receptors might be effective in other
overeating groups.
The complications of dopamine pharmacotherapy are well
illustrated by an analysis of the relationships between
dopamine and serotonin receptor binding affinities of
typical and atypical antipsychotic drugs and treatment
efficacy (Richtand et al., 2007), the latter in particular
being contraindicated for patients at risk of binge eating
(McElroy et al., 2010). For example, for typical anti-
psychotics, the best binding predictor of efficacy (lowest
effective dose) was the ratio of 5-HT2C/D2 affinities
(r = – 0.81), that is negatively related, such that the
activation of 5-HT2C receptors is associated with efficacy
(through D2 receptor antagonism) at lower antipsychotic
doses. The authors suggest that 5-HT2C agonists may
improve psychosis treatment by inhibiting mesolimbic
dopamine activity. By contrast, for atypical antipsychotics,
the relationship between efficacious dose and 5-HT2C
binding is in the opposite direction and significantly
different from the above, although not itself a significant
correlation (r = 0.47). Instead, the best predictors of
efficacious dose were combined affinities comprising
D2(5-HT2A/5-HT1A) (r = 0.80) and D2(5-HT2C/5-HT1A)
(r = 0.78) ratios. It may be that differing doses alter the
balance of postsynaptic and presynaptic D2 and 5-HT
receptor influences on psychosis, so that linear analyses
are too simplistic.
Antipsychotic treatment, particularly atypical drugs, has
been associated with weight gain and even with exaggerated binge eating in patients with eating disorders
(McElroy et al., 2010). Sentissi et al. (2009) compared
eating styles between schizophrenic patients (N = 153)
treated with either class of antipsychotic or no treatment.
Emotional eating tended to be higher in patients treated
with atypical versus typical antipsychotics, but they were
also clearly higher in external eating: BMI was higher in
both treated groups than the controls, and the highest (but
not significantly) in the atypical drug group. These data are
limited, but could suggest an association between emotional and external eating and BMI in patients that is better
addressed by typical antipsychotics.
The importance of serotonergic efferents in regulating
dopamine pathways has echoes in the treatment of binge
eating, both for patients with BN and for those with binge
eating disorder, and may be relevant for any attempt to
treat comfort eating. Depressive symptoms are known to
be common in these eating disorders as well as among
emotional eaters (Ouwens et al., 2009; Konttinen et al.,
2010; Vögele and Gibson, 2010). Therefore, it is not
surprising that the usual pharmacotherapy for BN is
antidepressant treatment, albeit at quite high doses;
furthermore, this approach is used to treat binge eating
disorder, although with less reliability (McElroy et al.,
2010). Clearly, one might in principle consider treatment
strategies for comfort eating, where health was compromised, on the basis of a range of drug classes that have
been shown to have some benefit in binge eating
behaviour, for example targeting orexigenic hormones,
opioid and cannabinoid hedonic pathways, and restraint
of the HPA axis (Pecoraro et al., 2006; Shin et al., 2009).
Moreover, as new therapies, such as peptide hormone
analogues, become licenced for weight management,
some of these may prove to be effective in reducing
emotional eating: for instance, the appetite-stimulating
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Psychobiology of comfort eating Gibson 457
hormone, ghrelin, has been found to mediate stressinduced eating in mice (Chuang et al., 2011), and human
emotional eaters show lack of postingestive decline in
ghrelin compared with nonemotional eaters, although
their baseline ghrelin level was also lower (BMI did not
differ; Raspopow et al., 2010). However, at least for the
time being, nonpharmacological interventions such as
stress management, relaxation, reward substitution and
promotion of physical activity may be preferential
(Konttinen et al., 2010).
Conclusion
Comfort or emotional eating (similar to stress-induced
eating) has been a core theme of explanations for
overeating and obesity for decades. Theories of emotional
eating and related eating styles have gradually evolved so
that there is increasing convergence with animal models of
overeating as well as broader aspects of personality theory,
such as impulsivity and reward sensitivity. Examination of
prevalence suggests that emotional eating is more common
in women than men, and in obese than nonobese
individuals; nevertheless, emotional eaters typically are in
the minority among nonclinical populations. An emerging
theme is one of vulnerability to depression, emotional
dysregulation and a need to find ways of escaping from
negative affect and rumination. A strikingly consistent
finding in both animal and human studies is that comfort
eating (or stress-induced eating) leads to preferential
consumption of sweet, and often fatty, energy-dense foods.
