Content uploaded by Tomoko Udo
Author content
All content in this area was uploaded by Tomoko Udo on Nov 19, 2018
Content may be subject to copyright.
Modeling the Effects of Positive and Negative Mood on the
Ability to Resist Eating in Obese and Non-obese Individuals
Tomoko Udo, PhD1, Carlos M. Grilo, PhD1, Kelly D. Brownell, PhD2, Andrea H. Weinberger,
PhD1, Ralph J. DiLeone, PhD1, and Sherry A. McKee, PhD1
1Department of Psychiatry, Yale University School of Medicine, New Haven, CT 06519, USA
2Rudd Center for Food Policy and Obesity, Yale University, New Haven, CT 06519, USA
Abstract
This pilot study adapted a well-established drug self-administration paradigm to examine the
effects of mood induction on the ability to resist high-calorie foods and subsequent food
consumption differently in 15 obese individuals (40.0% women, BMI: 35.1±3.70) and 15 non-
obese individuals (46.7% women, BMI: 23.0±1.96). Participants completed two laboratory
sessions (positive vs. negative mood conditions) consisting of 3-hour food deprivation, followed
by mood induction, and a 3-hour ad-lib eating period, where they were asked to choose between
favorite high-calorie snacks and monetary reinforcement. Obese individuals were less able to
resist eating and increased high-calorie food consumption during the positive mood condition than
the negative condition. Non-obese individuals were less able to resist eating during the negative
mood condition than the positive condition, but their total consumption was not affected by the
mood conditions. In obese individuals, food craving was associated with less ability to resist
eating and greater calorie consumption during the negative mood condition. This is the first study
to experimentally demonstrate that mood state may increase vulnerability to food consumption by
reducing the ability to resist eating. The ability to resist eating may be a novel dimension of eating
behaviors that has a significant contribution to understanding mood-eating relationships.
Keywords
Obesity; mood; eating behaviors; high-calorie foods; food craving
1. Introduction
Obesity is a leading health risk for chronic diseases and conditions in the United States, such
as cardiovascular diseases, type-II diabetes, and certain cancers (Ogden, Yanovski, Carroll,
& Flegal, 2007). In 2007-2008, the prevalence of obesity was 33.8% (Flegal, Carroll,
Ogden, & Curtin, 2010), with annual medical expenditure attributable to obesity estimated at
$147 billion (Flegal, et al., 2010). The modest efficacy of nutrition- and exercise-related
interventions highlights the need for new approaches to control body weight. An abundance
of food, particularly high-calorie palatable foods, and overeating have been argued to be
partly responsible for the current obesity epidemic (Pandit, de Jong, Vanderschuren, &
Adan, 2011). A recent epidemiology study has indeed found increasing trends in frequency
of snacking, energy density of snacks, and the contribution of snacks to total calorie
Corresponding Author: Sherry A. McKee, Ph.D., Department of Psychiatry, Yale University School of Medicine, 2 Church St. South,
Suite 109, New Haven, CT, USA, 06519. Phone: 1-203-737-3529, Fax: 1-203-737-4243, sherry.mckee@yale.edu.
NIH Public Access
Author Manuscript
Eat Behav
. Author manuscript; available in PMC 2013 April 23.
Published in final edited form as:
Eat Behav
. 2013 January ; 14(1): 40–46. doi:10.1016/j.eatbeh.2012.10.010.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
consumption (Piernas & Popkin, 2010). It is therefore critical to develop effective obesity
intervention strategies that focus on reduction of high-calorie food consumption.
While multiple factors contribute to obesity, overeating due to loss of control over food
intake is important. Eating behaviors are highly cue-dependent, and mood states are thought
to influence overeating behaviors. When queried about relapse situations, overweight dieters
frequently reported both temptations to overeat and overeating when experiencing positive
and negative mood states (Grilo, Shiffman, & Wing, 1989). In human laboratory studies,
negative mood states, including stress, have been shown to alter food preference to highly-
palatable foods (Oliver, Wardle, & Gibson, 2000) and to have disinhibiting effects on eating,
particularly in individuals with tendencies to engage in restrained (Habhab, Sheldon, &
Loeb, 2009; Lattimore & Caswell, 2004; Schotte, Cools, & McNally, 1990; Wallis &
Hetherington, 2004) or emotional eating (Oliver, et al., 2000; Wallis & Hetherington, 2004,
2009). Experimental studies have also demonstrated that increased positive mood enhanced
consumption of palatable food in normal weight individuals (Cools, Schotte, & McNally,
1992; Yeomans & Coughlan, 2009), although the effects of negative mood appeared to be
more potent (Yeomans & Coughlan, 2009). Collectively, these studies across diverse subject
groups and weight categories suggest that mood, regardless of valence, can trigger
overeating and loss of control over eating. Importantly, Grilo et al. (1989) found that coping
attempts may prevent overeating or overcome temptations to overeat in response to such
emotional cues; however, those findings are weakened by reliance on retrospective self-
report data and require experimental manipulation in a laboratory setting to arrive at firmer
conclusions. We are unaware of existing human laboratory studies that have modeled the
ability to resist eating high-calorie palatable foods, which could be a crucial component in
promoting weight loss efforts and healthy weight maintenance to counteract the current
obesity epidemic.
