Hunger, inhibitory control and distress-induced emotional eating

Research report

Hunger inhibitory control and distress-induced emotional eating

Tatjana van Strien ab Machteld A Ouwens c1 Carmen Engel a Carolina de Weerth a

a Behavioural Science Institute Radboud University Nijmegen PO Box 9102 6500 HC Nijmegen The Netherlandsb Faculty of Earth and Life Sciences Free University Amsterdam De Boelelaan 1085 1081 HV Amsterdam The Netherlandsc Department of Medical and Clinical Psychology Tilburg University PO Box 90153 5000 LE Tilburg The Netherlands

A R T I C L E I N F O

Article historyReceived 16 December 2013Received in revised form 17 April 2014Accepted 19 April 2014Available online 23 April 2014

KeywordsEmotional eatingHungerInhibitory controlStop Signal TestTrier Social Stress TestFood intake

A B S T R A C T

Self-reported emotional eating has been found to significantly moderate distress-induced food intakewith low emotional eaters eating less after a stress task than after a control task and high emotional eaterseating more The aim of the present study was to explore possible underlying mechanisms by assessingpossible associations with (1) ability to experience the typical post-stress reduction of hunger and (2)inhibitory control We studied these effects in 54 female students who were preselected on the basis ofextremely high or low scores on an emotional eating questionnaire Using a within subject design wemeasured the difference of actual food or snack intake after a control or a stress task (Trier Social StressTest) As expected the moderator effect of emotional eating on distress-induced food intake was foundto be only present in females with a failure to report the typical reduction of hunger immediately aftera stress task (an a-typical hunger stress response) Contrary to our expectations this moderator effect ofemotional eating was also found to be only present in females with high ability to stop motor impulses(high inhibitory control) These findings suggest that an a-typical hunger stress response but not poorinhibitory control may underlie the moderator effect of emotional eating on distress-induced food intakeHowever inhibitory control may play a role whether or not there is a moderator effect of self-reportedemotional eating on distress-induced food intake

copy 2014 Elsevier Ltd All rights reserved

Introduction

Distress is associated with both increased and decreased foodintake (Greeno amp Wing 1994) with eating less being the typical andpredominant response (Gold amp Chrousos 2002 Stone amp Brownell1994) Distress is normally associated with physiological re-sponses that are designed to prepare the individual for a fight orflight reaction inhibition of gastric motility and release of sugar intothe bloodstream As these physiological states are similar to the chiefperipheral physiological correlates of satiety distress generally leadsto decreased eating and subsequent weight loss (Gold amp Chrousos2002 Stone amp Brownell 1994) However so-called emotional eatersshow the atypical response to distress of eating similar or largeramounts of food (Oliver Wardle amp Gibson 2000 van Strien et al2013a van Strien Herman Anschutz Engels amp de Weerth 2012avan Strien amp Ouwens 2003 Wallis amp Hetherington 2004) Thoughthese effects are robust when participants have sufficiently extreme

emotional eating scores1 andor ego threat is involved (van Strienet al 2012a) little is known about the mechanisms underlying theseeffects

In an earlier study on possible mechanisms we examined whetherthis opposite pattern of emotional over-eating is associated withchanges in the stress reactivity of the hypothalamic pituitary adrenal(HPA) axis as indicated by changes in the stress hormone cortisolafter a distress (ego-threat) versus a control condition (van StrienRoelofs amp de Weerth 2013b) Cortisol stress reactivity was indeedfound to significantly moderate the relationship between emotion-al eating and the difference in food intake after distress versus control

Acknowledgement Anita Jansen and Chantal Nederkoorn of Maastricht Univer-

sity are kindly acknowledged for providing us this version of the Stop Signal Task Corresponding author

E-mail address tvanstrienpsychrunl (T van Strien)1 Present address GGzBreburg Tilburg PO Box 770 5000 AT Tilburg The

Netherlands

1 For two reasons it is important to have sufficient participants with extreme ob-servations on emotional eating The first reason is statistical According to McClellandand Judd (1993) pp 382ndash383 ldquojointly extreme observations are crucial for detect-ing interactionsrdquo In comparison with the lsquofour corners designrsquo in which 25 of casesare allocated to each extreme ldquoa normal like distribution of the two variables has arelative efficiency of only 06 for detecting an interaction and requires nearly 17 timesas many observations to have comparable efficiencyrdquo (Whisman amp McClelland 2005p 117) Interestingly most studies with no moderator effect for emotional eating usedno extreme values on emotional eating and only had a small number of subjects intheir study (van Strien et al 2012a 2012b) The second reason is conceptual Resultsof a functional magnetic resonance imaging study by Bohon Stice and Spoor (2009)suggest that emotional eating may be best described as a categorical rather than acontinuous variable Extreme scores on self-reported emotional eating predicted inthis study important individual differences in reward response during negative moods

httpdxdoiorg101016jappet2014040200195-6663copy 2014 Elsevier Ltd All rights reserved

Appetite 79 (2014) 124ndash133

Contents lists available at ScienceDirect

Appetite

journal homepage wwwelseviercom locate appet

tasks High emotional eaters with a blunted cortisol response atemore food after distress than those with an elevated cortisol stressresponse (see Tomiyama Dallman amp Epel 2011 and Tryon DeCantamp Laugero 2013 for similar findings) This finding suggests that emo-tional over-eating is indeed associated with a lowered HPA-axis func-tioning though it is as yet unclear whether this blunted cortisolresponse is secondary (ie the result of an adaptive down regula-tion) or primary to emotional eating (see van Strien et al 2013b)In the present study on the same sample plus additional participants2

we want to pursue our search for possible mechanisms further thistime by addressing the relationship of emotional eating with (1) self-reports of hunger after stress versus control and (2) inhibitory control

Emotional over-eating may be the result of a failure to experi-ence the typical post-stress reduction of hunger An earlier studyexamined the hunger ratings immediately after a distress vs controlcondition of low versus high emotional eaters (van Strien et al2012a) It was found that high emotional eaters in contrast to thelow emotional eaters did not report a substantial reduction of hungerimmediately after the stress task compared with the control taskThis finding suggests that high emotional eaters in contrast to thelow emotional eaters indeed suffer from an a-typical hunger stressresponse Whether this a-typical hunger stress response also pre-dicts distress-induced food intake was however not investigatedin that study Furthermore there are to our best knowledge no pub-lished studies that have systematically tested the relationshipbetween hunger stress reactivity and actual stress-induced foodintake in high and low emotional eaters The present study is there-fore the first to systematically examine whether the moderator effectof emotional eating on distress-induced food intake is affected bythe individualrsquos hunger stress reactivity

Failure to experience the typical post-stress reduction of hungerof high emotional eaters may be the result of poor interoceptiveawareness that is a confusion of physiological symptoms associ-ated with stress and negative emotions and those associated withhunger and satiety (Bruch 1964) Poor interoceptive awareness ishighly associated with alexithymia (Greek literally meaning lsquonowords for emotionsrsquo) most notably the alexithymia aspects of dif-ficulty in identifying feelings and describing feelings to other people(Garner 1991 Taylor Parker Bagby amp Bourke 1996) The associa-tion of poor interoceptive awareness and alexithymia with self-reported emotional eating has been widely established (Larsen vanStrien Eisinga amp Engels 2006 Pinaquy Chabrol Simon Loevet ampBarber 2003 van Strien Engels van Leeuwe amp Snoek 2005) within one study a moderator effect of alexithymia on actual distress-induced food intake (van Strien amp Ouwens 2007) However whetherpoor interoceptive awareness and alexithymia are also associatedwith hunger stress reactivity (typical versus a-typical hunger re-sponses after stress versus control) has not yet been assessed Sim-ilarly it is as yet unknown whether a possible moderator effect ofhunger stress reactivity on the moderator effect of emotional eatingon distress-induced food intake is mediated by poor interoceptiveawareness and alexithymia

