RESEARCHARTICLE Food Preference and Appetite after Switching between Sweet and Savoury Odours in Women MariëlleG.Ramaekers1,PieternelA.Luning1,CatrionaM.M.Lakemond1,MartinusA.J. S.vanBoekel1,GerritGort2,SanneBoesveldt3* 1 FoodQualityandDesign,WageningenUniversity,POBox17,6700AAWageningen,TheNetherlands, 2 Biometris,WageningenUniversityandResearchCentre,POBox100,6700ACWageningen,The Netherlands,3 DivisionofHumanNutrition,WageningenUniversity,POBox8129,6700EVWageningen, TheNetherlands *[email protected] a11111 Abstract Background Exposuretofoododoursincreasestheappetiteforcongruentfoodsanddecreasesthe appetiteforincongruentfoods.However,theeffectofexposuretoavarietyoffoododours, OPENACCESS asoftenoccursindailylife,isunknown. Citation:RamaekersMG,LuningPA,Lakemond CMM,vanBoekelMAJS,GortG,BoesveldtS(2016) Objective FoodPreferenceandAppetiteafterSwitching Investigatehowswitchingbetweensweetandsavouryodoursaffectstheappetiteforsweet betweenSweetandSavouryOdoursinWomen. PLoSONE11(1):e0146652.doi:10.1371/journal. andsavouryproducts. pone.0146652 Editor:JohnI.Glendinning,BarnardCollege, Design ColumbiaUniversity,UNITEDSTATES Thirtywomen(age:18-45y;BMI:18.5-25kg/m2)intenselysmelledthecontentsofcupsfilled Received:September7,2014 withbanana,meatorwater(no-odour)inawithin-subjectdesignwithfourcombinations: Accepted:December21,2015 no-odour/banana,no-odour/meat,meat/bananaandbanana/meat.Participantsreceived Published:January11,2016 onecombinationpertestday.Ineachcombination,twocupswithdifferentfillingswere smelledforfiveminutesaftereachother.Treatmentorderwasbalancedasmuchaspossi- Copyright:©2016Ramaekersetal.Thisisanopen accessarticledistributedunderthetermsofthe ble.Theeffectsofpreviousexposureandcurrentodourontheappetitefor(in)congruent CreativeCommonsAttributionLicense,whichpermits sweetandsavouryproducts,andodourpleasantnesswereanalysed.Achangefrommeat unrestricteduse,distribution,andreproductioninany tobananaodourorbananatomeatodourwasreferredtoasswitch,whereasachange medium,providedtheoriginalauthorandsourceare fromno-odourtomeatodourorno-odourtobananaodourwasno-switch. credited. DataAvailabilityStatement:Allrelevantdataare Results withinthepaperanditsSupportingInformationfiles. Thecurrentodour(P<0.001),asopposedtothepreviousexposure(P=0.71),determined Funding:Theauthorshavenosupportorfundingto report. theappetitefor(in)congruentsweetandsavouryproducts,alreadyoneminuteaftera switchbetweensweetandsavouryodours.Thepleasantnessoftheodourdecreaseddur- CompetingInterests:Theauthorshavedeclared thatnocompetinginterestsexist. ingodourexposure(P=0.005). PLOSONE|DOI:10.1371/journal.pone.0146652 January11,2016 1/16 FoodPreferenceandAppetiteafterSwitchingOdours Conclusions Afteraswitch,theappetiteforspecificproductsquicklyadjustedtothenewodourandfol- lowedthetypicalpatternasfoundduringodourexposureinpreviousstudies.Interestingly, theappetiteforthesmelledfoodremainedelevatedduringodourexposure,knownassen- sory-specificappetite,whereasthepleasantnessoftheodourdecreasedovertime,previ- ouslytermedolfactorysensory-specificsatiety.Thisseemingcontradictionmayresultfrom differentmechanismsunderlyingtheodour-inducedanticipationoffoodintakeversusthe decreaseinhedonicvalueduringprolongedsensorystimulation. Introduction Unhealthyeatinghabitssuchasunhealthyfoodchoicesorovereatingincreasetheprevalence ofobesity,diabetes,cancer,cardiovascularandotherdiseases[1–3].Therefore,itisimportant tounderstandhowvariousfactors,forexamplesensoryprocesses,influenceoureatingbehav- iour.Sensoryprocessesplayaroleinfoodselectioninseveralways.First,associationsbetween