The animal studies suggest that eating such foods confers
protection against stress, evidenced by suppression of the
HPA axis response, possibly mediated by hormonal
consequences of accumulating abdominal obesity. However,
the human data are less clear, in that the lifting of mood
after comfort eating appears to be transient: perhaps stressinduced binges develop in an attempt to prolong mood
improvement, although escape from negative rumination is
as likely. Also, some studies find comfort eaters to be high
reactors to stress, which itself has been linked to abdominal
obesity, although not consistently. However, chronic severe
stress eventually leads to reduced cortisol response to
stress, although this may adversely affect health in the long
term. Not surprisingly, many studies associate comfort
eating with a greater risk of obesity, although not all studies
are consistent in this: one reason may be that, like eating
disorders, individuals rarely express just one pure facet of
eating behaviour independent of others, and comfort eating
might be associated with other personality characteristics
that work against the risk of obesity. Still, chronic
overeating of energy-dense foods, as may occur with
comfort eating, is a risk factor for weight gain, and is likely
to be detrimental to health over the long term.
The extent of heritability of comfort eating remains
uncertain, as different populations of twin studies have
produced quite variable results; nevertheless, genetic
predispositions are likely, including polymorphisms such
as Taq1A that affect the regulation of dopamine pathways.
This variant might contribute to weakened resilience to
stress, as its presence promoted the emergence of comfort
eating in adolescents reporting high levels of parental
control. Similarly, there is increasing evidence that this
polymorphism dissociates the risk of overeating and obesity
in terms of either overactivation or underactivation of
dopamine pathways underlying incentive salience. Emotional eaters in particular seem to show heightened
activation of incentive pathways to food stimuli during
negative mood.
Even assuming that a patient’s health was being
sufficiently compromised by comfort eating, the obvious
complexities of neurochemical systems linking stress,
emotions, impulsivity and reward sensitivity resist reliable proposals for pharmacotherapy using today’s drugs.
At present, cognitive behavioural therapies would be the
first approach. Nevertheless, future developments may
use pharmacogenomic strategies to provide more personally focused drug therapies to prevent comfort eating
from threatening health and well-being.
Acknowledgements
Conflicts of interest
There are no conflicts of interest.
References
Akbaraly TN, Brunner EJ, Ferrie JE, Marmot MG, Kivimaki M, Singh-Manoux A
(2009). Dietary pattern and depressive symptoms in middle age. Br J
Psychiatry 195:408–413.
Anglé S, Engblom J, Eriksson T, Kautiainen S, Saha M-T, Lindfors P, et al. (2009).
Three Factor Eating Questionnaire-R18 as a measure of cognitive restraint,
uncontrolled eating and emotional eating in a sample of young Finnish
females. Int J Behav Nutr Phys Act 6:41–47.
Anschutz DJ, Van Strien T, Engels RC (2008). Exposure to slim images in mass
media: television commercials as reminders of restriction in restrained eaters.
Health Psychol 27:401–408.
Anschutz DJ, Van Strien T, Van De Ven MO, Engels RC (2009). Eating styles and
energy intake in young women. Appetite 53:119–122.
Appelhans BM, Woolf K, Pagoto SL, Schneider KL, Whited MC, Liebman R (2011).
Inhibiting food reward: delay discounting, food reward sensitivity, and palatable
food intake in overweight and obese women. Obesity 19:2175–2182.
Arnow B, Kenardy J, Agras WS (1995). The Emotional Eating Scale: the
development of a measure to assess coping with negative affect by eating. Int
J Eat Disord 18:79–90.
Ashcroft J, Semmler C, Carnell S, van Jaarsveld CH, Wardle J (2008). Continuity
and stability of eating behaviour traits in children. Eur J Clin Nutr 62:985–990.
Bekker MHJ, van de Meerendonk C, Mollerus J (2004). Effects of negative mood
induction and impulsivity on self-perceived emotional eating. Int J Eat Disord
36:461–469.
Bell ME, Bhatnagar S, Liang J, Soriano L, Nagy TR, Dallman MF (2000). Voluntary
sucrose ingestion, like corticosterone replacement, prevents the metabolic
deficits of adrenalectomy. J Neuroendocrinol 12:461–470.
Bello EP, Mateo Y, Gelman DM, Noain D, Shin JH, Low MJ, et al. (2011). Cocaine
supersensitivity and enhanced motivation for reward in mice lacking
dopamine D2 autoreceptors. Nat Neurosci 14:1033–1038.
Bjorntorp P (1999). Neuroendocrine perturbations as a cause of insulin
resistance. Diabetes Metab Res Rev 15:427–441.
