Estradiol levels modulate brain activity and negative responses to psychosocial stress across the menstrual cycle

Estradiol levels modulate brain activity and negative responses to psychosocial stress across the menstrual cycle

Kimberly Albert

, Jens Pruessner

, Paul Newhouse

aCenterforCognitiveMedicine,VanderbiltUniversity,Nashville,TN,UnitedStates

bGeriatricResearch,Education,andClinicalCenter,TennesseeValleyVAHealthSystem,TennesseeValley, TN,UnitedStates

cUniversityofVermont,Burlington,VT05405,UnitedStates

dMcGillUniversity,Montreal,QC,Canada

Summary Althoughovarianhormonesarethoughttohaveapotentialroleinthewell-known sexdifferenceinmoodandanxietydisorders,themechanismsthroughwhichovarianhormone changescontributetostressregulationarenotwellunderstood.Onemechanismbywhichovar- ianhormonesmightimpactmoodregulationisbymediatingtheeffectofpsychosocialstress, whichoftenprecedesdepressiveepisodesandmayhavemoodconsequencesthatareparticu- larlyrelevantinwomen.Inthecurrentstudy,brainactivityandmoodresponsetopsychosocial stress wasexaminedinhealthy,normallycyclingwomenateither thehighorlowestradiol phase ofthemenstrual cycle. Twenty eightwomen were exposed tothe Montreal Imaging StressTask(MIST),withbrainactivitydeterminedthroughfunctionalmagneticresonanceimag- ing,andbehavioralresponseassessedwithsubjectivemoodandstressmeasures.Brainactivity responsestopsychosocialstressdifferedbetweenwomeninthelowversushighestrogenphase ofthemenstrualcycle:womenwithhighestradiollevelsshowed significantlylessdeactiva- tioninlimbicregionsduringpsychosocialstresscomparedtowomenwithlowestradiollevels.

Additionally,womenwithhigherestradiollevelsalsohadlesssubjectivedistressinresponse totheMISTthanwomenwithlowerestradiollevels.Theresultsofthisstudysuggestthat,in normallycyclingpremenopausalwomen, highestradiollevelsattenuatethebrainactivation changesandnegativemoodresponsetopsychosocialstress.Normalovarianhormonefluctua- tionsmayaltertheimpactofpsychosociallystressfuleventsbypresentingperiodsofincreased

∗Correspondingauthorat:VanderbiltCenterforCognitiveMedicine,DepartmentofPsychiatry,VanderbiltUniversitySchoolofMedicine, 160123rdAve.,SouthNashville,TN37212,UnitedStates.Tel.:+16153277030;fax:+16158750686.

E-mailaddress:Paul.Newhouse@Vanderbilt.edu(P.Newhouse).

Konstanzer Online-Publikations-System (KOPS) URL: http://nbn-resolving.de/urn:nbn:de:bsz:352-2-qk8ebb21r8mj4

https://dx.doi.org/10.1016/j.psyneuen.2015.04.022

1. Introduction

The sex difference in affectiveand stress-related disease rates (2 to 3 times greater incidence and prevalence of majordepressioninwomencomparedtomen(Brometetal., 2011;Kessleretal.,2005))emergesatpubertyandremains until menopause (Kessler et al., 1994). The stress expo- sure model of depression suggests that MDD is the result of a vulnerability to depression, combined with the trig- ger of stressful life events (Hankin et al., 2007; Liu and Alloy,2010).Accordingly,psychosocialstressorsareamong thetopreportedantecedentstodepressionepisodes(Frank etal.,1994;Kendleretal.,1999;Kendleretal.,2000),and dysregulationofthehypothalamic—pituitary—adrenal(HPA) axis response is a consistent finding in major depression (Burkeetal.,2005;Lupienetal.,2009).Psychosocialstress maybeespeciallyimportantintheetiologyofmooddisor- dersinwomenasthedepressogeniceffectsoflifestressors arereportedlygreaterinwomenthanmen(Mezulisetal., 2010),evenwhenthereisnodifference inthenumberof stressfullifeeventsorinthesubjectiveperceptionofthese events (Korszun, 2009).Although thisfinding isconsistent andhasbeenreplicatedbystudiesacrosstheglobe(Bromet etal.,2011),thereasonsforthesexdifferenceinincidence andprevalenceofmoodandanxietydisordersarenotwell understood.Thecortisolresponsetostressshowsconsistent sexdifferences(Vamvakopoulos,1995),andchangesacross the menstrual cycle, decreasing during the latefollicular phasewhencirculatingestrogenishigh(Kirschbaumetal., 1999).Womenaremorelikelythanmentoshowvariations inHPAfunctioninresponsetostressors(Weissetal.,1999) andduringdepressive episodes(Korszun,2009).Thesedif- ferencesinstresssystemresponselikelycontributetomood disorderriskinwomen(Weissetal.,1999),andmaybemod- ulatedbyovarianhormonefluctuationsacrossthemenstrual cycle(Kajantie,2006;Rocaetal.,2005).Theroleofcorti- costeroidsinstressresponseandregulationiswellknown, however theeffectsof gonadalsteroids(e.g. estradiol)— which maybe specificallyimportant in women —areless wellcharacterized.