Inability to inhibit motor impulses (pressing a button) ndash a measurefor inhibitory control which has been considered an indirect measurefor impulsivity (Jansen et al 2009) ndash would also be a good candi-date for affecting the moderator effect of emotional eating ondistress-induced food intake In a series of studies summarized byJansen et al (2009) inability to inhibit motor impulses as mea-

sured with the stop signal task (by Logan Schachar amp Tannock 1997)predicted higher food intake a higher body weight and less weightloss after a weight reduction treatment However in more recentstudies inability to stop motor impulses only predicted overeatingand weight gain in combination with other factors such as a highimplicit preference for the test food high dietary restraint or strongfeelings of hunger (Jansen et al 2009 Nederkoorn HoubenHofmann Roefs amp Jansen 2010) Emotional eating could also be sucha factor given the association of emotional eating with various self-report measures of reduced inhibitory control or negative urgency(the tendency to act impulsively in the face of negative affect)(Bekker van de Meerendonk amp Mollerus 2004 Ebneter LatnerRosewall amp Chisholm 2012 Elfhag amp Morey 2008 Lattimore Fisheramp Malinowski 2011 Ouwens van Strien amp van Leeuwe 2009 Racineet al 2013) and the number of false alarms in a GoNoGo task ameasure of efficiency of inhibitory control in the presence of foodcues (Jasinska et al 2012) We know of no published study thattested the relationship between inability to stop motor impulses andactual stress-induced food intake in high and low emotional eatersHence to the best of our knowledge this would be the first studyto assess the effect of this implicit measure of impulsivity on themoderator effect of emotional eating on distress-induced food intake

This study aims at exploring the relationships of (1) hunger stressreactivity and (2) inhibitory control with actual distress-induced foodintake in high versus low emotional eaters As in most previousstudies on emotional eating the focus is on females and on eatingin response to negative emotions3 (see also Gibson 2012) Earlierwe found that emotional eating significantly moderated the distress-food intake association with low emotional eaters eating less afterthe stress than after the control task and high emotional eaters eatingmore For the present study we hypothesized that this moderatoreffect of emotional eating would be more likely to hold true for par-ticipants with an a-typical hunger stress response (failure to showthe typical reduction of hunger after stress) and for participants withhigh inability to stop motor impulses (poor inhibitory control) Spe-cifically participants who combined an a-typical hunger stress re-sponse or poor inhibitory control with high emotional eating wereexpected to eat more after the stressor than those with the typicalhunger stress response and adequate inhibitory control We left itan open empirical question whether this possible moderator effectof hunger stress reactivity on the moderator effect of emotionaleating on distress-induced food intake would be mediated by poorinteroceptive awareness or alexithymia

Methods

Design

We used a within-subjects design in which females who werepreselected on the basis of extremely high or low scores on an emo-tional eating questionnaire were subjected to a control task and astress task (Trier Social Stress Task ndash TSST) on 2 consecutive daysThe TSST involves speaking in front of a jury coupled with an ar-ithmetic challenge Because the stress condition is perceived by somesubjects as quite stressful we deliberately chose to always start with

2 Results of a sub-sample of 47 participants the data of which had been collect-ed in spring and autumn of 2010 have been reported in two previous publications(van Strien et al 2012a Study 2 van Strien et al 2013b) They address respective-ly the moderation of distress-induced eating by emotional eating scores and cor-tisol reactivity and distress-induced emotional eating The data of the additionalparticipants for the present study were collected in spring 2012

3 While recognizing that eating in response to positive emotions may occur as fre-quently as eating in response to negative emotions there is increasing evidence thateating in response to negative and to positive emotions may refer to different con-structs (Nolan Halperin amp Geliebter 2010 van Strien et al 2013a) One reason maybe that we tend to celebrate happy events with food Eating in response to positiveemotions unlike eating in response to negative emotions can therefore not be con-sidered an lsquoinaptrsquo or lsquoinappropriatersquo response Accordingly it can be expected thatself-assessed emotional eating in response to positive emotions has a main effecton food intake whereas self-assessed emotional eating in response to negative emo-tions only shows an effect on food intake in interaction with distress

125T van Strien et alAppetite 79 (2014) 124ndash133

the control condition and not to counterbalance the order of the twoconditions We were afraid that the control condition would sufferfrom carry over effects if we started with the stress condition andalso that we would lose too many participants because they wouldrefuse to come back the following day for the control condition4The study protocol was approved by the ethical board of the Facultyof Social Sciences of the Radboud University Nijmegen Before par-ticipating the participants filled out informed consent forms

Participants

Participants for the present study (n = 60) were recruited froma pool of female students taking introductory psychology or ped-agogy courses who had completed the emotional eating scale in classor on our research participant portal Participants with scores below182 or above 325 (corresponding to the 20th and 80th percentileof the Dutch norm group of females) were invited to participate ina study on lsquohealth and physiologyrsquo2

Procedure

A complete description of the procedure can be found in vanStrien et al (2013b)5 Briefly the two sessions were scheduled onconsecutive weekdays between 11 am and 15 pm For 1 h prior tothe experiment the participants were asked to refrain from eatingand drinking (an exception was made for water) On the first dayparticipants were subjected to the control condition in which theyhad to rate different fabrics (eg wool and fur) on various attri-butes (eg softness warmth) After 15 min the experimenter tookthem to a separate room to fill out several questionnaires at a tablewhich also held a glass of water and four bowls filled with respec-tively white grapes pieces of carrot MampMs and pieces of butter cakeWith the words ldquoPlease help yourself to the water and the food Youhave earned itrdquo the experimenter left the room to return after 20 minto take the participants to another room to perform the 15-min stopsignal task (a computer task)

On the second test day the participants were subjected to a modi-fied version of the Trier Social Stress Task (Kirschbaum Pirke ampHellhammer 1993) Briefly the task consisted of preparing (5 min)and delivering (5 min) a videotaped speech followed by a serial sub-traction task (5 min) The speech and subtraction task were pre-sented in front of a two-person jury who sat behind a table and worewhite doctorsrsquo coats The participant had to stand in stocking feeton a Wiicopy balance board presumably to measure kinaesthetic ac-tivity (but in actual fact to enhance stress) After the stress task theexperimenter asked to wait for the juryrsquos judgement of the parti-cipantrsquos performancemdashin this manner the stressfulness of the publicspeaking task was extended by a prolonged period of waiting forthe resultsmdashand to fill out a set of questionnaires After 15 min theexperimenter returned to communicate a positive judgement by thejury after which the experimenter took them to a separate room tofill out a further set of questionnaires Once again on the table a glass

of water and bowls with food were placed and participants wereinvited to help themselves to the water and the food in the samewords as the previous day After 20 min the experimenter re-turned to measure the weight and height of the participants in lightclothing and stocking feet Finally the participant was debriefedthanked and paid with course credits

It should be noted that the experimenter was kept blind to theemotional eating status of the participants and that all partici-pants were unaware that their food intake was being assessed

Materials and measures

Actual food intakeBefore and after participants ate the bowls with grapes carrots

MampMs and butter cakes were weighed with a professional balance(Kern 200) We used the simple addition (in grams) of the variousfood-types as outcome variable because food weight has been shownto be a stronger determinant of intake during an eating episode thanenergy content (Westerterp-Plantenga 2004) In addition to the totalweight of food consumed (grapes plus carrots plus MampMs plus buttercakes) we also used the addition (in grams) of snack food (MampMsplus butter cakes) Inspection of outliers defined as 3 SD above themean revealed that two persons showed outliers above the meanin their food intake data (one person had an outlying value on totalfood and another person had an outlying value on intake of snackfood) The outlying values were winsorized by replacing the out-lying values with the value of 3 SD above the mean

Mood and hungerOn both days ratings of affect and hunger upon arrival and at

three more time points post-task (but before food intake) were mea-sured (for the Flow chart for the times as related to the protocolsee van Strien et al 2013b) The Positive and Negative Affect Sched-ule (PANAS Watson Clark amp Tellegen 1988) was used to measureon a 5-point (lsquonot at allrsquo to lsquoextremelyrsquo) scale the degree to whichparticipants experienced 10 positive and 10 negative emotions Thedescriptor lsquohungryrsquo was inserted among the PANAS items so thathunger could be evaluated without alerting the participants to thetrue nature of the study6

Hunger stress reactivity was calculated by computing the dif-ference between the area under the curve with respect to increase(AUCi) in the stress condition and the AUCi in the control condi-tion (deltaAUCi) using the formula for AUCi by PruessnerKirschbaum Meinlschmid and Hellhammer (2003) A higherdeltaAUCi is indicative of a typical hunger stress response of lowerhunger values after the stress compared with the control condi-tion (The mean values of AUCi stress and AUCi control are both neg-ative and the mean of deltaAUC1 has a positive value (see thedescriptives of hunger food intake and other measures in the Resultssection))