Blass EM, Shide DJ, Weller A (1989). Stress-reducing effects of ingesting milk,
sugars, and fats – a developmental perspective. Ann NY Acad Sci 575:292–306.
Boggiano MM, Chandler PC, Viana JB, Oswald KD, Maldonado CR, Wauford PK
(2005). Combined dieting and stress evoke exaggerated responses to
opioids in binge-eating rats. Behav Neurosci 119:1207–1214.
Bohon C, Stice E, Spoor S (2009). Female emotional eaters show abnormalities
in consummatory and anticipatory food reward: a functional magnetic
resonance imaging study. Int J Eat Disord 42:210–221.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
458
Behavioural Pharmacology 2012, Vol 23 No 5&6
Booth DA, Higgs S, Schneider J, Klinkenberg I (2010). Learned liking versus
inborn delight: can sweetness give sensual pleasure or is it just motivating?
Psychol Sci 21:1656–1663.
Bruch H (1961). Social and emotional factors in diet changes. J Am Dent Assoc
63:461–465.
Butland B, Jebb S, Kopelman P, McPherson K, Thomas S, Mardell J, et al. (2007).
Tackling obesities: future choices – project report (foresight). London:
Government Office for Science, Department of Innovation Universities and Skills.
Buwalda B, Blom WA, Koolhaas JM, van Dijk G (2001). Behavioral and
physiological responses to stress are affected by high-fat feeding in
male rats. Physiol Behav 73:371–377.
Cappelleri JC, Bushmakin AG, Gerber RA, Leidy NK, Sexton CC, Lowe MR, et al.
(2009). Psychometric analysis of the Three-Factor Eating Questionnaire-R21:
results from a large diverse sample of obese and non-obese participants. Int J
Obes 33:611–620.
Cartwright M, Wardle J, Steggles N, Simon AE, Croker H, Jarvis MJ (2003).
Stress and dietary practices in adolescents. Health Psychol 22:362–369.
Chuang JC, Perello M, Sakata I, Osborne-Lawrence S, Savitt JM, Lutter M, et al.
(2011). Ghrelin mediates stress-induced food-reward behavior in mice. J Clin
Invest 121:2684–2692.
Dallman MF, Pecoraro N, Akana SF, La Fleur SE, Gomez F, Houshyar H, et al.
(2003). Chronic stress and obesity: a new view of ‘comfort food’. Proc Natl
Acad Sci USA 100:11696–11701.
Davis C, Fox J (2008). Sensitivity to reward and body mass index (BMI): evidence
for a non-linear relationship. Appetite 50:43–49.
De Lauzon B, Romon M, Deschamps V, Lafay L, Borys JM, Karlsson J, et al. (2004).
The Three-Factor Eating Questionnaire-R18 is able to distinguish among
different eating patterns in a general population. J Nutr 134:2372–2380.
De Lauzon-Guillain B, Basdevant A, Romon M, Karlsson J, Borys JM, Charles MA,
et al. (2006). Is restrained eating a risk factor for weight gain in a general
population? Am J Clin Nutr 83:132–138.
Dess NK, Choe S, Minor TR (1998). The interaction of diet and stress in rats:
high-energy food and sucrose treatment. J Exp Psychol Anim Behav Process
24:60.
Dubé L, LeBel JL, Lu J (2005). Affect asymmetry and comfort food consumption.
Physiol Behav 86:559–567.
Elfhag K, Erlanson-Albertsson C (2006). Sweet and fat taste preference in
obesity have different associations with personality and eating behavior.
Physiol Behav 88:61–66.
Elfhag K, Tholin S, Rasmussen F (2008). Consumption of fruit, vegetables, sweets
and soft drinks are associated with psychological dimensions of eating behaviour
in parents and their 12-year-old children. Public Health Nutr 11:914–923.
Epel ES, Moyer AE, Martin CD, Macary S, Cummings N, Rodin J, et al. (1999).
Stress-induced cortisol, mood, and fat distribution in men. Obes Res 7:9–15.
Epel ES, McEwen B, Seeman T, Matthews K, Castellazzo G, Brownell KD, et al.
(2000). Stress and body shape: stress-induced cortisol secretion is consistently
greater among women with central fat. Psychosom Med 62:623–632.
Epel ES, Lapidus R, McEwen B, Brownell K (2001). Stress may add bite to
appetite in women: a laboratory study of stress-induced cortisol and eating
behavior. Psychoneuroendocrinology 26:37–49.