The concurrence of increased major depression disor- der(MDD)andpost-traumaticstressdisorder(PTSD)riskin womenduringthereproductiveperiodoflifesuggeststhat thecyclicfluctuationofovarianhormonesduringthisperiod maycontributetotheriskforpsychopathology.Understand- ingtheroleofovarianhormonesinemotionalprocessingand mood regulation in women mayprovide important insight into the mechanisms underlying the stress response and potentially the increased incidence of MDD and PTSD in women.

Ovarian hormones may modulate the effects of stress onmood.Brain areasthat arecentraltomoodregulation (includingtheamygdalaandhippocampus)showsomeofthe largestdensities ofestrogenreceptorsinthehuman brain

(Merchenthaleretal.,2004;Ostlundetal.,2003);interest- ingly, thoseareas are alsoveryrich in cortisol receptors.

Estrogen may modulate the activity of these areas; large communityandclinicbasedstudiesindicatethatnegative moodcomplaints(Davydovetal.,2005;Gondaetal.,2008) (eveninhealthywomen)andsuicidalbehavior(Baca-Garcia etal.,2004;SaundersandHawton,2006)increaseinwomen during low estrogenphases of the menstrual cycle. Brain activity relatedtoprocessingnegativeemotionalinforma- tionisalsomodulatedbychangingestradiollevelsacrossthe menstrualcycle(Goldsteinetal.,2005;Merzetal.,2012), suggesting that estrogenmay alter the mood responseto negativeinformation,makingthisinformationmoreorless salienttocognitiveprocessesandsubsequentlymoodstates.

Although there is strong evidence that estrogen mod- ulates the response to negatively valenced stimuli, such as negativeimages(Andreano andCahill, 2010; Goldstein et al., 2005), the effect of estrogen on the response to psychosocial stressis lesswellunderstood. Animalmodels suggest that female rodents do not gain the same ben- eficial effect of acute stress on hippocampally-mediated or prefrontal tasks as maleanimals, and that high estro- gen levelsenhance thenegativeeffects ofstress inthese areas (Shansky et al., 2004; Shors and Leuner, 2003). In contrast studies in ovariectomized female rats indicate thatestrogenreplacementincreasesresiliencetostressin earnedhelplessnessmodelsandsupporthippocampalplas- ticity(BredemannandMcMahon,2014;Smithetal.,2010).

Themodulationofstresseffectsonbrainactivityandfunc- tionremainunclearandlikelydependonthetypeofstress and measure of function (Shansky, 2009). The effects of estradiolonthecognitiveconsequencesorbrainactivityin womenhavenotbeenwidelyinvestigated.Previousstudies have generallycomparedstress responseinwomenduring the early follicular phaseand mid lutealphase.However, this approach precludes examining the effects of estra- diolseparatelyfromprogesterone(Kirschbaumetal.,1992;

Kirschbaumetal.,1999).Additionally,fewofthesestudies have included subjective mood measures, and the mea- surement ofstressresponsesolelythroughfreecirculating (salivary) cortisol is complicated by menstrual effects on cortisol bindingglobulin andadrenocorticotropichormone sensitivity.Psychosocialstressdiffersfromprocessingemo- tional information in a number of ways — stress includes elements of self-esteem, uncontrollability, and personal threat.Further,studiesusingperformance—basedstress- ors provideevidence thatthereis asexdifference in the endocrine (cortisol) response to this type of stress, and thatcyclingovarianhormonesmaymodulatethisresponse in women (Kirschbaum et al., 1992; Kirschbaum et al., 1999).Social-evaluativethreatisoneelementofpsychoso- cialstressthatmaybeespeciallysalientforwomen,andhas face validityfor the reallife stressors womenexperience and thatcontributetomood disorderrisk (Kendleretal.,