thenutrientcompositionandthesensorypropertiesoffoods,suchasappearance,smelland taste,areformedduetorepeatedexposureinourdailylives[4].Theseassociationspartly determinethepleasurethatisderivedfromfoods[5,6],whereuponpleasantnessinfluences foodselection.Moreover,theseassociationsalsofacilitatetheestimationofthenutrientcom- positionoffoodsbasedonthesensoryproperties[4]andthisinformationcanbeusedforfood selection,forexampleincaseofnutrientdeficits[7,8].Furthermore,recentlyconsumedfoods modulatefoodpreference,whichislikelydrivenbytheneedforvariety.Forexample,thepref- erenceforsavouryproductsdecreasesaftereatingasavourymeal,aphenomenonreferredto assensory-specificsatiety[9,10].Finally,externalfactorssuchasexposuretosight,tasteor smelloffoodschangeourfoodpreference[11–15].Ithasbeenwidelydemonstratedthatexpo- suretofoodcuesincreasesthepreferenceforthecuedfood[14–20].Forexample,sweetodours increasedtheappetiteforsweetproductsandsavouryodourstheappetiteforsavouryproducts [17,18].Moreover,sweetodoursalsodecreasedtheappetiteforsavouryproductsandsavoury odoursforsweetproducts[17,18]. Theprecisemechanismbehindthesefindingshasnotyetbeenelucidated,althoughwehave proposedthatthisincreaseinappetiteforcongruentfoodsanddecreaseinappetiteforincon- gruentfoodsmaybecausedbycephalicphaseresponses[17].Cephalicphaseresponsespre- parethebodyfortheintakeanddigestionoffoods[21–23]andareelicitedbyfoododoursand otherfoodcues.Ingeneral,sweetnessisassociatedwithsugarcontentandsavourinesswith proteincontent[4,7],withdistinctroutesofdigestionfordifferentmacronutrients.Therefore, determiningthetypeoffoodbyexposuretofoodcuesbeforeingestion,maypreparethe bodyforthedigestionofthespecificmacronutrientsoftheanticipatedfoods[17,24].It maybethatoncethebodyispreparedfortheintakeofafoodwithacertain(macro) nutrientcomposition,itislessfavourabletoingestafoodwithaverydifferent(macro) nutrientcomposition[17]. Indailylifethough,forexamplebystrollingthrougha(super)market,exposuretoavariety offoodcuesthatprimeforawidevarietyoffoods,mayinduceconfusioninthebody.Previous exposurestofoodcuesmaypossiblyinterferewithexposurestonewfoodcues.Ifourbody indeedspecificallypreparesfortheintakeofcuedfoods,thenitmaytakesometimetoswitch theappetitefor(in)congruentfoodsaccordingtothecharacteristicsofnewfoodcues. PLOSONE|DOI:10.1371/journal.pone.0146652 January11,2016 2/16 FoodPreferenceandAppetiteafterSwitchingOdours Toourknowledge,theeffectofswitchingbetweendifferentfoodcuesongeneralappetite andfoodpreferencehasnotbeeninvestigatedbefore.Theobjectiveofthecurrentstudywasto determinehowswitchingbetweensweetandsavouryfoododoursaffectstheappetiteforsweet andsavouryproducts,foodpreferenceandgeneralappetite.Generalappetiteandtheappetite forsweetandsavouryproductsweremeasuredatseveraltimepointsduringodourexposureto exploreifpossiblechangesafterswitchingareimmediateortaketime.Theresultscouldpro- videinsightintheprocessesbehindtheeffectoffoodcuesonfoodpreferenceinreallife. MaterialsandMethods Participants Thirtyhealthywomenaged18–45(21.6±4.7)yandBMI18.5–25(21.9±1.3)kg/m2were recruitedfromWageningenandsurroundingsbetween7and14February2013,byusingane- maillistwithpotentialparticipantsthatwasset-upbytheDivisionofHumanNutrition, WageningenUniversity(Fig1).Exclusioncriteriawere:dislikeforbanana,bananapie(Dutch pastry),steamedmeatorbeefsoup(score<5onanine-pointscale),smoking,pregnancyor Fig1.Flowchartthatshowstheparticipantenrolmentinthisstudy. doi:10.1371/journal.pone.0146652.g001 