Epel ES, Jimenez S, Brownell K, Stroud L, Stoney C, Niaura R (2004). Are stress
eaters at risk for the metabolic syndrome? Ann NY Acad Sci 1032:208–210.
Epstein LH, Lin H, Carr KA, Fletcher KD (2012). Food reinforcement and obesity.
Psychological moderators. Appetite 58:157–162.
Foster MT, Warne JP, Ginsberg AB, Horneman HF, Pecoraro NC, Akana SF, et al.
(2009). Palatable foods, stress, and energy stores sculpt corticotropinreleasing factor, adrenocorticotropin, and corticosterone concentrations after
restraint. Endocrinol 150:2325–2333.
Fulkerson JA, Sherwood NE, Perry CL, Neumark-Sztainer D, Story M (2004).
Depressive symptoms and adolescent eating and health behaviors: a
multifaceted view in a population-based sample. Prev Med 38:865–875.
Ganley RM (1989). Emotion and eating in obesity – a review of the literature. Int J
Eat Disord 8:343–361.
Geliebter A, Aversa A (2003). Emotional eating in overweight, normal weight, and
underweight individuals. Eat Behav 3:341–347.
Gibson EL, Hamer M, Steptoe A, Wardle J (2008). Effects of work-related stress
and stress proneness on sweet and fatty food consumption in men. Appetite
51:755.
Gibson EL (2006). Emotional influences on food choice: sensory, physiological
and psychological pathways. Physiol Behav 89:53–61.
Gibson EL (2011). Emotional and behavioural aspects of chocolate eating. In:
Preedy VR, Watson RR, Martin CR, editors. Handbook of behavior, food and
nutrition. New York: Springer. pp. 601–620.
Gibson EL, Desmond E (1999). Chocolate craving and hunger state: implications
for the acquisition and expression of appetite and food choice. Appetite
32:219–240.
Gibson EL, Checkley S, Papadopoulos A, Poon L, Daley S, Wardle J (1999).
Increased salivary cortisol reliably induced by a protein-rich midday meal.
Psychosom Med 61:214–224.
Gold PW, Chrousos GP (2002). Organization of the stress system and its
dysregulation in melancholic and atypical depression: high vs low CRH/NE
states. Mol Psychiatry 7:254–275.
Goldfield GS, Legg C (2006). Dietary restraint, anxiety, and the relative
reinforcing value of snack food in non-obese women. Eat Behav 7:323–332.
Goldfield GS, Adamo KB, Rutherford J, Legg C (2008). Stress and the relative
reinforcing value of food in female binge eaters. Physiol Behav 93:579–587.
Greeno CG, Wing RR (1994). Stress-induced eating. Psychol Bull 115:444–464.
Hagan MM, Wauford PK, Chandler PC, Jarrett LA, Rybak RJ, Blackburn K (2002).
A new animal model of binge eating: key synergistic role of past caloric
restriction and stress. Physiol Behav 77:45–54.
Haynes C, Lee MD, Yeomans MR (2003). Interactive effects of stress, dietary
restraint, and disinhibition on appetite. Eat Behav 4:369–383.
Heatherton TF, Baumeister RF (1991). Binge eating as escape from selfawareness. Psychol Bull 110:86–108.
Herman CP, Mack D (1975). Restrained and unrestrained eating. J Pers 43:647–660.
Johnson PM, Kenny PJ (2010). Dopamine D2 receptors in addiction-like reward
dysfunction and compulsive eating in obese rats. Nat Neurosci 13:635–641.
Kandiah J, Yake M, Jones J, Meyer M (2006). Stress influences appetite and
comfort food preferences in college women. Nutr Res 26:118–123.
Kaplan HI, Kaplan HS, Leder HL (1957). The psychosomatic management of
obesity. NY State J Med 57:2815–2826.
Karlsson J, Persson LO, Sjostrom L, Sullivan M (2000). Psychometric properties
and factor structure of the Three-Factor Eating Questionnaire (TFEQ) in
obese men and women. Results from the Swedish Obese Subjects
(SOS) study. Int J Obes 24:1715–1725.
Keskitalo K, Tuorila H, Spector TD, Cherkas LF, Knaapila A, Kaprio J, et al. (2008).
The Three-Factor Eating Questionnaire, body mass index, and responses to
sweet and salty fatty foods: a twin study of genetic and environmental
associations. Am J Clin Nutr 88:263–271.
Kinzig KP, Hargrave SL, Honors MA (2008). Binge-type eating attenuates
corticosterone and hypophagic responses to restraint stress. Physiol Behav
95:108–113.
