This study might be of interest to anyone who struggles with binge eating while on a cutting diet. I'll be watching out for those "highly palatable foods" from now on, especially at night!

Int J Eat Disord. 2003 Sep; 34(2):183-97.

The role of palatable food and hunger as trigger factors in an animal model of
stress induced binge eating.

Hagan MM, Chandler PC, Wauford PK, Rybak RJ, Oswald KD.

OBJECTIVE: . . . . . rats with a history of caloric restriction and
only if highly palatable food (HPF) . . . . . Female rats were cycled through the R/S
protocol but this time were given just a taste of HPF with ad lib regular chow.
After another R/S cycle, rats were stressed during restriction (while hungry)
and were given HPF and chow.

RESULTS: Although binge eating did not occur if
only chow was available after stress, just a taste of HPF sufficed to increase
chow intake to more than 160% (p < 0. 001) of rats with a history of restriction
only, stress-only, or neither. Hunger increased the proportion of chow consumed
by both restricted groups, but stress magnified this hunger-induced overeating
by increasing HPF intake to 137% of restriction-only rats (p < 0. 001).

DISCUSSION: These effects suggest that binge eating in this model is motivated
by reward, not metabolic need, and parallels observations of binge triggers
described in clinical binge-eating disorders. . . . . . . . . . .

PMID: 12898554 [PubMed - in process]
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

In our previous work with the restriction/stress model, binge eating did not occur
if only regular food (rat chow)was available but only if HP food (cookies)was available.
Second, the rats were always stressed while sated from refeeding but never while hungry.
Therefore, the first test explored the effect of just a taste of HP food to trigger the binge-like
eating of chow when only chow was available. We predicted that, despite failure of
R + S rats to binge on chow when only chow is presented, they would be driven to
binge on chow if first allowed to ingest just a morsel of HP food. The second test explored
the effect of hunger on stress-induced eating by administering stress to rats while they were
calorie restricted. We predicted that, despite the expected hyperphagia of hungry rats
when given access to ad lib food (especially when it included HP food), the intake of
hungry rats that were stressed would exceed that of hungry rats that were not stressed.
In addition, we predicted that the magnitude of this binge would be greater than that
previously observed in the same rats while sated.
. . . . . . . . . . . . . . . . . . . . .

Results of Trigger Test 1:Effect of a Taste of Highly Palatable Food on Intake
of Chow After Stress When Only Chow Is Available

In contrast to Figure 2, which depicts no difference in chow intake by stressed rats with
a history of restriction (R + S group), just a taste-sized morsel of HP food was enough to
precipitate bingelike eating of chow in R + S rats (Figure 3A). The interaction between
stress and restriction was significant as early as 2 hours after stress (F [1, 25 ]<8. 1, p <0. 01;
data not shown). By 4 hours (Figure 3A)it remained significant (F [1, 25 ]<32. 82, p <0. 001),
with R + S rats eating 63% more chow kilocalories than the restricted group that was not
stressed (R + NS)and 68% to 77% more than the nonrestricted groups (all, p <0. 001). The
slightly lower intake of the NR + S (stress-only rats)was not significantly different from the
NS groups. Figure 3B depicts the continued R + S interaction at 24 hours (F [1, 25 ]<13. 7,
p <0. 001), with R + S rats consuming 54% more chow kilocalories than all other groups
(p <0. 001). In summary, when only nonpreferred regular food was available, just a morsel
of a preferred palatable food was enough to trigger binge eating of regular food when rats
with a history of restriction were stressed.

Results of Trigger Test 2:Effect of Hunger on Food Intake After Stress

Restriction interacted with stress to increase total food intake as early as 2 hours
(F [1, 25 ]<8. 5, p <0. 01; not shown)and remained significant at 4 hours (Figure 4A)
and 24 hours (Figure 4B)after stress (F [1, 25 ]<24. 5, p <0. 001). Parallel to the previously
reported behavior of the R + S rats during sated conditions (Figure 1), these same rats
continued to engage in binge eating while hungry. Despite the fact that both restricted
groups were tested while hungry (during restriction), Figure 4A shows that by 4 hours,
the R + S group ate 37% more total kilocalories than the nonstressed hungry group
(R + NS)and approximately 62% more than the nonrestricted, or sated, rats (all p <0. 001).
However, Figure 4A also shows that, unlike the case under sated conditions where the
nonstressed restricted group (R + NS)did not differ in intake from the nonrestricted
groups (Figure1A), hunger exerted an effect on intake, because R + NS rats ate 42% more
total kilocalories than NR + NS rats, mostly from chow (p <0. 01)and 32% more total
kilocalories than NR + S rats (p <0. 059). However, as shown in Figure 4B, by 24 hours,
the hyperphagia of the R + NS group dissipated but continued in R + S rats (p <0. 001).