For the current study, we adapted a well-established human drug self-administration
paradigm that we had previously developed, which assesses the effects of positive and
negative mood induction on the ability to resist smoking (McKee, et al., 2011). In this
paradigm, following mood induction, a lighter, cigarette, and ashtray are presented, and
participants are asked to resist smoking and are offered an increasing amount of monetary
compensation the longer they can resist smoking. The inclusion of money as an alternative
reinforcement was a critical part of this model to provide incentive for not smoking, and to
enhance the likelihood that the effects of stress on the reinforcing value of smoking would
be detected (McKee, 2009). The latency to start smoking (i.e., ability to resist) is the primary
outcome measure, following which ad-libitum smoking is evaluated. Using this unique
paradigm, our laboratory has reliably demonstrated that negative mood reduces the ability to
resist smoking and leads to more intense smoking behaviors (e.g., increased puffs, shorter
inter-puff interval, and greater peak puff velocity), compared to positive mood, in daily
smokers (McKee, et al., 2011). Adapting this model to examine eating behavior will provide
a framework to evaluate the effects of a mood manipulation on the ability to resist eating
high-calorie, palatable food. Similar to the smoking-lapse model, money was used as an
alternative reinforcer to provide a sensitive test of the relative reinforcing value of high-
calorie foods.
The goal of this pilot study was to examine the effects of both negative and positive mood
induction on two important aspects of overeating behavior in obese and non-obese
individuals: 1) the failure of the ability to resist eating high-calorie food and 2) subsequent
ad-lib eating of high-calorie food. It is unknown whether mood manipulations will affect the
ability to resist eating and subsequent food consumption differently in obese and non-obese
individuals. We also examined whether the relationship between mood-induced eating
Udo et al. Page 2
Eat Behav
. Author manuscript; available in PMC 2013 April 23.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
behaviors, self-reported positive and negative emotions, and food craving differed in obese
and non-obese individuals.
2. Methods
2.1. Participants
A total of 30 participants (mean age = 36.8 ± 12.6 years old; 43.3% women) completed the
study. Eligible participants had to be between 18 to 65 years of age and have a Body Mass
Index (BMI) between 30 and 45 (obese group) or below 30 (non-obese group) (CDC, 2011).
Exclusion criteria included: current diagnosis of Axis I psychiatric disorders (except nicotine
dependence), current diagnosis of anorexia nervosa and/or bulimia nervosa, significant
medical conditions, including metabolic disorders (e.g., diabetes, thyroid problems, and
abnormal fasting glucose value), and current use of psychotropic or illicit drugs. The ethnic
composition of the sample was non-Hispanic White (56.7%) and non-Hispanic African-
American (43.3%). The majority of participants had completed high school (80.0%) and
reported an income of less than $60,000 (65.5%).
2.2. Procedures
The experimental protocol was approved by the Yale Human Investigation Committee, and
the procedures were in compliance with the Declaration of Helsinki for human subjects.
Written informed consent was obtained from all the participants.
2.2.1. Intake assessment—The Structured Clinical Interview for DSM-IV Axis I
Psychiatric Disorders (SCID; First, Spitzer, Gibbon, & Williams, 1995) was used to exclude
individuals who met diagnostic criteria for current psychiatric disorders (except nicotine
dependence), including eating disorders. Participants were screened for metabolic disorders
assessed with basic blood chemistry tests. Eligible participants were then scheduled for the
laboratory sessions.
2.2.2. Script development session—A personalized guided imagery procedure was
used for negative and positive mood induction (Sinha, 2009). The imagery script for
negative mood induction was developed by having participants provide a detailed
description about a recent negative mood-inducing experience occurring in the past 6
months that they perceived as “most stressful.” Perceived stress was rated on a 10-point
Likert scale where 1 =
not at all stressful
and 10 =
the most stress they recently felt in their
life
. Ratings were made relative to other experiences in the past 6 months. Only situations
with perceived stress rating of ≥ 8 were accepted as appropriate for script development.
Examples of stressful experiences were marital conflict or losing employment. The imagery
script for positive mood induction was developed by having participants describe a personal
positive mood-inducing situation that only involved themselves, such as sitting at the beach
or reading in the park. Scripts were developed by a PhD-level clinician, audio-taped for
presentation, and were approximately 5 minutes in length (see also McKee et al., 2011). At
the end of this session, the participants completed an assessment of their preferred high-
calorie sweet and salty foods.
2.2.3. Laboratory session—Each participant individually completed two 9-hour
laboratory sessions (positive imagery vs. negative imagery session). The order of the
sessions was counterbalanced. Participants were compensated $390 for completing the entire
study. The average time between two laboratory sessions was 7.0 ± 2.9 days.