Kisler VA, Corcoran KJ (1997). Effects of negative outcome on food consumption
in college women with and without troubled eating patterns. Addict Behav
22:461–467.
Konttinen H, Männistö S, Sarlio-Lähteenkorva S, Silventoinen K, Haukkala A
(2010). Emotional eating, depressive symptoms and self-reported food
consumption. A population-based study. Appetite 54:473–479.
Kräuchi K, Reich S, WirzJustice A (1997). Eating style in seasonal affective
disorder: Who will gain weight in winter? Compr Psychiatry 38:80–87.
La Fleur SE, Houshyar H, Roy M, Dallman MF (2005). Choice of lard, but not total
lard calories, damps adrenocorticotropin responses to restraint. Endocrinol
146:2193–2199.
Laitinen J, Ek E, Sovio U (2002). Stress-related eating and drinking behavior and
body mass index and predictors of this behavior. Prev Med 34:29–39.
Lanfranco F, Giordano R, Pellegrino M, Gianotti L, Ramunni J, Picu A, et al.
(2004). Free fatty acids exert an inhibitory effect on adrenocorticotropin and
cortisol secretion in humans. J Clin Endocrinol Metab 89:1385–1390.
Le Foll B, Gallo A, Le Strat Y, Lu L, Gorwood P (2009). Genetics of dopamine receptors
and drug addiction: a comprehensive review. Behav Pharmacol 20:1–17.
Lluch A, Herbeth B, Mejean L, Siest G (2000). Dietary intakes, eating style and
overweight in the Stanislas Family Study. Int J Obes 24:1493–1499.
Lowe MR, Fisher EB Jr (1983). Emotional reactivity, emotional eating, and
obesity: a naturalistic study. J Behav Med 6:135–149.
Lowe MR, Kral TV (2006). Stress-induced eating in restrained eaters may not be
caused by stress or restraint. Appetite 46:16–21.
Loxton NJ, Dawe S, Cahill A (2011). Does negative mood drive the urge to eat?
The contribution of negative mood, exposure to food cues and eating style.
Appetite 56:368–374.
Macht M (2008). How emotions affect eating: a five-way model. Appetite 50:1–11.
Macht M, Mueller J (2007). Immediate effects of chocolate on experimentally
induced mood states. Appetite 49:667–674.
Marazziti D, Catena Dell’osso M (2008). The role of oxytocin in neuropsychiatric
disorders. Curr Med Chem 15:698–704.
McCann BS, Warnick GR, Knopp RH (1990). Changes in plasma lipids and
dietary intake accompanying shifts in perceived workload and stress.
Psychosom Med 52:97–108.
McElroy SL, Guerdjikova AI, O’Melia AM, Mori N, Keck PE Jr (2010).
Pharmacotherapy of the eating disorders. In: Agras WS, editor. The Oxford
handbook of eating disorders. Oxford: Oxford University Press. pp. 417–451.
Mercer ME, Holder MD (1997). Antinociceptive effects of palatable sweet
ingesta on human responsivity to pressure pain. Physiol Behav 61:311–318.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Psychobiology of comfort eating Gibson 459
Michaud C, Kahn JP, Musse N, Burlet C, Nicolas JP, MeJean L (1990).
Relationships between a critical life event and eating behaviour in high-school
students. Stress Med 6:57–64.
Michener W, Rozin P (1994). Pharmacological versus sensory factors in the
satiation of chocolate craving. Physiol Behav 56:419–422.
Michopoulos V, Toufexis D, Wilson ME (2012). Social stress interacts with diet
history to promote emotional feeding in females. Psychoneuroendocrinology
Available at: http://dx.doi.org/10.1016/j.psyneuen.2012.02.002.
Moyer AE, Rodin J, Grilo CM, Cummings N, Larson LM, Rebuffe-Scrive M (1994).
Stress-induced cortisol response and fat distribution in women. Obes Res
2:255–262.
Newman E, O’Connor DB, Conner M (2007). Daily hassles and eating behaviour:
the role of cortisol reactivity status. Psychoneuroendocrinology 32:125–132.
Nguyen-Rodriguez ST, Unger JB, Spruijt-Metz D (2009). Psychological determinants of emotional eating in adolescence. Eat Disord 17:211–224.
Nolan LJ, Halperin LB, Geliebter A (2010). Emotional Appetite Questionnaire.
construct validity and relationship with BMI. Appetite 54:314–319.
O’Connor DB, Jones F, Conner M, McMillan B, Ferguson E (2008). Effects of daily
hassles and eating style on eating behavior. Health Psychol 27:S20–S31.