1999,2001;Stroudetal.,2002).Understandingtheeffectof ovarianhormonesontheresponsetothesetypesofstressors isthusimportantinunderstandingtheroleofovarianhor- monesinthebrainmechanismsofemotionalprocessingand mood in women. Investigating the stress response, under controlled laboratory conditions, provides an opportunity toexaminetheeffects ofovarian hormones onemotional processing,stressresponse,andsubsequentmood.

In this study,we examined the effectof estradiollev- els across the menstrual cycle onboth the brain activity and behavioral response to a laboratory-based model of psychosocialstress. Wehypothesized that highcirculating estrogenwouldreduce theeffectofacutestress onbrain activityandbeassociatedwithlesssubjectivedistressfol- lowingthestress-inducingtask.

2. Methods

This cross sectional study included 28 normally cycling women(ages18—45)whowereexaminedduringearlyfollic- ularphase(day1—2ofmenstrualcycle),oratovulation(day 12—14),andwhoparticipatedinapsychologicalstressfMRI task.Duetothepsychosocialstresstaskusedinthisstudy, andtherequirementtodebriefparticipantsaboutthefocus ofthestudyoncethestresstaskwascompleted,wewere unabletorepeatthestresstaskforawithinsubjectscom- parison,andwerelimitedtobetweensubjects/groupcom- parisons.Approximatelyhalfoftheparticipantswerestud- iedattheUniversityofVermontandhalfwerecompletedat VanderbiltUniversity.The studycoordinatorwasthesame throughoutthestudyandtheMRIscannersweretheiden- ticalmake and model, with the samesoftwareversion in operation at thetime of thestudy.The study procedures wereidenticalatbothUniversities,andimagingandbehav- iordatawereanalyzedforstudylocationeffects.Thestudy wasreviewedandapprovedbyboththeUniversity ofVer- montandVanderbiltUniversityInstitutionalReviewBoards, andallparticipantsprovidedwritteninformedconsent.

2.2. Participants

Participants were recruited and told that they would be participating in a study to examine the effects of men- strual cycle hormone changes on mathematical ability in women. All participants were healthy, with regular men- strual cycles (21—35days), nohormonal contraceptive or centrally active medication use, and nohistory of severe painormoodchangesrelatedtotheirmenstrualcycles.We excludedparticipantswithcurrentorpastAxisIpsychiatric disordersusingtheStructuredClinicalInterviewforDSMDis- orders(Firstetal.,2001).BeckDepressionIndexscorewas required tobe less than 7, and Beck Anxiety Indexscore lessthan15(Becketal.,1961).Currentillicitdruguseand pregnancywereexclusions.

2.3. Screeningandcharacterization

To assureallparticipants wereof atleast average intelli- gence,weadministeredtheWechslerAbbreviatedScaleof

Intelligence(Wechsler,1999).The CompositeInternational Diagnostic Interview for premenstrual dysphoric disorder (Kessler and Ustun, 2004) was used to collect menstrual cycle history and torule out premenstrual dysphoric dis- order,andphysicalandpsychologicalsymptomsassociated with the menstrual cycle were assessed using the Moos MenstrualDistress Questionnaire(Moos, 1968).Personality factorswereassessedatscreeningusingtheNEOFiveFactor Inventory (Costa and McCrae,2000). Following screening, participantstrackedtheirmenstrualcyclesforthreemonths todetermine anaveragecyclelengthandtocalculatethe studydaydate. AttheUniversityofVermont,participants tracked using a paper calendar; at Vanderbilt University, womenfilledindailysurveysusingtheon-lineREDCapsys- tem.

Womenwererandomlyassignedtoreturnforthestudy dayduringtheearlyfollicularphase(days1—2ofthemen- strualcycle)oratovulation(day12—14);beforethestudy began,participantresearchIDnumberswereassignedaran- domlygeneratednumber,rankedaccordingtothisrandom numberandthenalternatinglyassignedtothetwogroups.