PLOSONE|DOI:10.1371/journal.pone.0146652 January11,2016 3/16 FoodPreferenceandAppetiteafterSwitchingOdours breastfeedingduringthelastsixmonths,lackofappetite,followinganenergy-restricteddiet orchangeinbodyweight>5kgduringthelast2months,hypersensitivitytoanyofthefoods understudyorbeingavegetarian.Itwasexplainedtotheparticipantsthattheinfluenceofsen- sorysignalsonfoodchoicewasinvestigated.Afterthestudy,theparticipantswereinformed aboutthefullstudyobjectives.Twenty-nineparticipantscompletedthestudyandonepartici- pantmissedthelastsession(no-odour/meat)duetoillness.Allparticipantssignedaninformed consentformbeforeparticipation.AllprocedureswereinaccordancewiththeHelsinkiDecla- rationof1975(asrevisedin1983).ThisstudywasexemptfromapprovalbytheMedicalEthi- calCommitteeofWageningenUniversity,becausetherulesofconductinthepresentresearch werenotinterventionist,withverylittleburdentotheparticipants.Thisstudywasregisteredat theDutchtrialregister(NTR3830;www.trialregister.nl/trialreg/admin/rctview.asp?TC=3830). Theparticipantsreceivedacompensationof25euro. Experimentaldesign Theparticipantsintenselysmelledthecontentsofcupsfilledwithbanana,meatorwater(no- odour)inawithin-subjectdesignwithfourcombinations:no-odour/banana,no-odour/meat, meat/bananaandbanana/meat(Fig2).Eachparticipantwasexposedtoonecombinationwith twosuccessiveodourexposuresandaone-minutebreakinbetweenexposures,oneachtest day.Thesefourcombinationsweresetuptocomparetheappetiteforspecificproductsaftera switchwiththeappetiteforspecificproductsafterno-switch(no-odourascontrol).Forexam- ple,tocomparetheappetiteforbanana,duringsmellingbananaaftersmellingno-odour(no- switch),withtheappetiteforbanana,duringsmellingbananaaftersmellingmeat(switch).We usedno-odour/bananaandno-odour/meatascontrolsinsteadofbanana/bananaandmeat/ meat,toexcludeundesireddifferencesinadaptationorspecificsatiationduetodifferencesin exposuretimetoaspecificodour.Eachodourwassmelledintenselyforfiveminutes.Appetite measurementsweretakenat1and5minutesduringexposuretothefirstodourandat7and 11minutesduringexposuretothesecondodour.Theorderofthecombinationswasbalanced overtheparticipants,andasmuchaspossibleoverthetestdays.Eachparticipantwassched- uledonfourseparatedays,preferablyonceperweekatthesametimeofthedaybetween11.20 hand13.40h.Theeffectsofthepreviousexposureandthecurrentodourontheappetitesfor (in)congruentsweetandsavouryproducts,foodpreference,generalappetiteandodourpleas- antnesswereanalysed.Achangefrommeattobananaodourorfrombananatomeatodour wasreferredtoasswitchbetweenodours,whereasachangefromno-odourtomeatodouror no-odourtobananaodourwasno-switch(Fig2). Odours Theparticipantssmelledthecontentsofcupsthatcontainedeither10gofwater(no-odour),a tablespoonofmediumripemashedbanana(bananaodour)oratablespoonofwarmsteamed meat(stoofvlees,Coertjens,Belgium;meatodour).Atissueandaplasticlidwereplacedover thefillingtopreventvisualcuesandodourcontaminationinthesensoryroom. Standardizinghungerstate Thevisitsofeachparticipantwerescheduledatthesametimeofthedaytostandardizethe individualhungerstate.Onthefirsttestday,participantswereinstructedtoconsumeanormal amountofbreakfast,atleast2.5hoursbeforethestartoftheexperiment.Afterthistime,only waterandweakteawereallowed.Ontheremainingtestdays,theparticipantswererequested toconsumethesamebreakfastasonthefirsttestday.Thediariesinwhichparticipants PLOSONE|DOI:10.1371/journal.pone.0146652 January11,2016 4/16 FoodPreferenceandAppetiteafterSwitchingOdours Fig2.Schematicrepresentationoftheexperimentaldesignincludingexplanationoftheterminologyusedinthedataanalysisandresultsections. *Currentodourispresentlysmelledbytheparticipants,whichcanbeeitherthefirstorthesecondodourinacombination.