O’Donnell L, O’Meara N, Owens D, Johnson A, Collins P (1987). Plasma
catecholamines and lipoproteins in chronic psychological stress. J R Soc
Med 80:339–342.
Oliver G, Wardle J (1999). Perceived effects of stress on food choice. Physiol
Behav 66:511–515.
Oliver G, Wardle J, Gibson EL (2000). Stress and food choice: a laboratory study.
Psychosom Med 62:853–865.
Ott V, Friedrich M, Prilop S, Lehnert H, Jauch-Chara K, Born J, et al. (2011). Food
anticipation and subsequent food withdrawal increase serum cortisol in
healthy men. Physiol Behav 103:594–599.
Ouwens MA, van Strien T, van Leeuwe JFJ (2009). Possible pathways between
depression, emotional and external eating. A structural equation model.
Appetite 53:245–248.
Parker G, Crawford J (2007). Chocolate craving when depressed: a personality
marker. Br J Psychiatry 191:351–352.
Parker G, Parker I, Brotchie H (2006). Mood state effects of chocolate. J Affect
Disord 92:149–159.
Pecina S, Schulkin J, Berridge KC (2006). Nucleus accumbens corticotropinreleasing factor increases cue-triggered motivation for sucrose reward:
paradoxical positive incentive effects in stress? BMC Biol 4:8.
Pecoraro N, Reyes F, Gomez F, Bhargava A, Dallman MF (2004). Chronic stress
promotes palatable feeding, which reduces signs of stress: feedforward and
feedback effects of chronic stress. Endocrinol 145:3754–3762.
Pecoraro N, Dallman MF, Warne JP, Ginsberg AB, Laugero KD, la Fleur SE, et al.
(2006). From Malthus to motive: how the HPA axis engineers the phenotype,
yoking needs to wants. Prog Neurobiol 79:247–340.
Phillips AL, Gibson EL, Slade L (2012). Anxious attachment predicts uncontrolled
eating independently of emotional eating. Appetite: (Abstract) [in press].
Pollard TM, Steptoe A, Canaan L, Davies GJ, Wardle J (1995). Effects of
academic examination stress on eating behavior and blood lipid levels. Int J
Behav Med 2:299–320.
Raspopow K, Abizaid A, Matheson K, Anisman H (2010). Psychosocial stressor
effects on cortisol and ghrelin in emotional and non-emotional eaters:
influence of anger and shame. Horm Behav 58:677–684.
Richtand NM, Welge JA, Logue AD, Keck PE Jr., Strakowski SM, McNamara RK
(2007). Dopamine and serotonin receptor binding and antipsychotic efficacy.
Neuropsychopharmacology 32:1715–1726.
Roberts C, Troop N, Connan F, Treasure J, Campbell IC (2007). The effects of
stress on body weight: biological and psychological predictors of change
in BMI. Obesity 15:3045–3055.
Rose N, Koperski S, Golomb BA (2010). Mood food: chocolate and depressive
symptoms in a cross-sectional analysis. Arch Intern Med 170:699–703.
Rozin P, Levine E, Stoess C (1991). Chocolate craving and liking. Appetite
17:199–212.
Ruderman AJ (1983). The restraint scale: a psychometric investigation. Behav
Res Ther 21:253–258.
Schachter S (1968). Obesity and eating. Internal and external cues differentially
affect the eating behavior of obese and normal subjects. Science 161:
751–756.
Selby EA, Anestis MD, Joiner TE (2008). Understanding the relationship between
emotional and behavioral dysregulation: emotional cascades. Behav Res
Ther 46:593–611.
Sentissi O, Viala A, Bourdel MC, Kaminski F, Bellisle F, Olie JP, et al. (2009). Impact
of antipsychotic treatments on the motivation to eat: preliminary results in 153
schizophrenic patients. Int Clin Psychopharmacol 24:257–264.
Serlachius A, Hamer M, Wardle J (2007). Stress and weight change in university
students in the United Kingdom. Physiol Behav 92:548–553.
Shin AC, Zheng H, Berthoud HR (2009). An expanded view of energy
homeostasis: neural integration of metabolic, cognitive, and emotional drives
to eat. Physiol Behav 97:572–580.
Smit HJ, Blackburn RJ (2005). Reinforcing effects of caffeine and theobromine as
found in chocolate. Psychopharmacology (Berl) 181:101–106.
Smit HJ, Gaffan EA, Rogers PJ (2004). Methylxanthines are the psychopharmacologically active constituents of chocolate. Psychopharmacology
(Berl) 176:412–419.