Laboratory sessions started at 7:30 AM. Participants were instructed not to eat past 10 PM
the night before the laboratory session. After providing a urine drug screen and baseline
Udo et al. Page 3
Eat Behav
. Author manuscript; available in PMC 2013 April 23.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
assessments of self-report emotion ratings and food craving, participants received a
standardized breakfast to control caloric intake and the time since last food consumption at
8:00 AM. This was followed by 3 hours of food deprivation. During the food deprivation
period, participants were allowed to watch TV and read. Emotion ratings and food cravings
were assessed before the imagery task at 10:45 AM (i.e., pre-imagery assessment).
At 10:55 AM, participants were instructed to clear their mind of any worrying thoughts and
to focus on deep breathing. At 11:00 AM, they were instructed as follows for the guided
imagery: “You will soon hear a situation being described to you. Your task is to close your
eyes and imagine yourself in the situation being described, ‘as if’ it were happening right
now. Allow yourself to become completely involved in the situation, by involving your
mind and body in actually doing what is being described. Continue imagining until you are
asked to stop.” Then, the participant listened to the script (negative or positive) over
headphones. Following the script, participants rated how clearly they were able to imagine
the scene on a 140-mm visual analog scale (VAS). Mean vividness ratings were 113.77 ±
19.71 for the negative mood imagery and 119.00 ± 15.63 for the positive mood imagery (
p
> .05). At 11:05 AM, post-imagery assessments of mood and food craving were completed.
At 11:30 PM, ad-lib food consumption began by presenting the individual participant’s
preferred three choices of high-calorie sweet foods (e.g., cookies, snack cakes, chocolate
pudding) and three choices of high-calorie salty foods (e.g., potato chips, popcorn with
butter, peanuts). Snacks were portioned to five servings of each item, with possible caloric
intake ranging from 3950 kcal to 5750 kcal, depending on the combination of food choices.
Participants were told that they could start eating at any time they wish over the next three
hours. However, for each minute they delayed or “resisted” eating, they would receive
monetary rewards. Monetary reinforcement was scheduled as $0.20/minute for the first
hour, $0.10/minute for the second hour, and $0.05/min for the third hour. The participants
could thus earn up to $21 over the 3-hour period. We determined the value of monetary
reinforcement based on our previous smoking-lapse models, and chose this de-escalating
schedule of reinforcement to model how the ability to resist eating high-calorie food
decreases over time. The end of delay period was marked when the participants could no
longer resist and decided to start eating, and their emotion ratings and food craving were
assessed at this time (immediately prior to food consumption). Participants were then
allowed to eat as much as they wished until the end of the 3-hour ad-lib eating period. After
the 3-hour ad-lib eating period, the food was removed, and thus participants no longer had
access to the food. Participants were required to remain in the lab for an additional 2 hours
to add a response cost if they chose not to consume any food during this period.
2.3. Measures
2.3.1. Eating behaviors—Ability to resist eating was calculated as the latency between
the presentation of the preferred foods and the decision to start eating. When subjects
decided to start eating, time was recorded in minutes and seconds (range = 0 – 180 min). In
addition, total calories consumed (sweet and salty foods combined) over the 3-hour ad-lib
eating period were recorded.
2.3.2. Self-report measures—The Revised Differential Emotion Scale (RDES; Izard,
1972) was used to measure positive and negative emotional states throughout the session.
The RDES consists of 30 negative (e.g., irritated, distressed, upset, sad or depressed, angry)
and positive (e.g., pleasant, happy, joyful, relaxed, comfortable) emotion words, and
participants were asked to rate them on a 100-mm VAS. To measure food craving, we
developed a food craving scale based on a 10-item Tiffany Questionnaire of Smoking Urges
(QSU-Brief; Cox, Tiffany, & Christen, 2001) by replacing the word “a cigarette” with
Udo et al. Page 4
Eat Behav
. Author manuscript; available in PMC 2013 April 23.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
“food,” and the word “smoke” with “eat.” Although the factor structure of this food craving
scale has not been tested, corresponding with the original QSU-Brief, we calculated a global
food craving score (Chronbach’s alpha = 0.97). In addition, the Eating Disorder
Examination Questionnaire with Instructions (EDE-Q-I; Goldfein, Devlin, & Kamenetz,
2005), a psychometrically sound measure with good test-retest reliability (Reas, Grilo, &
Masheb, 2006), was used to measure if participants had recently engaged in dieting (Item 1:
“how many days out of the past 28 days have you been deliberately trying to limit the
amount of food you eat to influence your shape or weight?”) or experienced binge eating
episodes (Item 15: “over the past 28 days, how many days have such episodes of overeating
occurred [i.e., you have eaten an usually large amount of food and have had a sense of loss
of control at the time]?). The Dutch Eating Behavior Questionnaire (DEBQ; Van Strien,
Frijters, Bergers, & Defares, 1986), a scale of eating habits with demonstrated validity (Van
Strien et al., 1986; Wardle, 1987), was used to characterize restraint and emotional eating in
our sample.