Eating behaviourEating behaviour was assessed with the Dutch Eating Behaviour

Questionnaire (DEBQ van Strien 2010) The DEBQ has 33 items 10on restrained eating (eg ldquoDo you try to eat less at mealtimes thanyou would like to eatrdquo) 10 on external eating (eg ldquoIf food smellsand looks good do you eat more than usualrdquo) and 13 on emotion-al eating (eg ldquoDo you have a desire to eat when you are irri-tatedrdquo) All items have to be rated on a 5-point scale with responsecategories that range from 1 lsquoneverrsquo to 5 lsquovery oftenrsquo This scale has

4 The participants were informed that the study was on the influence of stress onsensation and behaviour and that the study would take place on 2 consecutive daysThey thus knew that there would be a stressor on the second day but they did notknow exactly what kind of stressor this would be and also not that this stressor wouldinvolve ego threat (in Europe this stressor is perceived by most participants as verystressful) We cannot rule out that there was anticipatory stress on the second dayHowever all participants came back on the second day and this very low numberof drop outs also is a valuable outcome for us

5 At five identical time points on each study day saliva samples were collected bypassive drooling For the present study the cortisol values only served as manipu-lation check As in the previous study on the smaller sample (van Strien et al 2013b)cortisol values were significantly higher in the stress than in the control conditionindicating that the stress manipulation was successful (results on requestavailable)

6 The significant moderator effect of the stress condition on the hunger values overtime in addition to the significant moderator effect of emotional eating on the hungervalues in the stress condition can also be interpreted as indicating that our measureof hunger has good validity for the typical versus a-typical hunger responses to stress

126 T van Strien et alAppetite 79 (2014) 124ndash133

good internal reliability and good construct and predictive validi-ty (van Strien Herman amp Anschutz 2012b van Strien et al 2012avan Strien amp van de Laar 2008)

Poor interoceptive awareness and alexithymia

Poor interoceptive awareness was measured with a subscale ofthe Revised Eating Disorder Inventory (EDI-II) (Garner 1991) Thescale for poor interoceptive awareness has 10 items (eg I get con-fused as to whether or not I am hungry) Response categories rangedfrom 1 lsquoneverrsquo to 6 lsquoalwaysrsquo In contrast to the EDI manual (Garner1991) in which a transformation of responses into a four-point scaleis advocated the present study utilized untransformed responsesas scale transformation was found to reduce the validity of the EDIamong a non-clinical population (Schoemaker van Strien amp van derStaak 1994) Scales were constructed by calculating the means forall scales

The alexithymia aspects lsquoDifficulty identifying feelingsrsquo and lsquoDif-ficulty describing feelingsrsquo were measured with the TorontoAlexithymia Scale-20 the TAS-20 (Bagby Parker amp Taylor 1994Dutch translation Kooiman Spinhoven amp Trijsburg 2002) Thesubscale lsquodifficulty identifying feelingsrsquo has seven items (eg ldquoI havefeelings that I canrsquot quite identifyrdquo) The subscale lsquodifficulty describ-ing feelingsrsquo has five items (eg ldquoIt is difficult for me to find the rightwords for my feelingsrdquo) Response categories ranged from 1 lsquoneverrsquoto 5 lsquoalwaysrsquo

Response inhibition

A version of the Stop Signal Task (Nederkoorn et al 2010) wasused to measure response inhibition It contains two sorts of trialsa lsquogo trialrsquo (75) and a lsquostop trialrsquo (25) During the lsquogo trailsrsquo the par-ticipants have to react as quickly as possible to a stimulus wherethey have to press a certain button on the right when they see anlsquoXrsquo on the screen and a certain button on the left when they see onlsquoOrsquo on the screen (before 1500 ms) This response has to be inhib-ited when they hear a tone during the lsquostop trailsrsquo which means theparticipants do not push a button in response to the stimulus Atthe start of the task the delay between the go signal (lsquoXrsquo or lsquoOrsquo) is250 ms this increases with 50 ms when the participant inhibits suc-cessfully and decreases with 50 ms when the participant is not ableto stop making the task easier The task contains two practice blocksof eight and 16 trails and four blocks of 128 trails The stop signalreaction time (SSRT) is measured by the mean reaction time (SR)minus the mean delay on the stop trails (stop delay) The higher theSSRT value the more impulsive the participant Inspection of out-liers defined as 3 SD above or below the mean revealed that twopersons showed outliers below the mean in their SSRT assess-ment The outlying values were winsorized by replacing the out-lying values with the value of 3 SD below the mean

Statistical analyses

All analyses were carried out using SPSS version 190 (SPSS IncChicago) With repeated measures GLM we conducted various ma-nipulation checks by assessing the effect of time on the various cor-tisol valueshunger ratings in both the control and the stresscondition in addition to the effect of condition (control vs stress)on the hunger response over time In the same manner we also as-sessed the effect of condition (control vs stress) on the mood re-sponse over time Subsequently we assessed the effects of emotionaleating or impulsivity (dichotomized by means of median split) onthe hunger and mood response over time in both the control andthe stress condition GreenhousendashGeisser corrections were appliedwhere appropriate Further where appropriate (n gt (conditions + 30)HuynhndashFeldt lt 85) and the multivariate test was applied

Next we computed delta food and delta snack intake which isthe difference between intake of food (in grams) respectively intakeof snack (in grams) in the stress and the control condition a pos-itive value meaning a higher intake of total food respectively snackfood in the stress condition As a result the within-subjects factorcondition has disappeared but it is still expressed through the dif-ference scores With hierarchical regression analyses we tested theinteraction of (a) hunger stress reactivity and (b) inhibitory controlwith emotional eating on delta food and delta snack intake (in ad-dition to the main effects) In case of significant effects we as-sessed its robustness by additionally correcting for external eatingand dietary restraint (because of the high interrelations betweenemotional external eating and dietary restraint) In the analyses withpoor inhibitory control we controlled for pre-task hunger (the meanof pre-task hunger on the control and the stress day) In case of asignificant moderator effect for hunger stress reactivity we as-sessed whether this moderator effect is mediated by poor intero-ceptive awareness or the alexithymia aspects difficulty in identifyingfeelings or describing feelings by conducting a mediated modera-tion analysis following the three models elaborated in the presentResults section by Muller Judd and Yzerbyt (2005) To avoid mul-ticollinearity in the regression analyses all variables were cen-tered before computing interaction terms (Aiken amp West 1991)

Results

Preliminary analyses of the participants

Of the total of 60 participants two appeared to erroneously haveno extreme values on emotional eating one person experienced toomuch stress during the stress task and was allowed to premature-ly leave the laboratory on the stress day one person failed to providedata on the stop signal task two persons had missing values on atleast one of their hunger ratings andor missing or invalid cortisoldata So our final sample comprised of 54 participants 25 low emo-tional eaters (LEE) and 29 high emotional eaters (HEE) The meanage of our final sample was 201 years (SD = 232) and the mean BMI(body mass index weightheight times height) was 2117 (SD = 239) (withonly two participants having BMIs gt 25) The mean and standard de-viation (SD) of body mass index and scores on external eating re-strained eating and the Stop Signal Task of the total sample and thesubsamples of HEE vs LEE may be found in Table 1

Manipulation checks

Negative moodThe scales for negative mood at the time points 1ndash4 showed

massive skewness and curtosis Only by computing the differencebetween negative mood in the stress condition and the control con-dition (henceforth delta negative mood) the negative mood valuescould be normalized Positive values of delta negative mood meantmore negative mood in the stress condition Figure 1 shows the meandelta negative mood values at the four time points The finding thatthe mean delta negative mood values showed a peak after the dis-tress manipulation indicates that the stress manipulation had beensuccessful in affecting negative mood (see the supplementary datain the online version at doi101016jappet201404020 for statis-tical information on delta negative mood)

HungerFigure 2 shows the mean hunger values in the control and the

stress condition at the four time points The multivariate test re-vealed a significant overall moderator effect of the stress condi-tion on hunger values over time (F (351) = 4486 p = 007partial eta squared = 209) The finding that the mean hunger valueswere significantly lower after the stress than after the control


manipulation indicates that distress manipulation had been suc-cessful in affecting the self-reported hunger responses (see thesupplementary data in the online version at doi101016jappet201404020 for further information on hunger)