Snoek HM, Van Strien T, Janssens JMAM, Engels RCME (2007). Emotional,
external, restrained eating and overweight in Dutch adolescents. Scand J
Psychol 48:23–32.
Steptoe A, Wardle J (2005). Cardiovascular stress responsivity, body mass and
abdominal adiposity. Int J Obes 29:1329–1337.
Steptoe A, Lipsey Z, Wardle J (1998). Stress, hassles and variations in alcohol
consumption, food choice and physical exercise: a diary study. Br J Health
Psychol 3:51–63.
Steptoe A, Gibson EL, Hamer M, Wardle J (2007). Neuroendocrine and
cardiovascular correlates of positive affect measured by ecological momentary
assessment and by questionnaire. Psychoneuroendocrinology 32:56–64.
Stice E, Spoor S, Bohon C, Small DM (2008a). Relation between obesity and
blunted striatal response to food is moderated by TaqIA A1 allele. Science
322:449–452.
Stice E, Spoor S, Bohon C, Veldhuizen MG, Small DM (2008b). Relation of
reward from food intake and anticipated food intake to obesity: a functional
magnetic resonance imaging study. J Abnorm Psychol 117:924–935.
Stice E, Yokum S, Bohon C, Marti N, Smolen A (2010). Reward circuitry
responsivity to food predicts future increases in body mass: moderating
effects of DRD2 and DRD4. Neuroimage 50:1618–1625.
Stunkard A (1959). Obesity and the denial of hunger. Psychosom Med 21:281–290.
Stunkard AJ, Messick S (1985). The Three-Factor Eating Questionnaire to measure
dietary restraint, disinhibition and hunger. J Psychosom Res 29:71–83.
Sung J, Lee K, Song YM, Lee MK, Lee DH (2010). Heritability of eating behavior
assessed using the DEBQ (Dutch Eating Behavior Questionnaire) and
weight-related traits: the Healthy Twin Study. Obesity 18:1000–1005.
Tholin S, Rasmussen F, Tynelius P, Karlsson J (2005). Genetic and environmental
influences on eating behavior: the Swedish Young Male Twins Study. Am J
Clin Nutr 81:564–569.
Thompson J, Thomas N, Singleton A, Piggott M, Lloyd S, Perry EK, et al. (1997).
D2 dopamine receptor gene (DRD2) Taq1 A polymorphism: reduced
dopamine D2 receptor binding in the human striatum associated with the
A1 allele. Pharmacogenetics 7:479–484.
Tice DM, Bratslavsky E, Baumeister RF (2001). Emotional distress regulation
takes precedence over impulse control: if you feel bad, do it! J Pers Soc
Psychol 80:53–67.
Tomiyama AJ, Dallman MF, Epel ES (2011). Comfort food is comforting to those
most stressed: evidence of the chronic stress response network in high
stress women. Psychoneuroendocrinology 36:1513–1519.
Unusan N (2006). Linkage between stress and fruit and vegetable intake among
university students: an empirical analysis on Turkish students. Nutr Res
26:385–390.
Van Jaarsveld CH, Fidler JA, Steptoe A, Boniface D, Wardle J (2009). Perceived
stress and weight gain in adolescence: a longitudinal analysis. Obesity
17:2155–2161.
Van Strien T, Koenders PG (2012). How do life style factors relate to general
health and overweight? Appetite 58:265–270.
Van Strien T, Van de Laar FA (2008). Intake of energy is best predicted by
overeating tendency and consumption of fat is best predicted by dietary
restraint: a 4-year follow-up of patients with newly diagnosed type 2 diabetes.
Appetite 50:544–547.
Van Strien T, Frijters JE, Roosen RG, Knuiman-Hijl WJ, Defares PB (1985). Eating
behavior, personality traits and body mass in women. Addict Behav 10:333–343.
Van Strien T, Frijters JER, Bergers GPA, Defares PB (1986). The Dutch Eating
Behavior Questionnaire (DEBQ) for assessment of restrained, emotional, and
external eating behavior. Int J Eat Disord 5:295–315.
Van Strien T, Herman CP, Verheijden MW (2009). Eating style, overeating, and
overweight in a representative Dutch sample. Does external eating play
a role? Appetite 52:380–387.
Van Strien T, Snoek HM, van der Zwaluw CS, Engels RC (2010). Parental control
and the dopamine D2 receptor gene (DRD2) interaction on emotional eating
in adolescence. Appetite 54:255–261.
Van Strien T, Herman CP, Anschutz DJ, Engels RC, de Weerth C (2012).