**Previousexposureisthe exposurethatprecededthecurrentodour.***Codeisusedintheresultssectionanddefinesthecombinationofpreviousexposureandcurrentodour.**** Exposuregroupisusedforthedata-analysis. doi:10.1371/journal.pone.0146652.g002 recordedtheirbreakfast,changesinphysicalactivityandhealthproblemswerecheckedfor possibleconfoundersandtoincreasecommitmenttothestudyrules. Procedure Uponthefirstarrival,theparticipantswereinformedontheprocedure,includingademon- strationofhowtosmellintenselybyplacingthenoseabovethecupandbreathingthroughthe nose.Theparticipantsthensettleddowninsensoryboothswithoverpressure(Biotechnionof WageningenUniversity)withtwocupsinfrontofthem.Instructionsweregivenonacom- puterscreen(E-prime,v2.0).Participantsfirstfilledouttheappetitequestionnaireatbaseline (t=0).Subsequently,theywererequestedtoremovetheplasticlidfromthefirstcup,while keepingthetissueonthecup,andthenintenselysmellthecontentofthefirstcup.Theappetite PLOSONE|DOI:10.1371/journal.pone.0146652 January11,2016 5/16 FoodPreferenceandAppetiteafterSwitchingOdours questionnairewasrepeated1and5minutesafterthestartofthesmelling.Theparticipants wereencouragedbythetextonthecomputertocontinueintensesmellingduringthewhole fiveminutesofodourexposure.Afterfiveminutes,theparticipantsplacedbackthelidonthe firstcupandhadaone-minutebreak.Subsequently,theyfollowedthesameprocedurewith thesecondcupandfilledouttheappetitequestionnaireatt=7and11minutes.Afterfivemin- utesofintenselysmellingthecontentofthesecondcup,participantsplacedbackthelidand filledoutthefoodpreferencequestionnaire.Attheendofthesession,theparticipantshadto choosebetweenabananaandabreadrollwithsteamedmeatthatwereplacedonaplatein frontofthem.Theparticipantswereaskedabouttheirthoughtsonthestudyobjectiveatthe endofthestudyinanendevaluation. Measurements Anappetitequestionnaireandfoodpreferencequestionnairewerefilledoutduringthe experiment. Theappetitequestionnairemeasuredhungeranddesire-to-eatovertimeon100mmcom- puterizedvisualanaloguescales(VAS,notatall–very)[25].Besides‘general’appetite,the appetitefor15individualproductswasmeasuredon100mmVASinarandomisedorder(for example,‘Howlargeisyourappetiteforabananaatthismoment?;notatall–very)[25].These productsweredividedintobanana,meat,sweet,savouryandstapleproducts.Bananaand bananapiewereselectedasbananaproductsandbreadrollwithsteamedmeatandbeefsoup asmeatproducts.Sweetproductsweremango,sweetpastry‘tompouce’,strawberryyoghurt andM&M’s.Savouryproductswerebreadrollwithegg,tomatosoup,cheeseandsaltedpea- nuts.Stapleproductswerebreadbun,croissantandpancake.Inaddition,odourintensity(100 mmVAS,notatall–very)andfeelingwell(100mmVAS,notatall–very)wereaddedtothe appetitequestionnairetocheckiftheodoursbecameoverwhelmingoraffectedthepartici- pants’well-being.Finally,odourpleasantnesswasmonitoredovertime(100mmVAS,notat all–very). Thefoodpreferencequestionnaire(FPQ)wasacomputerizedtaskaspreviouslyusedin Ramaekersetal.[17]andbasedonworkofFinlaysonetal.