Effects of emotional eating on hunger values over time

Figure 3 shows the mean hunger values of high and low emo-tional eaters in the stress condition The tests of the within sub-jects effects (GreenhousendashGeisser) revealed a significant moderatoreffect of emotional eating on hunger values over time (F (2523131192) = 3740 p = 018 partial eta squared = 067) Though HEEshowed higher hunger ratings in the stress condition than LEE dif-ferences were only borderline significant (Bonferroni corrected) forthe post stress hunger at T2 (18 min) (immediately after the stressor)(mean (sd) of LEE vs HEE 404 (219) vs 507 (234) p = 06 (one-tailed) see the supplementary data in the online version atdoi101016jappet201404020 for further information on the mod-erator effect of emotional eating on hunger)

For impulsivity for both the control and the stress condition therewere no significant effects of impulsivity on hunger values over time(ps gt 10)

Descriptives of hunger food intake and other measures

The mean (SD) of hunger reactivity (AUCi) in the control and thestress condition was minus1456 (728) and minus1411 (617) respectively Themean (SD) of deltaAUCi the hunger reactivity of the stress minusthe control condition was 44 (788) with a range of minus1850ndash1850A total of 463 of the participants (n = 25) had values below zeroThe mean food intake (grams) in the control condition did not differsignificantly from the mean food intake in the stress condition (7247(5778) vs 7279 (6953) p = 96) and the same held true for intakeof snack food (the addition of grams of intake of cake and MampMs)2925 (2563) vs 3119 (3702) p = 65

Table 2 shows the correlations of the AUCi of hunger in the controlcondition the stress condition and the deltaAUCi (ie AUCi of hungerof the stress minus the control condition) with the food intake mea-sures in the control and in the stress condition in addition to thecorrelations with emotional eating and poor inhibitory control

Emotional eating and poor inhibitory control were not inter-related Please note further the significant negative correlations ofAUCi stress and deltaAUCi with the food intake measures in the stress

Table 1Characteristics of the total sample and the subsamples of low and high emotional eaters (LEE and HEE)

Total sample LEE (n = 25) HEE (n = 29) Cohenrsquos da

M SD M SD M SD

BMIb 2117 239 2030 139 2192 280 75External eating 325 71 289 65 357 62 108Restrained eating 238 96 218 105 256 86 40Stop Signal Task 26053 6448 26014 6385 26086 66148 00

a Effect size of the difference between low and high emotional eaters by Cohenrsquos d (20 = small 50 = medium and 80 = large)b Only two participants had a BMI gt 25 and removal of these two participants did not affect any of the obtained results

-02

-01

0

01

02

03

04

05

06

-40 -30 -20 -10 0 10 20 30 40

Delta Negative Mood

Del

ta N

ega t

ive

Moo

d

Time (minutes)

Fig 1 Values of delta negative mood (the difference between negative mood in thestress condition and the control condition) over time showing a peak after the dis-tress manipulation

4

45

5

55

6

65

-40 -30 -20 -10 0 10 20 30 40

Control ConditionStress Condition

Hun

ger

Time (minutes)

Fig 2 The moderator effect of the stress condition (control vs stress) on hungervalues over time showing that the mean hunger values were significantly lower afterthe stress than after the control manipulation (significant differences for T2 T3 andT4 (18 27 and 34 min) see supplementary data in the online version at doi101016jappet201404020)


condition indicating that increased stress-induced food intake is as-sociated with an a-typical hunger stress response (a low hungerstress reactivity means that the typical lower hunger in the stresscondition is less pronounced)

Table 3 shows the correlations of the various AUCis of hunger withpoor interoceptive awareness and the alexithymia aspects difficul-ty identifying and describing feelings Except for difficulty describ-ing feelings which showed a significant negative correlation withthe AUCi of hunger in the control condition (r = minus31 p lt 05) the twoalexithymia aspects or poor interoceptive awareness were not sig-nificantly related to any of the AUCis of hunger This would suggestthat poor interoceptive awareness and the two alexithymia aspectson the one hand and self-reported hunger reactivity on the otherhand may be different and unrelated constructs A further findingwas that poor interoceptive awareness and the alexithymia aspectidentifying feelings but not the alexithymia aspect describing feel-ings were significantly related to emotional eating Finally only thealexithymia aspect difficulty describing feelings was positively relatedto food intake in the control and in the stress condition (r = 31 p lt 05and r = 31 p lt 05 respectively (not shown in Table 3))

Hunger stress reactivity emotional eating and food intake

With hierarchical regression analyses we next tested the inter-action of hunger stress reactivity with emotional eating on delta foodintake and delta snack intake In the analysis with the dependentvariable delta food intake there was no emotional eating times hungerstress reactivity interaction effect (p gt 10) There were however sig-nificant main effects for both emotional eating and hunger stressreactivity (deltaAUCi) (unstandardized begraveta (B)s respectively 32923p = 011 and minus2233 p = 007) We next inspected the nature of thesemain effects on the food intake in the stress vs control conditionfor low vs high emotional eaters and for participants with ana-typical vs typical hunger stress response (classified by means ofa median split a procedure which is however vulnerable tomisclassification of research participants and spurious effectsMaxwell amp Delaney 1993) As in the earlier study on the smallersample (van Strien et al 2012a 2012b) there was a typical mod-erator effect of emotional eating on food intake with low emotion-al eaters eating less after distress than after control (adjusted means[SE]5698 [1189] and 7634 [1181]) and high emotional eaters eatingmore (adjusted means [SE]8224 [1109] and 6982 [1102]) Fur-thermore there also was a moderator effect for hunger stress reac-tivity on food intake with people with a low (a-typical) hunger stressreactivity eating more after distress than after control (adjustedmeans [SE]9032 [1132] and 7554 [1127]) and people with the typ-ically high hunger stress reactivity (reported hunger goes down afterdistress) eating less after distress (adjusted means [SE]4889 [1165]and 7062 [1157])

In the analysis with the dependent variable delta snack intakethere was a significant main effect for emotional eating(unstandardized begraveta (B) 20185 p = 016) in addition to asignificant hunger stress reactivity times emotional eating interactioneffect (B = minus2513 p = 013) This interaction effect remained signif-icant when additionally controlling for external eating and dietaryrestraint (B = minus2461 p = 018) Inspection of the interaction effect inthe initial model of snack food intake according to Holmbeck (2002)revealed that the regression for people with low hunger stress re-activity (a-typical hunger stress response) indicated a positive as-

Table 2Pearson correlations of the hunger-rating reactivity measures (AUCi control AUCi stress and hunger reactivity (deltaAUCi)) emotional eating and poor inhibitory control(SSRT) with food and snack intake in the control and stress condition (SSRT)

AUCi control AUCi stress Delta AUCi Emotional eating Poor inhibitory control

Food intake control minus14 minus16 00 minus06 05Snack intake control 15 minus17 minus27 00 22Food intake stress minus03 minus41 minus29 22 09Snack intake stress 11 minus32 minus36 29 10Emotional eating 14 10 minus05Poor inhibitory control 14 minus05 minus17 01

p lt 05 p lt 01

4

42

44

46

48

5

52

54

56

-40 -30 -20 -10 0 10 20 30 40

Low Emotional EatersHigh Emotional Eaters

Hun

ger

Time (minutes)

Fig 3 The moderator effect of emotional eating (low vs high) on the hunger valuesin the stress condition showing that high emotional eaters had higher hunger ratingsin the stress condition than low motional eaters (but only a borderline significantdifference (p = 06) at T2 (18 min))

Table 3Pearson correlations of poor interoceptive awareness and the two alexithymia aspectsalexithymia identifying feelings and alexithymia describing feelings with the hunger-rating reactivity measures (AUCi control AUCi stress and hunger stress reactivity(deltaAUCi)) emotional eating and poor inhibitory control

Poor interoceptiveawarenessa

Alexithymiaidentifyingfeelings

Alexithymiadescribingfeelings

AUCi control 04 06 minus31AUCi stress 03 00 minus11Delta AUCi minus01 minus04 22Emotional eating 43 36 05Poor inhibitory control minus06 07 06

p lt 05 p lt 01a n = 52


sociation of emotional eating on delta snack intake (B = 40135p = 001) In contrast results of the regression for people with highhunger stress reactivity (typical hunger stress response) indicatedno significant association of emotional eating on delta snack intake(B = 504 p = 963) This result indicates that the moderator effectof emotional eating on stress induced snack intake only holds truefor people with low hunger stress reactivity (an a-typical hungerstress response) Regression lines depicting levels of emotional eatingfor the females with an a-typical versus typical hunger stress re-sponse in relation to delta snack intake are plotted in Fig 4