Inspection of the proportion of total kilocalories made up of chow and HP food in
Figure 4 (hatched bars vs. solid dark bars, respectively)revealed that while hungry,
restricted rats consumed a higher proportion of chow to HP food compared with their
own intake of chow while sated (Figure 1)and compared with the nonrestricted groups
(Figure 4). This is consistent with a significant main effect of restriction (F [1, 25 ]<46. 9
and 52. 3, p <0. 001 at 4 and 24 hours, respectively)but a lack of interaction with stress on
chow intake (F [1, 25 ]<1. 7 and 0. 9, ns, at 4 and 24 hours, respectively). Therefore, the
bingelike intake of hungry R + S rats was due entirely to a significant interaction of
restriction and stress on HP food intake (F [1, 25 ]<14. 1 and 22. 3, p <0. 001, at 4 and 24
hours, respectively). At 4 hours, R + S rats ate 47% more HP food than the non-stressed
hungry rats (R + NS; Figure 4A dark bars). By 24 hours (Figure 4B), although the total
intake of HP food by R + S rats did not differ from controls (83 2 vs. 78. 2 4. 7 kcal,
respectively, ns)the increased intake of chow as a function of restriction produced a
significantly higher total kilocalorie bout of feeding in R + S rats compared with all other
groups (p <0. 001). In summary, caloric restriction increased intake of chow. This
increase, in addition to that produced by a synergistic R + S synergistic effect on HP
food intake, rendered a bout of feeding larger than that produced by hunger or stress


The purpose of this study was to explore the effect of a morsel of palatable food and of
hunger, two known clinical binge ‘‘triggers ’’in an animal model of binge eating. In this
model, binge eating is developed by means of a history of cyclic caloric restriction and
refeeding but is only behaviorally evident after stress. An analogous synergistic response
to restriction and stress may be germane in the development of stress-induced overeating
among individuals who are dieting and of binge eating among bulimic and BED patients
with a history of dieting and trauma or daily stress. By testing the effect of a HP food
‘‘trigger ’’and hunger, we were able to further define this animal model while adhering
to conditions veridical of human binge eating.

Just a morsel of HP food was enough to trigger binge eating of chow in rats that
experienced both caloric restriction and stress (R + S rats). Without the trigger or the
presence of ad lib HP food, these rats did not overeat. This response was surprising but
predicted by human observations. For example, bulimics are likely to binge if they allow
themselves to eat a favored or ‘‘forbidden ’’food, which is typically palatable and energy
dense (Kales, 1990; Russell, 1979). If they permit themselves to eat the food they crave,
a binge on any food is likely to follow (Waters et al. , 2001b). In restrained eaters, binge
eating is commonly precipitated after savoring a tasty food or preload (Abraham &
Beumont, 1982; Hetherington &Rolls, 1991; Rogers &Hill, 1989; Russell, 1979).

A previous study showed that HP food, alone, without stress, could rekindle over-
eating in rats after a 30-day hiatus from 12 weeks of restriction-refeeding cycles but only
in rats with prior exposure to the HP food (Hagan &Moss, 1997). Rats also learn to
overeat if their time with a preferred high-fat food is limited once they are exposed to it
(Corwin, Wojnicki, Fisher, Dimitriou, Rice, &Young, 1998). Therefore, the contingency of
HP food to produce binge eating of chow may be due to a learned contrast effect between
regular and HP food. However, the importance of a contrast effect is weakened by an
earlier test in these rats in which HP food was reintroduced after 24 hours of chow only
after stress, and R + S rats failed to overeat (Hagan et al. , 2002). This second 24 hr period
was not immediately preceded by stress. Therefore, the proximity of stress to access of
HP food may be a more important contingent than the contrasting qualities of HP and
regular food.

The action of a morsel of HP food to trigger binge eating is reminiscent of the
reinstatement of drug use that is produced by just a single ‘‘priming ’’dose of a drug
(or by acute stress, such as foot shock)after prolonged extinction of responding for the
drug (Erb, Shaham, &Stewart, 1996; Stewart, 2000). This occurs even if the self-
administered substance is saline, indicating it is the learned pharmacological properties
of the drug that are being sought (Gerber &Stretch, 1975). Similarly, HP food may have
primed binge eating of chow, not for chow per se, but for the pharmacological effects
associated with overeating after stress. This would explain why rats do not overeat when
there is a 24-hour gap between stress and availability of HP food (Hagan et al. , 2002). Just
as drug cravings and the onset of food binges are highly correlated with negative affect
(Hyman, Hyman, &Malenka, 2001; Russell, 1979; Waters, Hill, &Waller, 2001a), the
pharmacological effect of HP food or a palatable primer may not be salient, desired, or
even necessary to rats unless it is immediately available after the negative experience of
foot shock.