The study day took approximately 3h tocomplete. Upon arrival at the Clinical Research Center,participants com- pletedtheStateandTraitAnxietyInventories(Spielberger, 2010)andtheBeckDepressionandAnxietyInventories(Beck etal.,1961).

2.4. Stresstask

We employed theMontreal Imaging Stress Task (MIST) for psychosocialstress induction(Pruessneretal.,2008).The MISTproducesmoderatepsychosocialstressthroughacom- bination of motivated performance and social-evaluative threat.WepresentedtheMISTasanarithmetictaskinthe MRI scanner, and instructed participants that they should achieve an 80—90% correctperformance for theirdata to beusablein thecontextofthisexperiment.Unbeknownst tothesubjects,theMISTcontains analgorithmproducing scriptthatautomaticallyadjuststhedifficultyofthemath tasks to the aptitude of the participant, this way main- taining a low performance rate (between 40 and 50%)by changingeithertheproblemdifficultyortheallottedtime.

A‘‘control’’condition,inwhichtheparticipantssolvearith- meticproblemswithnotimelimitandnoperformancelimit, serves asa contrast to control for brain activity changes induced byarithmetictaskdemands (visualstimuli,motor response,andmentalarithmetic).Socialevaluativethreat comesfromscriptedexperimenterinteraction;theexperi- mentersentertheMRIroombetweenruns andinformthe participantsthat theyarenot doingwell enoughand that theyneedtoimprovetheirperformancefortheexperiment tobesuccessful.

Inthis study,participants completedthree runs of the MIST;theMRIsessionswerescheduledduringtheafternoon when baseline cortisol levelsare low.After the first run, experimenter1enteredthescannerroomandprovidedthe scriptedfeedbackandaskedtheparticipanttocompletea second run of the task.After the second run the experi- menter1enteredthescannerroomandtoldtheparticipant that the experimenter 2 (‘‘doctor’’) would like tospeak with them about their performance, at which point the

experimenter2enteredthescannerroomandprovidedthe scriptedfeedbackandaskedtheparticipanttocompletea thirdrunofthetask.Thisprotocolwasdesignedtomaintain both performanceandsocial evaluativethreatthroughout the MIST and toprevent participants fromhabituating to theperformancechallengeorgivinguponcompletingthe task. Specifically the interaction of the participants and theexperimenterswasstructuredtogeneratepsychosocial stress.Tohabituatetheparticipantstothescanningenviron- mentanddecreasescanner-relatedstressatthestudyday, participantsenteredanMRIsimulatorduringthescreening visit; in the MRI simulator,participants watched a nature video while beingexposed tothe MRIenvironment (simu- lated MRI sounds, head coil, andbeing placed in the MRI bore)andthestimuluspresentationsystem.

TheMISTwaspresentedinablockdesign;eachcondition waspresentedtwiceper3runs(‘‘stress’’:100s,‘‘control’’:

50s,‘rest’’:30s).ParticipantspracticedtheMISTtask(con- troltrialsonly)intheMRIsimulator,beforetheMRIsession.

2.5. Subjectivemeasures

Participantscompletedsubjectivemeasures:theStressand Arousal Checklist(King etal.,1983)(before andafterthe MIST),theProfileofMoodStates(McNairandDroppleman, 1971),andaVisualAnalogueScaleformoodandtaskexpe- rience(bothaftertheMIST).

2.6. Endocrinemeasurements

Salivasampleswerecollected,usingSalimetricssalivettes placedunderthe tongue,at seventimes duringthestudy daytoassessthecortisolresponsetothestresstask(Supp.

Fig.1).Salivettesremainedunderthetonguefor2min,and then werestoredat−80^◦C freezeruntilanalysis.Cortisol levelswereassessed usingtheSalimetrics HighSensitivity Salivary Cortisol Enzyme Immunoassay Kit for quantita- tive determination of salivary cortisol with a sensitivity of 0.007␮g/dL for 25␮Lsample volume(Salimetrics LLC, Philadelphia,PA).Allcortisolmeasurementswereassayedin duplicate,withmeanintra-assaycoefficientofvariationof 7.5%andinter-assaycoefficientofvariationof14.3%.Corti- solareaunderthecurvemeasurementsforeachparticipant (overthestudyday)werecalculatedusingGraphpadPrism.