We next assessed whether the moderator effect of hunger stressreactivity on the association between emotional eating and deltasnack intake is mediated by poor interoceptive awareness and thetwo alexithymia aspects difficulty identifying feelings and diffi-culty describing feelings according to the three models proposedby Muller et al (2005) (see p 856 and Table 3 p 858) The proce-dure and outcomes of the various mediated moderation tests (andassociated tables) can be found in the supplementary data in theonline version at doi101016jappet201404020 In brief the resultssuggest that the moderator effect of hunger stress reactivity on theassociation between emotional eating and delta snack intake wasnot mediated by either poor interceptive awareness or any of thetwo alexithymia aspects difficulty identifying feelings or difficultydescribing feelings to other people

Inhibitory control emotional eating and food intake

With hierarchical regression analyses we subsequently tested theinteraction of inhibitory control with emotional eating on delta foodintake and delta snack intake controlling for pre-task hunger (themean of pre-task hunger in the control and the stress condition)

In the analysis with delta food intake there was no significantmain effect for inhibitory control (p gt 10) There was however a sig-nificant main effect of emotional eating (B = 3632 p = 009) and alsoa significant inhibitory control times emotional eating interaction effect(B = minus430 p = 043) The main effect of emotional eating and the emo-tional eating times inhibitory control interaction effect remained sig-nificant when additionally controlling for external eating and dietaryrestraint (B = 44453 p = 008 B = minus449 p = 036)

Inspection of this interaction effect in the initial model (only con-trolling for pre-task hunger) following Holmbeck (2002) revealedthat the regression for low inhibitory control (high impulsivity) in-dicated no effect of emotional eating on delta food intake (B = 8970p = 630) In contrast results of the regression for high inhibitorycontrol (low impulsivity) indicated a significant effect of emotion-al eating on delta food intake (B = 63571 p = 001) Regression linesdepicting levels of emotional eating for the females with a low versushigh inhibitory control in relation to delta food intake are plottedin Fig 5

Discussion

Self-reported emotional eating has been found to significantlymoderate distress-induced food intake with low emotional eaterseating less after a stress task than after a control task and high emo-tional eaters eating more Aim of the present study was to explorepossible underlying mechanisms by assessing possible associa-tions of this moderator effect with (1) ability to experience the typicalpost-stress reduction of hunger or (2) inhibitory control The mod-erator effect of emotional eating on distress induced food intake wasexpected to be stronger for females with a failure to report the typicalreduction of hunger immediately after a stress task (an a-typical

-20

-10

0

10

20

30

Low High

Typical hunger stress responseA-typical hunger stress response

Del

ta S

nack

Inta

ke

Emotional Eating

Fig 4 Regression lines depicting levels of emotional eating (low vs high) for par-ticipants with the a-typical vs typical hunger stress response in relation to delta snackintake (snack intake (intake of cake plus MampMs in grams) in the stress condition minussnack intake in the control condition) showing that high emotional eaters with ana-typical hunger stress response have relatively increased snack intake after stress

-40

-30

-20

-10

0

10

20

30

Low High

Low Inhibitory ControlHigh Inhibitory Control

Del

ta F

ood

Inta

ke

Emotional Eating

Fig 5 Regression lines depicting levels of emotional eating (low vs high) for par-ticipants with low vs high inhibitory control (low inhibitory control = high scoreson the stop signal task) in relation to delta food intake (food intake ie intake ofcake plus MampMs plus grapes plus carrots in grams in the stress condition minusfood intake in the control condition) showing that high emotional eaters with highinhibitory control have relatively increased snack intake after stress


hunger stress response) or for females with high inability to stopmotor impulses (low inhibitory control)

The results confirmed the previous moderator effect of emo-tional eating on distressed induced food intake in our earlier studyon a smaller part of the present sample (n = 47 vs n = 54) (van Strienet al 2012a) This moderator effect was found to hold true for bothintake of total food and for intake of snacks For intake of snacksbut not for intake of total food there was a significant moderatoreffect of hunger stress reactivity on the moderator effect of emo-tional eating on distress-induced snack intake In line with our ex-pectations the moderator effect of emotional eating on distress-induced snack intake was found to only hold true for participantswith low hunger stress reactivity (an a-typical hunger stress re-sponse of not reporting the typical reduction of hunger after dis-tress) Females who combined a high degree of emotional eatingwith an a-typical hunger stress response showed the highest snackintake For total food intake but not for snack food intake there isalso a significant moderator effect of inhibitory control on the mod-erator effect of emotional eating on distress-induced food intakeContrary to our expectations only participants with high inhibito-ry control (low impulsivity) showed the typical moderator effect ofemotional eating on distress-induced total food intake

The finding that the highest snack intake was found in femaleswho combined a high degree of emotional eating with an a-typicalhunger stress response would suggest that emotion-induced foodintake is only prevalent in females who do not experience the typicalpost-stress reduction of hunger This means that emotional eatingdoes not seem to be an example of lsquoeating in the absence of hungerrsquoIn the same line emotional eating also moderated the hunger valuesafter the stressor with a trend for high emotional eaters to showhigher hunger values immediately after the stressor than low emo-tional eaters (see Gluck Geliebter Hung amp Yahav 2004 for a similarfinding for patients with binge eating disorder ie a condition thatis closely related to emotional eating (van Strien et al 2005))

Precisely why some of our high emotional eaters did not expe-rience the typical post stress reduction of hunger is as yet unclearBruch (1964) suggested that this may be the result of a confusionof physiological symptoms associated with stress and negative emo-tions and those associated with hunger and satiety due to poor in-teroceptive awareness or alexithymia The present data do notsupport this suggestion Neither poor interoceptive awareness noralexithymia were related to hunger stress reactivity nor did theymediate the moderator effect of hunger stress reactivity on the mod-erator effect of emotional eating on distress-induced food intakeIt thus seems that self-reported hunger stress reactivity andalexithymia and poor interoceptive awareness are different and un-related constructs though further research in this respect is neededFor future research into the underlying mechanism of an a-typicalhunger stress response the present moderator effect for hunger stressreactivity per se on total food intake (though not on snack intake)may be of special interest with people with an a-typical hungerstress response eating more food after the stressor than those withthe typical hunger stress response (the typical reduction of hungerafter stress) It would also be of interest to additionally use physi-ological measures of satiety such as glucose level and ghrelin (LoeberGrosshans Herpertz Kiefer amp Herpertz 2013 Raspopow AbizaidMatheson amp Anisman 2010 2014) In this respect it should be notedthat Raspopow et al (2010) found some interesting differencesbetween low vs high (DEBQ) emotional eaters (median split clas-sification) in their ghrelin responses following food intake withghlerin levels declining in the non-emotional eaters but not in theemotional eaters (see also Raspopow et al 2014)

The present finding that only people with high inhibitory control(low impulsivity) showed the moderator effect of emotional eatingon distress-induced food intake was contrary to our expectationsThe absence of a moderator effect for emotional eating in the females

with low inhibitory control (high impulsivity) may possibly be ex-plained by the relatively high food intake after the distress manip-ulation by the low emotional eaters Instead of showing the typicalresponse of eating less food in the stress condition compared withthe control condition they ate similar amounts of food in both con-ditions (that is they showed non-regulation of food intake) Butexactly why low emotional eaters with low inhibitory control (highimpulsivity) showed non-regulation of food intake in the stress con-dition cannot be answered by the present data An explanation interms of a difference in hunger-ratings does not seem very plausi-ble given the absence of a moderator effect of inhibitory control onthe hunger values over time in both the control and the distress con-dition (see the additional results on the hunger-ratings in thesupplementary data in the online version at doi101016jappet201404020) People with low versus high inhibitory controldid not differ in their reported hunger values they both showed thetypical reaction of reporting less hunger after the stress than afterthe control condition so we have to search for other factors that mayunderlie this non-regulation of food intake of the low emotionaleaters with low inhibitory control (high impulsivity)