Perhaps the more important conclusion from the HP food trigger test is that the rat is
incapable of the psychological processes involved in the constructs that are offered to
explain binge eating in humans such as ‘‘cognitive disinhibition, ’’‘‘abstinence violation, ’’
or ‘‘counterregulation ’’(Guertin, 1999; Hetherington &Rolls, 1991; Knight &Boland,
1989; Rogers &Hill, 1989). Rats are not concerned with ‘‘blowing ’’a diet or giving in to a
fattening ‘‘forbidden ’’food simply because they have already ingested some of it. This is
not to say that higher cognitive processes do not play a role in the precipitation of human
binge eating. Rather, our observations that HP food can trigger binge eating of less
preferred food in rats suggests that biological or pharmacological drives may be more
influential in human binge eating than is currently considered. We predict that these
biological processes, now able to be deciphered, will explain much more of the ‘‘lack of
control ’’that characterizes human binge eating (APA, 1994)than can be explained by
higher cognitive processes alone.

The results of the hunger trigger test were consistent with our predictions that, despite
a greater than usual intake of the restricted groups because of energy depletion, stress
would still drive the most intake in rats with a history of restriction. That is, the
synergistic effect of restriction and stress holds up regardless of energy state, overriding
the overeating of hungry animals that are not environmentally stressed. This, together
with the fact that R + S rats were actually able to consume more kilocalories while sated
(Figure 1B)than they consumed after being calorie deprived for 5 consecutive days
(Figure 4B)emphasizes the powerful magnitude of food intake that characterizes this
animal model of binge eating.

In terms of potency as a binge trigger in rats, hunger is not equal to HP food. Unlike
the requirement of an HP trigger to produce binge eating of chow, hunger is not required
to produce binge eating in this model as evidenced by Figures 1 and 3 (data obtained
when rats were sated). Interestingly, hunger, compared with ‘‘forbidden ’’food and
negative affect, also plays a rather weak binge-triggering role in humans. For example,
bulimics cite that hunger is not a trigger to binge (Abraham &Beumont, 1982;
Hetherington, Stoner, Andersen, &Rolls, 2000; Waters et al. , 2001a), and hunger can
even increase after a meal (Walsh, Kissileff, Cassidy, &Datzic, 1989). A DSM-diagnostic
feature of BED is eating large amounts of food when not hungry (APA, 1994), and among
bariatric surgery patients, binge eating can persist despite decreased hunger (Saunders,
2001). Chronic dietary restraint, more than hunger from acute food deprivation, was
suspected to be a more potent precipitating factor in eating disorders (Hetherington et al. ,
2000). This notion is also true of our model. Chronic dieting (a history of restriction-
refeeding)is more influential than hunger or even stress, because R + Sratsbingedwhile
sated. Also, just one exposure to stress was found to be required to produce powerful binge
eating if the rats had a protracted (7-cycle)history of restriction (Hagan et al. , 2002).

Despite a minor role of hunger to elicit binge eating, the hunger trigger test supported
previous observations that rats eat a higher proportion of their total intake from chow
when energy depleted. Paradoxically, chow contains less fat and provides less energy
than the HP cookies. It may be that rats resort to ingesting more chow because they have
a lifetime experience with the positive nutritive consequences of chow. In humans,
hunger is more likely to increase selection of ‘‘good, ’’‘‘healthy food ’’over foods typically
eaten as dessert or snacks. The latter are more likely to be chosen over healthier food on a
full stomach. Interestingly, bulimic binges may be precipitated by HP food, which is
preferred during a binge (Weltzin, Hsu, Pollice, &Kaye, 1991; Woell, Ficher, Pirke, &
Wolfram, 1989), yet studies have found that at the end, bulimic binges are macronutrient
proportional to normal American-typical meals (Elmore &deCastro, 1991; Walsh et al. ,
1989; Woell et al. , 1989). In contrast, the few studies on macronutrient composition of
binges in BED report a disproportionately higher intake of fat (Yanovski, Leet, Yanovski,
Flood, Gold, Kissileff, &Walsh, 1992)and of meat, a fatty protein (Cooke, Guss, Kissileff,
Devlin, &Walsh, 1997). By definition, caloric restraint is more common among bulimics
than BED individuals (APA, 1994). Therefore, our animal model may be more veridical of
bulimia when rats are stressed during restriction, because this yields a more normal meal
macronutrient composition. Our model may be more veridical of BED when rats are
stressed after refeeding, because this yields increased intake of HP food (cookies), the
fattier food. Admittedly, we can only be simulating a subtype of BED that is anteceded by
dieting (Grilo &Masheb, 2000; Stice et al. , 2001), because a history of caloric restriction is
imperative to he synergistic restriction/stress formula for binge eating in this animal
model. As far as simulating stress, it is doubtful that anyone fitting the dieting-first
subgroup of BED did not or does not also continue to encounter some form of envi-
ronmentally induced stress.