[26,27](E-prime,v2.0;Psychology softwaretools,Sharpsburg,PA,USA)measuringfoodpreferenceattheendofthe10minutes ofexposuretotheodours.Oneachtrial,twofoodsweresimultaneouslyshownonacomputer screenusingdigitalcolourphotographs.Theparticipantswereaskedtochoosethefoodthat theywouldliketoeatmostatthatmoment.Thefoodsonthephotographswerethesameas thefoodsintheappetitequestionnaire,plusadditionallybananasweets,andlittlesnacksau- sages.Allbananaproductswerecomparedagainstallnon-bananaproducts,includingmeat products.Inaddition,allmeatproductswerecomparedagainstallnon-meatproducts,leading to84comparisons.Thefrequencyofeachchosenproductwasdetermined. Dataanalysis Allvariableswereanalysedwithmixedlinearmodelstoaccountforthecorrelationsbetween therepeatedmeasures(SASversion9.1.3;SASInstituteInc.,Cary,NC),wherethefixedpartof themodelcapturesthetreatmentstructureandtherandompartthecorrelationsamongobser- vations[28].AlldegreesoffreedomwerecalculatedaccordingtothemethodbyKenwardand Roger[29].P-values<0.05(two-sided)wereconsideredsignificant. Allfiguresshowmeanvalues±SDoftherawdata.Resultsinthetextareestimated means±SE,usingamixedmodel.Resultsinthetextontransformeddatawereback-trans- formedtotheoriginalscaletofacilitateinterpretation. PLOSONE|DOI:10.1371/journal.pone.0146652 January11,2016 6/16 FoodPreferenceandAppetiteafterSwitchingOdours Appetitequestionnaire. Participantsreceivedfourcombinationsofodours,butthedata fromtheappetitequestionnairewerenotcomparedbetweencombinations.Instead,westudied theeffectsofpreviousexposure(no-smelling,no-switch,switch),currentodour(no-odour, banana,meat),exposuretime(exposuretimetothecurrentodour,either1or5min;Fig2) andtheirinteractionsonthefollowingvariables:generalappetite,odourpleasantness,odour intensity,feelingwellandchangeinappetiteforspecificproducts. Wedefinedafactor‘exposuregroup’asacombinationofpreviousexposure,currentodour andtime(1,5,7,11min),becausenotallpossible3x3x4combinationsexisted.‘Exposure group’representsthe14actualcombinations(Fig2).Wecomparedgroupmeansof‘exposure group’byusingcontraststoanalysetheeffectsofpreviousexposure(exposuregroups3+4vs7 +8+9+10vs11+12+13+14),currentodour(1+2vs3+4+9+10+13+14vs5+6+7+8+11+12), exposuretimetothecurrentodour(1vs5minand7vs11min;exposuregroups1+3+5+7+9 +11+13vs2+4+6+8+10+12+14),productandtheirinteractions.Toinvestigatewhetherprevi- ousexposurehadaneffect,B,nBandmBwerecomparedagainsteachother(exposuregroups 5+6vs7+8vs11+12)andM,nMandbMwerecomparedagainsteachother(exposuregroups 3+4vs9+10vs13+14;seeFig2forthecodes).Incasetheeffectofpreviousexposurewasnot significant,thedatafromB,nBandmBweregroupedasbananaodour(exposuregroups5+6 +7+8+11+12),M,nMandbMasmeatodour(exposuregroups3+4+9+10+13+14)andNas no-odour(exposuregroups1+2).Whentherewasasignificanteffectofpreviousexposure, onlytheratingsattime1and5minwereusedforanalysisoftheeffectofcurrentodour(expo- suregroups1+2vs3+4vs5+6). Generalappetite,odourpleasantness,odourintensityandfeelingwell:Beforestatistical analysis,generalappetitewascalculatedastheaverageofthehungerandthedesire-to-eat scores.Subsequently,ratingsforgeneralappetite,odourpleasantness,odourintensityandfeel- ingwellwerelogittransformed,usingln((y/100+0.01)/(1-y/100+0.01))tostabilizethevari- ance.Thefixedpartofthemixedmodelconsistedofthefactor‘exposuregroup’.Furthermore, generalappetiteattime=0wasusedascovariateforgeneralappetite.Therandompartofthe mixedmodelsconsistedofrandomeffectsforsessionsandparticipants.Forgeneralappetite, weusedanautoregressiveorder-1correlationmatrixforthecorrelationsamongrepeatedmea- surementsat1,5,7and11min.Forvariablesodourpleasantness,odourintensityandfeeling wellacompoundsymmetrycorrelationmatrixwasused.Correlationsandresidualvariances wereallowedtodifferbetweencurrentodoursforodourpleasantnessandodourintensity. Changeinappetiteforspecificproducts:Beforestatisticalanalysis,themeanchangein