Moderation of distress-induced eating by emotional eating scores. Appetite
58:277–284.
Vicennati V, Pasqui F, Cavazza C, Garelli S, Casadio E, di Dalmazi G, et al.
(2011). Cortisol, energy intake, and food frequency in overweight/
obese women. Nutrition 27:677–680.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
460 Behavioural Pharmacology 2012, Vol 23 No 5&6
Vitaliano PP, Scanlan JM, Zhang J, Savage MV, Hirsch IB, Siegler IC (2002). A
path model of chronic stress, the metabolic syndrome, and coronary heart
disease. Psychosom Med 64:418–435.
Vögele C, Gibson EL (2010). Mood, emotions and eating disorders. In: Agras
WS, editor. The Oxford handbook of eating disorders. Oxford: Oxford
University Press. pp. 180–205.
Volkow ND, Wang GJ, Baler RD (2011). Reward, dopamine and the control of
food intake: implications for obesity. Trends Cogn Sci 15:37–46.
Wallis DJ, Hetherington MM (2004). Stress and eating: the effects of ego-threat
and cognitive demand on food intake in restrained and emotional eaters.
Appetite 43:39–46.
Wansink B, Cheney MM, Chan N (2003). Exploring comfort food preferences
across age and gender. Physiol Behav 79:739–747.
Ward A, Mann T (2000). Don’t mind if I do: disinhibited eating under
cognitive load. J Pers Soc Psychol 78:753–763.
Wardle J (1987). Eating style: a validation study of the Dutch Eating Behaviour
Questionnaire in normal subjects and women with eating disorders. J
Psychosom Res 31:161–169.
Wardle J, Gibson EL (2002). Impact of stress on diet: processes and
implications. In: Stansfeld SA, Marmot MG, editors. Stress and the heart:
psychosocial pathways to coronary heart disease. London: BMJ Books.
pp. 124–149.
Wardle J, Marsland L, Sheikh Y, Quinn M, Fedoroff I, Ogden J (1992). Eating style
and eating behaviour in adolescents. Appetite 18:167–183.
Wardle J, Steptoe A, Oliver G, Lipsey Z (2000). Stress, dietary restraint and food
intake. J Psychosom Res 48:195–202.
Wardle J, Sanderson S, Guthrie CA, Rapoport L, Plomin R (2002). Parental
feeding style and the inter-generational transmission of obesity risk. Obes
Res 10:453–462.
Wardle J, Chida Y, Gibson EL, Whitaker KL, Steptoe A (2011). Stress and
adiposity: a meta-analysis of longitudinal studies. Obesity 19:771–778.
Waters A, Hill A, Waller G (2001). Bulimics’ responses to food cravings: is bingeeating a product of hunger or emotional state? Behav Res Ther 39:877–886.
Weidner G, Kohlmann CW, Dotzauer E, Burns LR (1996). The effects of
academic stress on health behaviors in young adults. Anxiety Stress Coping
9:123–133.
Weinstein SE, Shide DJ, Rolls BJ (1997). Changes in food intake in response to
stress in men and women: psychological factors. Appetite 28:7–18.
Westenhoefer J, Broeckmann P, Munch AK, Pudel V (1994). Cognitive control of
eating behaviour and the disinhibition effect. Appetite 23:27–41.
Westover AN, Marangell LB (2002). A cross-national relationship between sugar
consumption and major depression? Depress Anxiety 16:118–120.
Wilkinson LL, Rowe AC, Bishop RJ, Brunstrom JM (2010). Attachment anxiety,
disinhibited eating, and body mass index in adulthood. Int J Obes 34:
1442–1445.
Willner P, Benton D, Brown E, Cheeta S, Davies G, Morgan J, et al. (1998).
‘Depression’ increases ‘craving’ for sweet rewards in animal and human
models of depression and craving. Psychopharmacology (Berl) 136:
272–283.
Yeomans MR, Coughlan E (2009). Mood-induced eating. Interactive effects of
restraint and tendency to overeat. Appetite 52:290–298.
Yeomans MR, Leitch M, Mobini S (2008). Impulsivity is associated with the
disinhibition but not restraint factor from the Three Factor Eating Questionnaire. Appetite 50:469–476.
Zellner DA, Loaiza S, Gonzalez Z, Pita J, Morales J, Pecora D, et al. (2006). Food
selection changes under stress. Physiol Behav 87:789–793.
Zellner DA, Saito S, Gonzalez J (2007). The effect of stress on men’s food
selection. Appetite 49:696–699.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.