What is more important about the change in food selection created by hunger is that it
highlights the difference between eating for metabolic need versus reward. Others have
observed the same pattern of increased chow vs. increased HP food for metabolic need
vs. reward, respectively (Glass, Billington, &Levine, 2000; Hagan &Moss, 1995; Weldon,
O ’Hare, Cleary, Billington, &Levine, 1996). When presented with a choice in this model,
R + S rats overwhelmingly consume HP food over chow and, as shown here, must be
primed with HP food to overeat chow. Therefore, the binge eating produced in this
animal model very likely reflects consumption primarily for reward.

Binge eating for reward is consistent with clinical binge eating in humans, which is
behaviorally and emotionally addictive-like in nature (Davis &Claridge, 1998; Jansen,
1998). Analogous to substances of abuse, binge eating is reported to provide ‘‘escape ’’or
‘‘medicating ’’qualities (Heatherton &Baumeister, 1991; Polivy et al. , 1994; Waters et al. ,
2001a). In rats and humans, stress is associated with an increased preference and craving
for sucrose (Bertiere, Sy, Baigts, Mandenoff, &Apfelbaum, 1984; Oliver, Wardle, &
Gibson, 2000)and typically craved foods, those high in fat and sugar, have stress-
reducing effects (Blass, Shide, &Weller, 1989). In this model, a history of restriction
and environmental stress may produce analogous negative states, which the rats attempt
to ameliorate by means of increased ingestion of rewarding food. Opioid and dopamine
systems, which signal drug reward in the central nervous system, are implicated to
mediate feeding induced by caloric restriction (Hagan &Moss, 1991; Pothos, Creese, &
Hoebel, 1995)and in the rewarding effects of palatable foods (Colantuoni, Schwenker,
McCarthy, Rada, Ladenheim, Cadet, Schwarz, Moron, &Hoebel, 2001; Drewnowski,
Krahn, Demitrack, Nairn, &Gosnell, 1995). Peptide YY (PYY)is a potent orexigenic
peptide that motivates intake of HP food despite its pairing with aversive consequences
(Hagan, 2002; Hagan &Moss, 1995). PYY levels are also elevated in the central nervous
system of abstaining bulimics (Kaye, Berrettini, Gwirtsman, &George, 1990). Together,
these comprise key neurochemical targets for future investigation in the neuroadaptation
of R + S rats to binge.

Several behavioral studies are still needed to define this model and to address the
shortcomings of this study, namely the influence of prior tests on the present results. The
same rats were used in a series of studies, so additional control groups would be needed
to ascertain that prior testing was not responsible for the effects reported here. However,
repeated testing failed to affect the intake of the three control groups that were always
run alongside the R + S groups. Throughout all tests, only the R + S group ate
abnormally. Of future interest is to determine whether hunger, like a morsel of HP
food, is sufficient to prime binge eating of chow when only chow is available after stress
and whether binge eating can be obtained with only chow if rats are never exposed to HP
food. We suspect that other stressors besides foot shock will elicit binge eating, but this
must also be tested. Another pressing question is whether these binge-eating effects will
replicate in male rats, especially in light of evidence that many stressors evoke different
patterns of food intake in female vs. male rats (McIntosh, Anisman, &Merali, 1999; Pare,
Blair, Kluczynski, &Tejani-Butt, 1999; Zylan &Brown, 1996). The role of gonadotropic
hormones is relevant and may shed light on the disproportionately greater incidence of
eating disorders among females (APA, 1994).

In conclusion, these data provide additional evidence that a history of caloric restric-
tion and stress, despite satiation and normal body weight, interact to promote abnormal
increases in food intake that are characteristic of binge eating. Just a morsel of palatable
food could precipitate binge eating of a food that is otherwise not overeaten. Also,
hunger, although not necessary to trigger binge eating, caused a significant increase in
regular food intake that, together with an R + S–induced overeating of HP food, drove
intake above and beyond what would be expected to be caused by hunger alone, a
response consistent with eating for reward. These results parallel the role of HP food
and hunger as binge-precipitating factors in humans and strengthen the validity of this
animal model to further our understanding of BEDs. Ultimately, this model will facilitate
the identification of the exact physiological mediators and changes caused by dieting and
stress to promote binge eating and, subsequently, should facilitate the identification of
those critical in arresting it.


Macho is not a gender issue.