2.4. Statistical Analysis
Repeated measures analysis of covariance (ANCOVA) was conducted to compare the
latency to start eating between positive and negative mood imagery conditions (within-
subject effect) by obesity status (between-subject effect). Repeated measures ANCOVA was
also used to examine the within-subject effect of the imagery condition and time (pre-mood
induction, post-mood induction, and end of delay [i.e., made the decision to eat, but had not
yet started to eat]) on positive and negative emotion ratings, as well as global food craving.
When participants successfully resisted eating for the entire 3 hours, values for emotion
ratings and food craving scores were censored at the end of the 3-hour eating period. In
addition, Pearson correlation analyses were conducted, separately for obese and non-obese
groups, to examine the relationship of eating behaviors with emotion ratings and food
craving. In all analyses, gender, session order, age, eating behaviors in the past 28 days (i.e.,
dieting and binge eating), and eating habits (i.e., restraint and emotional eating) were
evaluated as covariates and were retained only if they reduced residual variance.
3. Results
3.1. Sample characteristics
Table 1 summarizes the sample characteristics by obesity status. There were no significant
differences in demographic characteristics between obese and non-obese individuals, except
that the obese group was significantly older and, as designed, had greater BMI scores.
3.2. Manipulation check
Positive emotion ratings (imagery condition-by-time;
F
[2, 52] = 33.75,
p
< .01, partial η2 =
0.57 [large], with a significant quadratic trend) and negative emotion ratings (imagery
condition-by-time;
F
[2, 52] = 13.94,
p
< .01, partial η2 = 0.35 [large], with a significant
quadratic trend) differed by imagery condition, regardless of obesity status. Confirming the
imagery manipulation, positive emotion ratings increased after positive mood induction and
decreased after negative mood induction (Figure 1, upper); negative emotion ratings
increased after negative mood induction, but showed no changes after positive mood
induction (Figure 1, lower). Both positive and negative emotion ratings returned to the pre-
imagery level at the end of delay when participants made the decision to eat, but had not yet
started to eat.
Udo et al. Page 5
Eat Behav
. Author manuscript; available in PMC 2013 April 23.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
3.3. Eating behaviors
Obesity status interacted with imagery condition on latency to eat,
F
(1, 24) = 7.39,
p
< .05,
partial η2 = 0.24 (large).1 Obese individuals were less able to resist eating in the positive
mood condition compared to the negative mood condition (Figure 2, upper). Conversely,
non-obese individuals were less able to resist eating in the negative mood condition
compared to the positive mood condition (Figure 2, upper). Obesity status also interacted
with imagery condition on total calorie consumption,
F
(1, 24) = 4.36,
p
< .05, partial η2 =
0.15 (large).2 Obese individuals consumed significantly more calories compared to non-
obese individuals, particularly after the positive mood induction (Figure 2, lower). Obese
individuals also consumed more calories in the positive mood condition than in the negative
mood condition. No other significant between- and within-subject effects were found for
“resistance to eat” or total calories consumed (all
p
> .05).
3.4. Food craving
For global food craving scores, there was a significant imagery condition-by-time
interaction (Figure 3),
F
(2, 25) = 3.66 (multivariate test),
p
< .05, partial η2 = 0.23 (large).
Regardless of obesity status, food craving ratings increased after negative mood induction,
and showed little changes after positive mood induction. Following the decision to eat, food
craving increased in both imagery conditions.
3.5. Correlation analysis
Table 2 summarizes the results of correlation analysis.3 In the negative mood condition, less
ability to resist eating was significantly correlated with higher negative emotion ratings
before mood induction and greater food craving at the end of delay in obese individuals.
Greater total calorie consumption was significantly correlated with greater food craving after
negative mood induction in obese individuals. In the positive mood condition, greater total
calorie consumption was significantly correlated with greater food craving at all three
assessment time points. No significant correlations were found among non-obese individuals
in the negative or positive mood condition.
4. Discussion
This pilot laboratory study was the first to examine the effects of positive and negative mood
induction on the ability to resist high-calorie foods, in addition to ad-libitum consumption of
preferred high-calorie food. Interestingly, obese individuals showed less ability to resist
eating high-calorie foods and subsequently consumed more calories in the positive mood
condition than in the negative mood condition. There are only two studies that examined the
effects of positive mood induction on eating behaviors in laboratory settings, with both
examining calorie consumption in normal weight individuals (Cools, et al., 1992; Yeomans
& Coughlan, 2009). Our findings suggest that positive mood may contribute to increased
consumption of high-calorie food in obese individuals by reducing their ability to resist
eating. Negative mood did not reduce the ability to resist eating in our obese sample.