Is this non-regulation of food intake perhaps an example of eatingin the absence of feelings of hunger (An echo of the old internalndashexternal distinction with people with low inhibitory control beingmore easily tempted by external food stimuli regardless of inter-nal feelings of hunger and satiety (Herman amp Polivy 2008 Schachter1971)) Or is the food intake of the low emotional eaters with lowinhibitory control (high impulsivity) rather ruled by an unknownthird factor Earlier studies have found associations between prob-lems with impulse regulation and enhanced cortisol suppression(Diaz-Marsaacute et al 2008) Associations have also been found betweenmeasures of impulsivity and shortages in brain dopamine (Blum et al2000 Dawe amp Loxton 2004) Perhaps this also holds true for ourpresent females with poor ability to inhibit motor responses In casepoor ability to inhibit motor impulses is indeed associated with areward deficit as outcome of a shortage in brain dopamine peoplewith low inhibitory control may be more receptive to the reinforc-ing value of food to antagonize this dopamine deficit (Volkow et al2003) Similarly negative emotions which have been associated withdecreased dopamine activity in subjects with low dopamine recep-tor availability are more likely to evoke the desire to consume foodto counteract the dopamine deficit (Volkow et al 2003 p 140) Apossible reward deficit in the females with poor ability to inhibitmotor impulses may therefore have had as outcome that they usedfood as lsquoself-medicationrsquo to blunt the decreased dopamine activityas result of the negative emotions in the ego threat condition evenwhen they were low emotional eaters

Contrary to earlier findings with other measures of impulsivitythe present indirect measure for impulsivity (inability to inhibitmotor impulses as measured with the stop signal task) showed nosignificant association with emotional eating As indicated by Racineet al (2013) research into the role of impulsivity in problematic eatingbehaviour (as is emotional eating) is plagued by the fact that im-pulsivity encompasses multiple different constructs ranging fromsensation seeking and lack of perseverance to negative urgency (thetendency to act harshly in response to negative affect) This meansthat present findings are limited to this very indirect measure forimpulsivity and cannot be generalized to other aspects of impul-sivity or impulsivity in its broadest sense Further limitations arethe relatively small sample size which may have reduced the powerto reveal small (interaction) effects Additionally emotional eatinghas been closely associated with binge eating and depressive feel-ings (Ouwens et al 2009 Racine et al 2013 van Strien et al 2005)Therefore it is highly probable that our subjects with high emo-tional eating had other symptomatology such as depressive symp-toms This is a limitation of the present study that should deserveattention in future studies with more participants Also the present


experiment was conducted on predominantly normal weight femaleemotional eaters hence the study needs replication on high andlow emotional eaters with overweight A further limitation is thatwe cannot rule out the possibility that social desirability or acqui-escence may have affected scores on emotional external and re-strained eating Finally the present findings would need replicationoutside the laboratory but in that case other and less complicatedimplicit measures for impulsivity with no need for use of a com-puter need to be developed

A strength of the present study is the use of a within subjectdesign where participants acted as their own control group thoughit is a weakness that the two conditions were not counterbal-anced (for fear of carry over effects andor high dropout) A furtherstrength is that the use of groups with extreme scores on emotion-al eating is associated with higher efficiency of detecting interac-tion effects (McClelland amp Judd 1993) Additionally the use of groupswith extreme scores provides interesting insights into the physiol-ogy and psychology of females with (pre-)clinical levels ofdisregulated eating A further strength is that in all analyses we con-trolled for the other two eating styles so the present moderator effectof impulsivity on the interaction effect of emotional eating on stress-induced food intake seems robust for the possible confounding effectsof external and restrained eating

Conclusion

Failure to report the typical reduction of hunger after stress (ana-typical hunger stress response) but not poor inhibitory control(high scores on the stop signal test) are possible mechanisms un-derlying distress induced (emotional) overeating Emotional eatingthus does not seem an example of lsquoeating without hungerrsquo

Appendix Supplementary material

Supplementary data to this article can be found online atdoi101016jappet201404020

References

Aiken L S amp West S G (1991) Multiple regression Testing and interpreting interactionsNewbury Park Sage Publication

Bagby R Parker J D A amp Taylor G J (1994) The twenty-item Toronto AlexithymiaScale I Item selection and cross-validation of the factor structure Journal ofPsychosomatic Research 38 23ndash32

Bekker M H van de Meerendonk C amp Mollerus J (2004) Effects of negative moodinduction and impulsivity on self-perceived emotional eating The InternationalJournal of Eating Disorders 36 461ndash469

Blum K Braverman E R Holder J M Lubar J F Monastra V J Miller D et al(2000) Reward deficiency syndrome A biogenetic model for the diagnosis andtreatment of impulsive addictive and compulsive behaviours Journal ofPsychoactive Drugs 32 1ndash68

Bohon C Stice E amp Spoor S (2009) Female emotional eaters showabnormalities in consummatory and anticipatory food reward A functionalmagnetic resonance imaging study The International Journal of Eating Disorders42 210ndash221

Bruch H (1964) Psychological aspects in overeating and obesity Psychosomatics 5269ndash274

Dawe S amp Loxton N J (2004) The role of impulsivity in the development ofsubstance use and eating disorders Neuroscience and Biobehavioral Reviews 28343ndash351

Diaz-Marsaacute M Carrasco J L Basurte E Saacutelz J Loacutepez-Ibor J J amp Hollander E(2008) Enhanced cortisol suppression in eating disorders with impulsivepersonality features Psychiatry Research 158 93ndash97

Ebneter D Latner J Rosewall J amp Chisholm A (2012) Impulsivity in restrainedeaters Emotional and external eating are associated with attentional and motorimpulsivity Eating and Weight Disorders 17 e62ndashe65

Elfhag K amp Morey L C (2008) Personality traits and eating behaviour in the obesePoor self-control in emotional and external eating but personality assets inrestrained eating Eating Behaviors 9 285ndash293

Garner D M (1991) Eating disorder inventory-2 Professional manual Odessa FLPsychological Assessment Resources

Gibson E L (2012) The psychobiology of comfort eating Implications forneuropharmacological interventions Behavioral Pharmacology 23 442ndash460

Gluck M E Geliebter A Hung J amp Yahav E (2004) Cortisol hunger and desireto binge eat following a cold stress test in obese women with binge eatingdisorder Psychosomatic Medicine 66 876ndash881

Gold P W amp Chrousos G P (2002) Organization of the stress system and itsdysregulation in melancholic and atypical depression High vs low CRHNE statesMolecular Psychiatry 7 254ndash275

Greeno C G amp Wing R R (1994) Stress induced eating Psychological Bulletin 115444ndash464

Herman C P amp Polivy J (2008) External cues in the control of food intake in humansThe sensory-normative distinction Physiology amp Behavior 94 722ndash728

Holmbeck G N (2002) Post-hoc probing of significant moderational and meditationaleffects in studies of pediatric populations Journal of Pediatric Psychology 2787ndash96

Jansen A Nederkoorn C van Baak L Keirse C Guerrieri R amp Havermans R (2009)High restrained eaters only overeat when they are also impulsive BehaviourResearch and Therapy 47 105ndash110

Jasinska A J Yasuda M Burant B Gregor G Khatri S Sweet M et al (2012)Impulsivity and inhibitory control deficits are associated with unhealthy eatingin young adults Appetite 59 738ndash747

Kirschbaum C Pirke K M amp Hellhammer D H (1993) The amplsquoTrier Social StressTestrsquo A tool for investigating psychosocial stress responses in a laboratory settingNeuropsychobiology 28 76ndash81

Kooiman C G Spinhoven P amp Trijsburg R W (2002) The assessment of alexithymiaA critical review of the literature and a psychometric study of the TorontoAlexithymia Scale-20 Journal of Psychosomatic Research 53 1085ndash1090

Larsen J K van Strien T Eisinga R amp Engels R C M E (2006) Gender differencesin the association between alexithymia and emotional eating in obese individualsJournal of Psychosomatic Research 60 237ndash243

Lattimore P Fisher N amp Malinowski P (2011) A cross-sectional investigation oftrait disinhibition and its association with mindfulness and impulsivity Appetite56 241ndash248

Loeber S Grosshans M Herpertz S Kiefer F amp Herpertz S C (2013) Hungermodulates behavioral disinhibition and attention allocation to food-associatedcues in normal-weight controls Appetite 71 32ndash39

Logan G D Schachar R J amp Tannock R (1997) Impulsivity and inhibitory controlPsychological Science 8 60ndash64

Maxwell S E amp Delaney H D (1993) Bivariate median splits and spurious statisticalsignificance Psychological Bulletin 113 181ndash190

McClelland G H amp Judd C M (1993) Statistical difficulties of detecting interactionsand moderator effects Psychological Bulletin 114 376ndash390