Bloodsampleswere collectedfromeachparticipantat the endof thestudy daytodetermine theestradiollevel andtoverifymenstrualphase.Estradiolwasassessedfrom sera using the Immulite 1000 endocrine panel immunoas- saysystemwithmeanintra-assaycoefficientofvariationof 5.9%. For analyses by menstrual phase, we required that both theparticipant’s menstrualcycle trackingandestra- diollevelindicatedthattheywereintheassignedmenstrual phase.Participants’whoseestradiollevelsdidnotconcord withtheirrandomlyassignedmenstrualphaseatthestudy day; women whose estradiol levels were higher or lower thanthereferencevaluesfortheirassignedphase,werenot includedintheloworhighestradiolgroups,astheycould notbeaccuratelyassignedtoeithergroup.Thesedatawere includedinallotheranalyses.

ProgesteronelevelswereassessedusingtheSalimetrics SalivaryProgesterone(P4)EnzymeImmunoassayKitwith a

sensitivityof5pg/mLfor25␮Lsamplevolume(Salimetrics LLC, Philadelphia, PA). The firstsaliva sample of the day wasusedforprogesteronemeasurements.Allprogesterone measurementswereassayedinduplicate,withmeanintra- assaycoefficientofvariationof<1%.

2.7. MRIscanparameters

Participants werescannedona Philips3.0TAchievascan- ner,with an8channelhead coil.Allparticipantsreceived thefollowing MRsequences:(1)AsagittalT1-weighted3D TurboFieldEchoSensitivityEncoding(TFESENSE)sequence perpendicular to the anteriorcommissure (AC)—–posterior commissure(PC)line,repetitiontime(TR)of9.9ms,echo time(TE)of4.6ms,aflipangle of8^◦,numbersignalaver- ages(NSA)1.0,afieldofview(FOV)of256mm,a256×256 matrix,and1.0mmslicethicknesswithnogapfor140con- tiguous slices. (2) A T2-weighted Gradient and Spin Echo (GRASE) sequence wasrunparalleltothe AC—PC line,TR 2470ms,TE80ms,NSA3.0, FOVof 230mm, 0.7mm slice thicknesswith5.0mmgapfor28slices.(3)ThreeEchoplanar Blood Oxygenation Level Dependent (EpiBOLD) functional sequences, transverse orientation, TR 2500ms, TE 35ms, flip angle 90^◦, 1 NSA for FOV 240, 240×128 matrix, and 4.0mm slice thickness with nogap, with ascending inter- leavedacquisition,for35contiguousslices.

2.8. fMRIanalysis

fMRI datawasprocessedusingStatistical ParametricMap- ping(SPM8)(WellcomeDepartmentofCognitiveNeurology, 2008). Preprocessing included: realignment of the three functionalrunsandcorrectionforbulk-headmotion,coreg- istration of functional and anatomical images for each participant,segmentationoftheanatomicalimage,andnor- malizationoftheanatomicalandfunctionalimagestothe standardMNItemplate,andspatialsmoothingwithaGaus- sianfilter(8mmatfullwidth,halfmaximum).

Artifact detectionwasperformed onfunctionalimages usingtheARTtoolboxinSPM,andoutliersforsignalintensity (z>3) and motion (movement>3mm in either translation or rotation) were entered as nuisance regressors at the first level, single-participant analysis. At first level anal- ysis, T-maps were created from linear contrasts for task conditions and between condition comparisons; these T- maps were used in the second level whole-brain random effectsanalysisofparticipantgroupeffectswithtwo-sample t-tests. The preprocessed functional images had a voxel size of 2×2×2mm and cluster threshold correction was usedtocontrolfor multiplecomparisons (fromvoxelwise p=0.005)tocorrectedp<0.001withk=58(correctedvoxel- wise p=0.000001) (from alpha simulation in REST (Song etal.,2011)toolboxforSPM,usingagreymattermask).

Twosetsofsecond levelanalyseswereconducted,one analysiswithgroupsdefinedbymenstrualphaseandestra- diol level at the study day, and a second analysis with groups definedby subjective distresstotheMIST (change inStressArousalChecklistscore).Separateregressionanal- yseswererunforestradiollevelandsubjectivedistressby hippocampalactivity(usingabilateralhippocampusregion of interest mask created fromthe AutomatedAnatomical

Figure1 Psychosocial stresseffectfMRIstress condition—–control condition:allparticipants (n=28).Greater deactivationin limbicregionsduringthestressconditioncomparedtothecontrolcondition(correctedtop<0.001).