Vicennati et al. (2009) has suggested that not all obese individuals are susceptible to
negative-mood related eating behaviors. However, in our non-obese sample, the negative
mood induction reduced the ability to resist eating when compared with the positive mood
induction, but this did not translate to increased calorie consumption. Findings suggest that
1The model included session order, age, and binge eating symptoms as they contributed to reducing residuals, and they were all
statistically significant (
p
< .05).
2The model included session order, age, gender, and emotional eating as they contributed to reducing residuals, but no covariates were
statistically significant (
p
> .05).
3All correlation analyses included gender and session order as covariates.
Udo et al. Page 6
Eat Behav
. Author manuscript; available in PMC 2013 April 23.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
negative mood may be a more salient cue to consume high-caloric snacks in this sample.
Together, the present findings support that mood, regardless of valence, can negatively
affect the ability to resist eating high-calorie food, and there are great individual differences
in whether positive or negative mood contributes to increased susceptibility to high-calorie
food intake.
Regardless of obesity status, food craving increased after the negative mood induction,
while it showed little changes after the positive mood induction. When an individual made a
decision to eat, food craving significantly increased in both positive and negative mood
conditions, suggesting the strong role of food craving in one’s decision to eat. Further,
correlation analyses revealed that reduced ability to resist eating high-calorie food was
significantly associated with food craving in obese individuals at the point when they
decided to eat following negative mood manipulation. Thus, negative mood may still have
an adverse impact on the ability to resist eating high-calorie food in obese individuals with
heightened food craving reactivity in response to negative mood. As intuitive as the finding
may be, this is the first laboratory study to demonstrate the relationship between food
craving and the ability to resist eating in obese individuals. The finding also adds
experimental evidence to a retrospective self-report study with overweight dieters which
demonstrated that situations which caused emotional ‘upset’ were antecedents of overeating
(Grilo, et al., 1989). While there were no baseline differences in negative emotion ratings
between our obese and non-obese samples, there was a negative association with pre-
induction levels of negative emotion and the ability to resist eating in the negative mood
condition among obese individuals. This is somewhat parallel to Jansen et al. (2008), who
found increased calorie consumption after exposure to palatable food cues in non-disordered
overweight/obese women with high levels of negative affect as compared to overweight/
obese women with low levels of negative affect. Thus, for obese individuals, pre-existing
levels of negative emotion may are important to consider when evaluating individual
differences in the ability to resist eating after experiencing negative mood inducing
situations.
The study further demonstrated significant correlations between total calorie consumption
and food craving in obese individuals. In the negative mood condition, higher level of food
craving following mood induction was associated with greater total calorie consumption.
Thus, increases in food craving may play a key role in negative mood-related eating in obese
individuals. In contrast, higher levels of food craving at all three assessment points were
associated with greater total calorie intake during the positive mood condition. Thus, greater
total calorie consumption in the positive mood condition may be attributable to generally
elevated food craving due to the three hours of food deprivation. Surprisingly, the
relationship between mood-induced food craving and eating behaviors has not been well-
explored in human laboratory studies on obesity. Given absence of significant correlations in
non-obese individuals, the findings from correlation analyses suggest that, similar to
addictive behaviors, food craving may have important mechanistic implications in mood-
induced high-calorie food consumption, especially in obese individuals.
In both obese and non-obese individuals, the negative mood induction increased negative
emotion ratings and decreased positive emotion ratings. Once participants decided to eat, but
had not yet consumed food, both positive and negative emotion ratings reverted back to pre-
induction levels. A similar pattern of findings was demonstrated when examining the effects
of mood induction on smoking behavior (McKee et al., 2011). Reward anticipation has been
shown to activate the ventral striatum (O’Doherty, 2004), which has also been linked to
anticipatory increase in positive affect (Burgdorf & Panksepp, 2006). We can speculate that
improvement in emotional states once participants made the decision to eat may reflect
reward anticipation. Alternatively, it may be related to alleviation of stress/negative emotion
Udo et al. Page 7
Eat Behav
. Author manuscript; available in PMC 2013 April 23.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
that had increased while resisting temptation to eat. While further research on underlying
mechanisms is needed, consistent findings between two distinct appetitive behaviors suggest
that changes in emotion in this self-administration paradigm may be useful in characterizing
the common aspect of reward anticipation in addictive behaviors.
Our study has several strengths and limitations that serve as context for our findings. We
adapted a validated smoking paradigm (McKee et al., 2011), which utilizes a well-studied
method to induce negative mood, to evaluate the ability to resist eating and ad-lib
consumption of preferred high-caloric snacks. In contrast to previous research focusing
those with binge eating disorders or with specific characteristics (e.g., high restraint eaters,
emotional eaters), this study examined mood-eating relationship in obese and non-obese
individuals without eating disorders. In addition, a substantial proportion of our sample was
ethnic minorities. Together, while the study results cannot be generalized to clinical
population, they provide insight into mood-induced overeating in the general population,
which has important implications for current obesity epidemic. As this was a pilot study, the
study sample was relatively small but comparable to similar investigations (Appelhans,
Pagoto, Peters, & Spring, 2010; Wallis & Hetherington, 2004, 2009). The advantage of well-
controlled within-subject laboratory investigations is that robust effects can be demonstrated
with relatively modest sample sizes. In our investigation, robust effect sizes of our primary
outcomes were demonstrated.