appetiteforbananaproductswascalculatedbyaveragingthechangescores(currentratings minusratingsattime=0)oftheratingsoftheappetiteforbananaandbananapie.Similarly, theratingsoftheappetiteformeat,sweet,savouryandstapleproductswereaveragedsepa- ratelyovertherespectiveproducts.Thefixedpartofthemixedmodelconsistedofthefactor ‘exposuregroup’(Fig2),anextrafactor‘product’(banana,meat,sweet,savoury,staple)and theinteractionof‘product’with‘exposuregroup’.Therandompartofthemixedmodelcon- sistedofrandomeffectsforsessionsandparticipantsandanautoregressiveorder-1correlation matrixforthecorrelationsamongrepeatedmeasurementsat1,5,7and11min.Additionally, anunstructuredcovariancematrixwasaddedtoallowforunequal(co)variancesamongprod- uctgroupsperpersonandtime.Correlationsandresidualvarianceswereallowedtodiffer betweencurrentodours. Foodpreferencequestionnaire(FPQ). Theeffectofcombination(no-odour/banana,no- odour/meat,meat/bananaandbanana/meat)onfoodpreferencewasinvestigatedusingthe FPQ.Thefoodpreferencedataweretransformedusingarcsine(sqrt(frequency/max)),with ‘max’representingthemaximumnumberoftimesaproductcouldbechoseninaset,tostabi- lizevariances.AllcomparisonsbetweenproductsintheFPQweresplitintosevensets,with PLOSONE|DOI:10.1371/journal.pone.0146652 January11,2016 7/16 FoodPreferenceandAppetiteafterSwitchingOdours eachsetcontainingcomparisonsoftwotypesofproducts.Forexample,the‘banana-staple’set containsthecomparisons:bananavsbreadbun,bananavscroissant,bananavspancake, bananapievsbreadbun,bananapievscroissant,bananapievspancake,bananasweetsvs breadbun,bananasweetsvscroissant,bananasweetsvspancake.Fixedeffectfactorswere combination,setandtheirinteraction.Anunstructuredcovariancematrixspecifiedthe(co) variancesbetweensets.Perset,wewereinterestedincomparisonsbetweencombinations. Results Appetiteforbanana,meat,sweet,savouryandstapleproducts Fig3showstheaveragechangesintheappetitesforbanana,meat,sweet,savouryandstaple products,furtheronnamedastheappetiteforspecificproducts.Previousexposureandits interactionswithexposuretime,currentodourorproductdidnotsignificantlyaffecttheappe- titeforspecificproducts(allP>0.05;Table1).Therefore,allratingsweregroupedpercurrent odour(banana,meatandno-odour).Theinteractionbetweencurrentodourandproductwas significant(P<0.001):thechangesintheappetitesforbanana,meat,sweet,savouryandstaple productsdidnotdifferamongeachotherintheno-odourcombination(P=0.13),butdiffered duringexposuretobanana(P<0.001)andmeat(P<0.001)odour.Exposuretobananaodour increasedtheappetiteforbananaproducts(P<0.001),decreasedtheappetiteformeat (P=0.026)andsavoury(P=0.028)productsandhadnosignificanteffectonsweet(P=0.10) andstaple(P=0.48)products,comparedwithno-odour.Exposuretomeatodourincreased theappetiteformeat(P<0.001)products,decreasedtheappetiteforbanana(P<0.001)and sweet(P<0.001)productsandhadnoeffectonsavoury(P=0.46)andstapleproducts (P=0.72),comparedwithno-odour. Preferenceforbanana,meat,sweet,savouryandstapleproducts, measuredwithFPQ Thepreferenceof84pairsoffoodpictureswasassessed,comparingbananaandmeatproducts againsteachotherandagainstsavoury,sweetandstapleproducts(Table2).Whensetsof bananaandmeatproductswereoffered,thebananaproductswerechosenmoreoftenwhen thelastsmelledodourwasbanana,thanwhenthelastsmelledodourwasmeat(allP<0.05).In thebanana-savouryandbanana-staplesets,thebananaproductsweremoreoftenchosenafter theno-odour/bananacombination,thanaftercombinationsno-odour/meatandbanana/meat (allP<0.05).Inthemeat-sweetset,themeatproductswerechosenlessoftenaftercombination