Muller D Judd C M amp Yzerbyt V Y (2005) When moderation is mediated andmediation is moderated Journal of Personality and Social Psychology 6 852ndash863

Nederkoorn C Houben K Hofmann W Roefs A amp Jansen A (2010) Controlyourself or just eat what you like Weight gain over a year is predicted by aninteractive effect of response inhibition and implicit preference for snack foodsHealth Psychology 29 389ndash393

Nolan L J Halperin L B amp Geliebter A (2010) Emotional Appetite QuestionnaireConstruct validity and relationship with BMI Appetite 54 314ndash319

Oliver G Wardle J amp Gibson E L (2000) Stress and food choice A laboratory studyPsychosomatic Medicine 62 853ndash865

Ouwens M A van Strien T amp van Leeuwe J F J (2009) Possible pathways betweendepression emotional and external eating A structural equation model Appetite53 245ndash248

Pinaquy S Chabrol H Simon C Loevet J P amp Barber P (2003) Emotional eatingalexithymia and binge eating disorder in obese women Obesity Research 11195ndash201

Pruessner J C Kirschbaum C Meinlschmid G amp Hellhammer D H (2003) Twoformulas for computation of the area under the curve represent measures of totalhormone concentration versus time-dependent change Psychoneuroendocrinology28 916ndash931

Racine S E Burt A A Sisk C L Boker S Keel P K Neale M et al (2013) Exploringthe relationship between negative urgency and dysregulated eating Etiologicassociations and the role of negative affect Journal of Abnormal Psychology 122433ndash444

Raspopow K Abizaid A Matheson K amp Anisman H (2010) Psychological stressoreffects on cortisol and ghrelin in emotional and non-emotional eaters Influenceof anger and shame Hormones and Behavior 58 677ndash684

Raspopow K Abizaid A Matheson K amp Anisman H (2014) Anticipation of aphysiological stressor differentially influences ghrelin cortisol and food intakeamong emotional and non-emotional eaters Appetite 74 35ndash43

Schachter S (1971) Some extraordinary facts about obese humans and rats TheAmerican Psychologist 26 129ndash144

Schoemaker C van Strien T amp van der Staak C P F (1994) Validation of the EatingDisorder Inventory in a non-clinical population using transformed anduntransformed responses The International Journal of Eating Disorders 15387ndash393

Stone A A amp Brownell K D (1994) The stress eating paradox Multiple dailymeasurements in adult males and females Psychology and Health 8 425ndash436

Taylor G J Parker J D A Bagby R M amp Bourke M P (1996) Relationships betweenalexithymia and psychological characteristics associated with eating disordersJournal of Psychosomatic Research 412 561ndash568

Tomiyama A J Dallman M F amp Epel E S (2011) Comfort food is comforting tothose most stressed Evidence of the chronic stress response network in highstress women Psychoneuroendocrinology 36 1513ndash1519


Tryon M S DeCant R amp Laugero K D (2013) Having your cake and eating it tooA habit of comfort food may link chronic social stress exposure and acutestress-induced cortisol hyporesponsiveness Physiology amp Behavior 114ndash115 32ndash37

van Strien T (2010) Dutch eating behaviour questionnaire Manual Amsterdam BoomTest Publishers

van Strien T Cebolla A Etchemendy E Gutieumlrez-Maldonado J Ferrer-Garcia MBotella C et al (2013a) Emotional eating and food intake after sadness and joyAppetite 66 20ndash25

van Strien T Engels R C M E van Leeuwe J amp Snoek H M (2005) The Stice modelof overeating Tests in clinical and non-clinical samples Appetite 45 205ndash213

van Strien T Herman C P amp Anschutz D (2012b) The predictive validity of theDEBQ-external eating scale for eating in response to food commercials whilewatching television The International Journal of Eating Disorders 45 257ndash262

van Strien T Herman C P Anschutz D Engels R C M E amp de Weerth C (2012a)Moderation of distress-induced eating by emotional eating scores Appetite 58277ndash284

van Strien T amp Ouwens M A (2003) Counterregulation in female obese emotionaleaters Schachter Goldman and Gordonrsquos (1968) test of psychosomatic theoryrevisited Eating Behaviors 3 329ndash340

van Strien T amp Ouwens M A (2007) Effects of distress alexithymia and impulsivityon eating Eating Behaviors 8 251ndash257

van Strien T Roelofs K amp de Weerth C (2013b) Cortisol reactivity and distress-induced emotional eating Psychoneuroendocrinology 38 677ndash684

van Strien T amp van de Laar F A (2008) Intake of energy is best predicted byovereating tendency and consumption of fat is best predicted by dietary restraintA 4-year follow-up of patients with newly diagnosed type-2 diabetes Appetite50 544ndash547

Volkow N D Wang G J Maynard L Jayne M Fowler J S Zhu W et al (2003)Brain dopamine is associated with eating behaviours in humans The InternationalJournal of Eating Disorders 33 136ndash142

Wallis D J amp Hetherington M M (2004) Stress and eating The effects of ego-threatand cognitive demand on food intake in restrained and emotional eaters Appetite43 39ndash46

Watson D Clark L A amp Tellegen A (1988) Development and validation of briefmeasures of positive and negative affect The PANAS scales Journal of Personalityand Social Psychology 54 1063ndash1070

Westerterp-Plantenga M S (2004) Effects of energy density of daily food intake onlong-term energy intake Physiology amp Behavior 81 765ndash771

Whisman M A amp McClelland G H (2005) Designing testing and interpretinginteractions and moderator effects in family research Journal of Family Psychology19 111ndash120


tasks High emotional eaters with a blunted cortisol response atemore food after distress than those with an elevated cortisol stressresponse (see Tomiyama Dallman amp Epel 2011 and Tryon DeCantamp Laugero 2013 for similar findings) This finding suggests that emo-tional over-eating is indeed associated with a lowered HPA-axis func-tioning though it is as yet unclear whether this blunted cortisolresponse is secondary (ie the result of an adaptive down regula-tion) or primary to emotional eating (see van Strien et al 2013b)In the present study on the same sample plus additional participants2

we want to pursue our search for possible mechanisms further thistime by addressing the relationship of emotional eating with (1) self-reports of hunger after stress versus control and (2) inhibitory control

Emotional over-eating may be the result of a failure to experi-ence the typical post-stress reduction of hunger An earlier studyexamined the hunger ratings immediately after a distress vs controlcondition of low versus high emotional eaters (van Strien et al2012a) It was found that high emotional eaters in contrast to thelow emotional eaters did not report a substantial reduction of hungerimmediately after the stress task compared with the control taskThis finding suggests that high emotional eaters in contrast to thelow emotional eaters indeed suffer from an a-typical hunger stressresponse Whether this a-typical hunger stress response also pre-dicts distress-induced food intake was however not investigatedin that study Furthermore there are to our best knowledge no pub-lished studies that have systematically tested the relationshipbetween hunger stress reactivity and actual stress-induced foodintake in high and low emotional eaters The present study is there-fore the first to systematically examine whether the moderator effectof emotional eating on distress-induced food intake is affected bythe individualrsquos hunger stress reactivity

Failure to experience the typical post-stress reduction of hungerof high emotional eaters may be the result of poor interoceptiveawareness that is a confusion of physiological symptoms associ-ated with stress and negative emotions and those associated withhunger and satiety (Bruch 1964) Poor interoceptive awareness ishighly associated with alexithymia (Greek literally meaning lsquonowords for emotionsrsquo) most notably the alexithymia aspects of dif-ficulty in identifying feelings and describing feelings to other people(Garner 1991 Taylor Parker Bagby amp Bourke 1996) The associa-tion of poor interoceptive awareness and alexithymia with self-reported emotional eating has been widely established (Larsen vanStrien Eisinga amp Engels 2006 Pinaquy Chabrol Simon Loevet ampBarber 2003 van Strien Engels van Leeuwe amp Snoek 2005) within one study a moderator effect of alexithymia on actual distress-induced food intake (van Strien amp Ouwens 2007) However whetherpoor interoceptive awareness and alexithymia are also associatedwith hunger stress reactivity (typical versus a-typical hunger re-sponses after stress versus control) has not yet been assessed Sim-ilarly it is as yet unknown whether a possible moderator effect ofhunger stress reactivity on the moderator effect of emotional eatingon distress-induced food intake is mediated by poor interoceptiveawareness and alexithymia