One important limitation is that the present study did not include measures of the
hypothalamic-pituitary-adrenal axis (HPA) reactivity. Epel et al. (2001) and Newman et al.
(2007) found that only women who showed high cortisol response consumed more calories
after psychological stress, suggesting a critical role of individual differences in
glucocorticoids response in the effects of mood induction on overeating and obesity.
Inclusion of a neutral mood condition might also have provided a reference point to clarify
whether positive or negative mood induction led to overconsumption of highly-palatable
foods. However, we have evaluated the use of neutral conditions (both neutral imagery and
no-imagery) in our smoking lapse paradigm, and find little difference between positive and
neutral conditions. Finally, our food craving scale was developed through modifying a well-
validated smoking craving questionnaire (Cox et al., 2001). Although the internal
consistency of the scale was high, further evaluation of its psychometric properties is
necessary. Future studies should also investigate whether factors suggested as risk for loss of
control over eating and overconsumption of palatable food, such as impulsivity (Guerrieri,
Nederkoorn, Schrooten, Martijn, & Jansen, 2009; Guerrieri, et al., 2007), reward sensitivity
(Davis, et al., 2007; Davis, Strachan, & Berkson, 2004), and dietary restraint (Hill, 2004),
may influence eating behaviors in our model.
4.1. Conclusion
This pilot study demonstrated that the influence of positive and negative mood induction on
the ability to resist eating and ad-lib consumption of high-caloric foods varied by obesity
status. We also demonstrated strong associations between food craving and these eating
behaviors, particularly after following a negative mood induction in obese individuals. This
is the first study to utilize a human laboratory model of eating behavior that incorporates a
behavioral measure of the ability to resist eating. Frequent snacking of high-caloric foods
has been associated with obesity (Berteus Forslund, Torgerson, Sjostrom, & Lindroos, 2005;
de Graaf, 2006). In addition to increased caloric intake per eating episode, reduced ability to
resist eating may lead to increases in frequency of high-caloric food intake, which can in
turn contribute to excessive calorie intake in the long-term. Collectively, our experimental
findings support clinical observations for overweight dieters (Grilo et al., 1989) that
different mood states can elicit urges to overeat. Our study also highlights that the ability to
resist eating in response to mood cues represents a critical factor for understanding the
Udo et al. Page 8
Eat Behav
. Author manuscript; available in PMC 2013 April 23.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
mechanisms underlying cue-induced eating behaviors. This model requires further
replication and future investigations should also examine relationships between individual
differences in eating habits and physiological measures of stress response. The self-
administration paradigm adapted in this study has been previously used to evaluate the
effects of medication on smoking behavior (McKee, 2009) and this unique approach to
study eating behaviors in laboratory settings may ultimately have important clinical
applicability in developing behavioral and pharmacologic methods to reduce craving and
enhance one’s ability to resist high-calorie foods.
References
Appelhans BM, Pagoto SL, Peters EN, Spring BJ. HPA axis response to stress predicts short-term
snack intake in obese women. Appetite. 2010; 54:217–220. [PubMed: 19925839]
Berteus Forslund H, Torgerson JS, Sjostrom L, Lindroos AK. Snacking frequency in relation to energy
intake and food choices in obese men and women compared to a reference population. International
Journal of Obesity. 2005; 29:711–719. [PubMed: 15809664]
Burgdorf J, Panksepp J. The neurobiology of positive emotions. Neuroscience and Biobehavioral
Reviews. 2006; 30:173–187. [PubMed: 16099508]
Center for Disease Control and Prevention. Healthy Weight - it’s not a diet, it’s a lifestyle!. 2011.
Retrieved from http://www.cdc.gov/healthyweight/assessing/index.html
Cools J, Schotte DE, McNally RJ. Emotional arousal and overeating in restrained eaters. Journal of
Abnormal Psychology. 1992; 101:348–351. [PubMed: 1583231]
Cox LS, Tiffany ST, Christen AG. Evaluation of the brief questionnaire of smoking urges (QSU-brief)
in laboratory and clinical settings. Nicotine & Tobacco Research. 2001; 3:7–16. [PubMed:
11260806]
Davis C, Patte K, Levitan R, Reid C, Tweed S, Curtis C. From motivation to behaviour: a model of
reward sensitivity, overeating, and food preferences in the risk profile for obesity. Appetite. 2007;
48:12–19. [PubMed: 16875757]
Davis C, Strachan S, Berkson M. Sensitivity to reward: implications for overeating and overweight.