no-odour/banana,thanaftercombinationsno-odour/meatandbanana/meat(allP<0.05).The assessmentofthe84pairstook3to4minpersession. Actualfoodchoice Aftertheodourexposure,thirteenparticipantsalwayschoseabananaandnineparticipants alwayschoseabreadbunwithsteamedmeat.Eightoutof30participantsswitchedtheirfood choicebetweensessions.Thechoiceforbananavsbreadrollsteamedmeatwasasfollows:19 vs11inno-odour/banana,19vs11inmeat/banana,12vs17inno-odour/meatand17vs13in banana/meat. Generalappetite Generalappetiteratingswereonaverage79mm.Therewerenosignificantdifferencesingen- eralappetitebetweenthefourcombinationsattime=0min(F =1.8;P=0.16).General 3,83.4 appetiteratingsattimes1,5,7and11werestronglyrelatedtotheappetiteratingsattime=0 PLOSONE|DOI:10.1371/journal.pone.0146652 January11,2016 8/16 FoodPreferenceandAppetiteafterSwitchingOdours Fig3.Meanchangeinappetiteforbanana,meat,staple,sweetandsavouryproductsovertime. Scoresduringexposureto(A)no-odour/banana,(B)no-odour/meat,(C)meat/bananaand(D)banana/meat, measuredon100mmVAS.ThenumbersbetweenthebracketsrepresenttheSDatrespectively1,5,7and 11minutes.Thedashedlinerepresentstheswitchbetweenodours. doi:10.1371/journal.pone.0146652.g003 PLOSONE|DOI:10.1371/journal.pone.0146652 January11,2016 9/16 FoodPreferenceandAppetiteafterSwitchingOdours Table1. F-valueswithdegreesoffreedomandP-valuesofallfactorswithfixedeffectsforthechange inappetiteforspecificproducts. Changeinappetiteforproducts Fixedeffects Exposuretime F =0.6 P=0.55 2,336 Previousexposure F =0.8 P=0.54 4,281 Currentodour F =4.0 P=0.019 2,303 Product F =9.6 P<0.001 4,254 ExpTime*Previous F =1.2 P=0.32 4,298 ExpTime*Current F =2.6 P=0.073 2,279 ExpTime*Product F =1.1 P=0.38 8,473 Previous*Current F =0.7 P=0.50 2,258 Previous*Product F =0.80 P=0.71 16,502 Current*Product F =16 P<0.001 8,445 All3-or4-wayinteractions P>0.20 doi:10.1371/journal.pone.0146652.t001 (F =250;P<0.001;Table3).Generalappetitedidnotchangesignificantlyovertime 1,124 (P=0.52).Theeffectofpreviousexposureongeneralappetitewasborderlinesignificant (P=0.08)andtherefore,theeffectofodourwasinvestigatedonlyattimes1and5min.There werenosignificantdifferencesbetweenmeat,bananaandno-odourattimes1and5min(Mvs BvsN;P=0.25).Theinteractionbetweenpreviousexposureandcurrentodourwasnotsignif- icant(F =1.2;P=0.31). 2,167 Odourpleasantness Fig4showsthepleasantnessofthecurrentlysmelledodour.Therewasasignificantinteraction betweencurrentodourandexposuretime(F =3.5;P=0.033;Table3).Ratingsdecreased 2,152 onaverageby4mmfrom1minto5minexposuretimeduringexposuretomeatandbanana odour(P=0.005).Noeffectofexposuretimewasfoundduringexposuretono-odour (P=0.61).Previousexposurehadnosignificanteffectonratedodourpleasantness(F = 4,68.4 0.3;P=0.88).Thepleasantnessratingsofmeatandbananaodourswerehigherthanofno- Table2. Meanpercentageoftimesaproductwaschosenpersetofproducts,afterexposuretodifferentcombinations,measuredwiththefood preferencequestionnaire. Set Product1 Banana- Banana- Banana- Banana- Meat- Meat- Meat- Product2 Meat* Savoury* Sweet* Staple* Savoury# Sweet# Staple# Nrofcomparisonsperset 18 12 12 9 12 12 9 No-odour/banana 62a 60a 43 48a 46 33a 33 Meat/banana 58a 50ab 42 40ab 47 41ab 33 No-odour/meat 42b 47b 38 33b 53 56b 44 Banana/meat 42b 44b 43 33b 54 54b 46 SuperscriptwithdifferentlettersdenotesignificantdifferencesatP<0.05 *Meanpercentageoftimesthatabananaproductwaschosen #Meanpercentageoftimesthatameatproductwaschosen doi:10.1371/journal.pone.0146652.t002 PLOSONE|DOI:10.1371/journal.pone.0146652 January11,2016 10/16