Inability to inhibit motor impulses (pressing a button) ndash a measurefor inhibitory control which has been considered an indirect measurefor impulsivity (Jansen et al 2009) ndash would also be a good candi-date for affecting the moderator effect of emotional eating ondistress-induced food intake In a series of studies summarized byJansen et al (2009) inability to inhibit motor impulses as mea-

sured with the stop signal task (by Logan Schachar amp Tannock 1997)predicted higher food intake a higher body weight and less weightloss after a weight reduction treatment However in more recentstudies inability to stop motor impulses only predicted overeatingand weight gain in combination with other factors such as a highimplicit preference for the test food high dietary restraint or strongfeelings of hunger (Jansen et al 2009 Nederkoorn HoubenHofmann Roefs amp Jansen 2010) Emotional eating could also be sucha factor given the association of emotional eating with various self-report measures of reduced inhibitory control or negative urgency(the tendency to act impulsively in the face of negative affect)(Bekker van de Meerendonk amp Mollerus 2004 Ebneter LatnerRosewall amp Chisholm 2012 Elfhag amp Morey 2008 Lattimore Fisheramp Malinowski 2011 Ouwens van Strien amp van Leeuwe 2009 Racineet al 2013) and the number of false alarms in a GoNoGo task ameasure of efficiency of inhibitory control in the presence of foodcues (Jasinska et al 2012) We know of no published study thattested the relationship between inability to stop motor impulses andactual stress-induced food intake in high and low emotional eatersHence to the best of our knowledge this would be the first studyto assess the effect of this implicit measure of impulsivity on themoderator effect of emotional eating on distress-induced food intake

This study aims at exploring the relationships of (1) hunger stressreactivity and (2) inhibitory control with actual distress-induced foodintake in high versus low emotional eaters As in most previousstudies on emotional eating the focus is on females and on eatingin response to negative emotions3 (see also Gibson 2012) Earlierwe found that emotional eating significantly moderated the distress-food intake association with low emotional eaters eating less afterthe stress than after the control task and high emotional eaters eatingmore For the present study we hypothesized that this moderatoreffect of emotional eating would be more likely to hold true for par-ticipants with an a-typical hunger stress response (failure to showthe typical reduction of hunger after stress) and for participants withhigh inability to stop motor impulses (poor inhibitory control) Spe-cifically participants who combined an a-typical hunger stress re-sponse or poor inhibitory control with high emotional eating wereexpected to eat more after the stressor than those with the typicalhunger stress response and adequate inhibitory control We left itan open empirical question whether this possible moderator effectof hunger stress reactivity on the moderator effect of emotionaleating on distress-induced food intake would be mediated by poorinteroceptive awareness or alexithymia

Methods

Design

We used a within-subjects design in which females who werepreselected on the basis of extremely high or low scores on an emo-tional eating questionnaire were subjected to a control task and astress task (Trier Social Stress Task ndash TSST) on 2 consecutive daysThe TSST involves speaking in front of a jury coupled with an ar-ithmetic challenge Because the stress condition is perceived by somesubjects as quite stressful we deliberately chose to always start with

2 Results of a sub-sample of 47 participants the data of which had been collect-ed in spring and autumn of 2010 have been reported in two previous publications(van Strien et al 2012a Study 2 van Strien et al 2013b) They address respective-ly the moderation of distress-induced eating by emotional eating scores and cor-tisol reactivity and distress-induced emotional eating The data of the additionalparticipants for the present study were collected in spring 2012

3 While recognizing that eating in response to positive emotions may occur as fre-quently as eating in response to negative emotions there is increasing evidence thateating in response to negative and to positive emotions may refer to different con-structs (Nolan Halperin amp Geliebter 2010 van Strien et al 2013a) One reason maybe that we tend to celebrate happy events with food Eating in response to positiveemotions unlike eating in response to negative emotions can therefore not be con-sidered an lsquoinaptrsquo or lsquoinappropriatersquo response Accordingly it can be expected thatself-assessed emotional eating in response to positive emotions has a main effecton food intake whereas self-assessed emotional eating in response to negative emo-tions only shows an effect on food intake in interaction with distress



Participants


Procedure





















Response inhibition





Results



Manipulation checks
















M SD M SD M SD



-02

-01

0

01

02

03

04

05

06

-40 -30 -20 -10 0 10 20 30 40

Delta Negative Mood

Del

ta N

ega t

ive

Moo

d

Time (minutes)


4

45

5

55

6

65

-40 -30 -20 -10 0 10 20 30 40


Hun

ger

Time (minutes)











p lt 05 p lt 01

4

42

44

46

48

5

52

54

56

-40 -30 -20 -10 0 10 20 30 40


Hun

ger

Time (minutes)







p lt 05 p lt 01a n = 52








Discussion


-20

-10

0

10

20

30

Low High


Del

ta S

nack

Inta

ke

Emotional Eating


-40

-30

-20

-10

0

10

20

30

Low High


Del

ta F

ood

Inta

ke

Emotional Eating














Conclusion




References





























































Participants


Procedure





















Response inhibition





Results



Manipulation checks
















M SD M SD M SD



-02

-01

0

01

02

03

04

05

06

-40 -30 -20 -10 0 10 20 30 40

Delta Negative Mood

Del

ta N

ega t

ive

Moo

d

Time (minutes)


4

45

5

55

6

65

-40 -30 -20 -10 0 10 20 30 40


Hun

ger

Time (minutes)











p lt 05 p lt 01

4

42

44

46

48

5

52

54

56

-40 -30 -20 -10 0 10 20 30 40


Hun

ger

Time (minutes)







p lt 05 p lt 01a n = 52








Discussion


-20

-10

0

10

20

30

Low High


Del

ta S

nack

Inta

ke

Emotional Eating


-40

-30

-20

-10

0

10

20

30

Low High


Del

ta F

ood

Inta

ke

Emotional Eating














Conclusion




References
































































Response inhibition





Results



Manipulation checks
















M SD M SD M SD



-02

-01

0

01

02

03

04

05

06

-40 -30 -20 -10 0 10 20 30 40

Delta Negative Mood

Del

ta N

ega t

ive

Moo

d

Time (minutes)


4

45

5

55

6

65

-40 -30 -20 -10 0 10 20 30 40


Hun

ger

Time (minutes)











p lt 05 p lt 01

4

42

44

46

48

5

52

54

56

-40 -30 -20 -10 0 10 20 30 40


Hun

ger

Time (minutes)







p lt 05 p lt 01a n = 52








Discussion


-20

-10

0

10

20

30

Low High


Del

ta S

nack

Inta

ke

Emotional Eating


-40

-30

-20

-10

0

10

20

30

Low High


Del

ta F

ood

Inta

ke

Emotional Eating














Conclusion




References






































































M SD M SD M SD



-02

-01

0

01

02

03

04

05

06

-40 -30 -20 -10 0 10 20 30 40

Delta Negative Mood

Del

ta N

ega t

ive

Moo

d

Time (minutes)


4

45

5

55

6

65

-40 -30 -20 -10 0 10 20 30 40


Hun

ger

Time (minutes)











p lt 05 p lt 01

4

42

44

46

48

5

52

54

56

-40 -30 -20 -10 0 10 20 30 40


Hun

ger

Time (minutes)







p lt 05 p lt 01a n = 52








Discussion


-20

-10

0

10

20

30

Low High


Del

ta S

nack

Inta

ke

Emotional Eating


-40

-30

-20

-10

0

10

20

30

Low High


Del

ta F

ood

Inta

ke

Emotional Eating














Conclusion




References




































































p lt 05 p lt 01

4

42

44

46

48

5

52

54

56

-40 -30 -20 -10 0 10 20 30 40


Hun

ger

Time (minutes)







p lt 05 p lt 01a n = 52








Discussion


-20

-10

0

10

20

30

Low High


Del

ta S

nack

Inta

ke

Emotional Eating


-40

-30

-20

-10

0

10

20

30

Low High


Del

ta F

ood

Inta

ke

Emotional Eating














Conclusion




References


































































Discussion


-20

-10

0

10

20

30

Low High


Del

ta S

nack

Inta

ke

Emotional Eating


-40

-30

-20

-10

0

10

20

30

Low High


Del

ta F

ood

Inta

ke

Emotional Eating














Conclusion




References







































































Conclusion




References




























































Hunger, inhibitory control and distress-induced emotional eating

Documents

Transcript of Hunger, inhibitory control and distress-induced emotional eating