Appetite. 2004; 42:131–138. [PubMed: 15010176]
de Graaf C. Effects of snacks on energy intake: an evolutionary perspective. Appetite. 2006; 47:18–23.
[PubMed: 16675059]
Epel E, Lapidus R, McEwen B, Brownell K. Stress may add bite to appetite in women: a laboratory
study of stress-induced cortisol and eating behavior. Psychoneuroendocrinology. 2001; 26:37–49.
[PubMed: 11070333]
First, MB.; Spitzer, RL.; Gibbon, M.; Williams, JB. Structured Clinical Interview for DSM-IV.
American Psychiatric Press; Washington, D. C.: 1995. Patient Edition
Flegal KM, Carroll MD, Ogden CL, Curtin LR. Prevalence and trends in obesity among US adults,
1999-2008. JAMA. 2010; 303:235–241. [PubMed: 20071471]
Goldfein JA, Devlin MJ, Kamenetz C. Eating Disorder Examination-Questionnaire with and without
instruction to assess binge eating in patients with binge eating disorder. International Journal of
Eating Disorders. 2005; 37:107–111. [PubMed: 15732081]
Grilo CM, Shiffman S, Wing RR. Relapse crises and coping among dieters. Journal of Consulting and
Clinical Psychology. 1989; 57:488–495. [PubMed: 2768608]
Guerrieri R, Nederkoorn C, Schrooten M, Martijn C, Jansen A. Inducing impulsivity leads high and
low restrained eaters into overeating, whereas current dieters stick to their diet. Appetite. 2009;
53:93–100. [PubMed: 19467278]
Guerrieri R, Nederkoorn C, Stankiewicz K, Alberts H, Geschwind N, Martijn C, Jansen A. The
influence of trait and induced state impulsivity on food intake in normal-weight healthy women.
Appetite. 2007; 49:66–73. [PubMed: 17261343]
Habhab S, Sheldon JP, Loeb RC. The relationship between stress, dietary restraint, and food
preferences in women. Appetite. 2009; 52:437–444. [PubMed: 19135112]
Udo et al. Page 9
Eat Behav
. Author manuscript; available in PMC 2013 April 23.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Hill AJ. Does dieting make you fat? British Journal of Nutrition. 2004; 92:S15–18. [PubMed:
15384316]
Izard, CE. Patterns of emotions: A new analysis of anxiety and depression. Academic Press; San
Diego, CA: 1972.
Jansen A, Vanreyten A, van Balveren T, Roefs A, Nederkoorn C, Havermans R. Negative affect and
cue-induced overeating in non-eating disordered obesity. Appetite. 2008; 51:556–562. [PubMed:
18495294]
Lattimore P, Caswell N. Differential effects of active and passive stress on food intake in restrained
and unrestrained eaters. Appetite. 2004; 42:167–173. [PubMed: 15010181]
McKee SA. Developing human laboratory models of smoking lapse behavior for medication
screening. Addiction Biology. 2009; 14:99–107. [PubMed: 18855800]
McKee SA, Sinha R, Weinberger AH, Sofuoglu M, Harrison EL, Lavery M, Wanzer J. Stress
decreases the ability to resist smoking and potentiates smoking intensity and reward. Journal of
Psychopharmacology. 2011; 25:490–502. [PubMed: 20817750]
Newman E, O’Connor DB, Conner M. Daily hassles and eating behaviour: the role of cortisol
reactivity status. Psychoneuroendocrinology. 2007; 32:125–132. [PubMed: 17198744]
O’Doherty JP. Reward representations and reward-related learning in the human brain: insights from
neuroimaging. Currrent Opinion in Neurobiology. 2004; 14:769–776.
Ogden CL, Yanovski SZ, Carroll MD, Flegal KM. The epidemiology of obesity. Gastroenterology.
2007; 132:2087–2102. [PubMed: 17498505]
Oliver G, Wardle J, Gibson EL. Stress and food choice: a laboratory study. Psychosomatic Medicine.
2000; 62:853–865. [PubMed: 11139006]
Pandit R, de Jong JW, Vanderschuren LJ, Adan RA. Neurobiology of overeating and obesity: the role
of melanocortins and beyond. European Journal of Pharmacology. 2011; 660:28–42. [PubMed:
21295024]
Piernas C, Popkin BM. Snacking increased among U.S. adults between 1977 and 2006. Journal of
Nutrition. 2010; 140:325–332. [PubMed: 19955403]
Reas DL, Grilo CM, Masheb RM. Reliability of the Eating Disorder Examination-Questionnaire in
patients with binge eating disorder. Behaviour Research and Therapy. 2006; 44:43–51. [PubMed:
16301013]
Schotte DE, Cools J, McNally RJ. Film-induced negative affect triggers overeating in restrained eaters.
Journal of Abnormal Psychology. 1990; 99:317–320. [PubMed: 2212283]
Sinha R. Modeling stress and drug craving in the laboratory: implications for addiction treatment
development. Addiction Biology. 2009; 14:84–98. [PubMed: 18945295]
Van Strien T, Frijters JE, Bergers GP, Defares PB. The Dutch Eating Behavior Questionnaire (DEBQ)
for assessment of restrained, emotional, and external eating behavior. International Journal of
Eating Disorders. 1986; 5:295–315.
Vicennati V, Pasqui F, Cavazza C, Pagotto U, Pasquali R. Stress-related development of obesity and
cortisol in women. Obesity. 2009; 17:1678–1683. [PubMed: 19300426]
Wardle J. Eating style: a validation study of the Dutch Eating Behaviour Questionnaire in normal
subjects and women with eating disorders. Journal of Psychosomatic Research. 1987; 31:161–169.
[PubMed: 3473234]
Wallis DJ, Hetherington MM. Stress and eating: the effects of ego-threat and cognitive demand on
food intake in restrained and emotional eaters. Appetite. 2004; 43:39–46. [PubMed: 15262016]
Wallis DJ, Hetherington MM. Emotions and eating. Self-reported and experimentally induced changes
in food intake under stress. Appetite. 2009; 52:355–362. [PubMed: 19071171]
Yeomans MR, Coughlan E. Mood-induced eating. Interactive effects of restraint and tendency to
overeat. Appetite. 2009; 52:290–298. [PubMed: 19022307]
Udo et al. Page 10
Eat Behav
. Author manuscript; available in PMC 2013 April 23.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Figure 1.
Means and standard errors of positive emotion ratings (upper) and negative emotion ratings
(lower) by imagery conditions by time (pre-mood induction, post-mood induction, decision
to eat).
Udo et al. Page 11
Eat Behav
. Author manuscript; available in PMC 2013 April 23.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Figure 2.
Means and standard errors of latency to decide eating (upper) and total calorie consumption
(lower) by obesity status (obese, non-obese) by imagery conditions (positive, negative).
Udo et al. Page 12
Eat Behav
. Author manuscript; available in PMC 2013 April 23.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Figure 3.
Means and standard errors of global food craving score by imagery conditions (positive,
negative) by time (pre-mood induction, post-mood induction, decision to eat).
Udo et al. Page 13
Eat Behav
. Author manuscript; available in PMC 2013 April 23.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Udo et al. Page 14
Table 1
Sample characteristics for obese and non-obese participants.
Obese (n = 15) Non-Obese (n = 15)
Age 41.9 (11.6)
†
31.7 (11.9)
% female 40.0 46.7
% White 53.3 46.7
% African-American 60.0 40.0
BMI
Ranges 35.1 (3.70)
†
30-44
23.0 (1.96)
20-26
Number of days engaging the following behaviors in the past 28 days at intake (from EDE-Q-I)
Recent dieting (item 1) 6.45 (9.16) 7.00 (10.81)
Binge eating (item 15) 2.71 (6.19) 0.87 (1.25)
Eating habits (from DEBQ)
Restraint eating 24.67 (9.90) 23.60 (10.38)
Emotional eating 31.33 (17.77) 31.93 (13.83)
Negative and positive emotion at the time of arrival to the laboratory session
Negative emotion (negative) 11.20 (8.79) 8.73 (3.67)
Positive emotion (negative) 68.26 (22.17) 59.79 (21.03)
Negative emotion (positive) 7.48 (3.78) 8.90 (3.44)
Positive emotion (positive) 63.63 (19.65) 65.51 (24.21)
Notes
. Numbers in parentheses indicate standard deviation. BMI = body-mass-index. EDE-Q-I = the Eating Disorder Examination Questionnaire
with Instructions (Goldfein et al., 2005). DEBQ = Dutch Eating Behavior Questionnaire (Van Strien et al., 1986), with score ranges from 10-50
(restraint eating) and 13-65 (emotional eating), respectively.
†
= significantly different from non-obese group at
p
< .05.
Eat Behav
. Author manuscript; available in PMC 2013 April 23.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Udo et al. Page 15
Table 2
Correlation between eating behaviors, mood, and food craving.
Negative Mood Induction Positive Mood Induction
Negative emotion ratings Food craving Positive emotion ratings Food craving
Pre Post End Pre Post End Pre Post End Pre Post End
Obese
Latency −.60
†
−.31 −.53 −.19 −.32 −.67
†
−.10 −.12 .02 −.40 −.53 −.35
Calories .44 .20 .24 .42 .63
†
.45 −.27 −.05 −.35 .66
†
.77
†
.64
†
Non-obese
Latency .03 .06 −.05 −.20 −.34 −.49 −.15 .18 .01 .17 .14 .06
Calories .03 .36 −.15 −.17 .00 .20 .07 −.04 .17 −.20 −.37 −.36
Notes
. Pre = pre-mood induction; Post = post-mood induction; End = end of delay (i.e., decision to eat).
†p
< .05.
Eat Behav
. Author manuscript; available in PMC 2013 April 23.