Labeling (AAL) in WFU Pickatlas (Maldjian et al., 2003)).

Percent signal change in right and left hippocampus was calculatedwiththeREXtoolboxforSPM; usingmaskscre- atedfromtheaveragepeakclusterinthehippocampusROI (using the contrast ‘‘control’’ condition>‘‘stress’’ condi- tioninthefullsubjectset)withclusterthresholdcorrection (fromvoxelwisep=0.05)tocorrectedp<0.001withk=101 (correctedvoxel-wisep=0.001).

3. Results

Average participant age was 30.4 years (SD=8.2 years).

There was no difference in age between subject groups either by estradiol level (t (8)=−1.01, p=0.33), or dis- tress(t(26)=−0.83,p=0.41).NoparticipantshadMenstrual Distress Questionnaire scores that indicated severe pre- menstrual somatic or affective changes. There were no differences between any of the participant groups at screening, including Beck Depression Inventory scores, Beck Anxiety Inventory scores, NEOFive Factor Inventory personality measures, or any of the Menstrual Distress Questionnairemeasures,andnone oftheparticipantshad elevated scores on these scales (Supp. Table 1). For the analysisoftheeffectofmenstrualphaseonstressrelated brainactivity,weincludedonlyparticipantswhoseestradiol levelat the studyday confirmedtheir assignedmenstrual phase.Thereferencevalueformedianearlyfollicularphase estradiol levels using the Immulite 1000 endocrine panel immunoassay system is 31pg/mL. Women were included in thelow estradiolphase groupif theywereassigned to theearlyfolliculargroupandhadanestradiollevelonthe studydaylessthan40pg/mL(mean24.91pg/mL).Women wereincludedinthehighestradiolphasegroupiftheywere assignedtotheovulatorygroupandhadanestradiollevelon thestudydaygreaterthan50pg/mL(mean103.52pg/mL).

This resulted in 8 participants (4 from the low estradiol, and4fromthehighestradiolgroup)notbeingincludedin stress-brainimaginganalysesforlowvs.highestradiolphase groups;theseparticipants’datawereincludedinallother analyses.

3.2. Brainactivity

3.2.1. Psychosocialstressresponse

Secondlevelcomparisonusingaonesamplet-testofaver- age images of all subjects (n=28) (using the 1st level

contrastofcontrolcondition>stresscondition)revealedsig- nificantly(uncorr.p=0.005,clustersize=58,corr.p<0.001) greater activation in cingulate, temporal (including hip- pocampus), and frontal regions and activation in parietal and precentral gyrus regions during the control condition comparedtothestresscondition(Fig.1andSupp.Table2).

Thisactivitypatternisconsistentwiththeactivitypattern previouslyshownbystudiesusingtheMIST(Pruessneretal., 2008);reducedactivationofhippocampusandpara-limbic regions during the stress condition. These results confirm thattheoverallbrainactivityresponsetopsychosocialstress in a sample including only women is similar tothe brain activityseeninstudiesthatincludebothmenandwomen.

Menstrual cyclephase/estradioleffects.Estradiollevel was significantly higherin the high estradiol phase group (mean=103.52pg/mL, SD=34.09) than the low estradiol phasegroup(mean=24.91pg/mL, SD=9.17)(t(18)=7.04, p<0.001). All women had salivary progesterone levels under105pg/mL (mean=49.72pg/mL, SD=35.54), consis- tent with the follicular phase (Chatterton et al., 2005).

Progesteronelevelsdid notdifferbetween the lowestra- diolphasegroup(mean=48.90pg/mL,SD=36.10)andhigh estradiol phase group (mean=50.54pg/mL, SD=36.91) (t (18)=0.099,p=0.92)(Supp.Fig.2).

ThefMRIestradiolphasegroupanalysisfortheeffectof psychosocialstress,includingparticipantswhoshowedcon- cordanceof assignedmenstrualphaseandestradiollevel, (high estradiol n=10, low estradiol n=10), showed a sig- nificant(uncorr.p=0.05,cluster size=529,corr. p<0.005) effectofestradiollevelonbrainactivityduringpsychosocial stress and perceived distressfollowing theMIST, however thesefindingsdidnotsurvivemoreconservativecorrection formultiplecomparisons.Womenwith lowestradiollevels hadsignificantlylessactivityinthelefthippocampusthan womenwithhighestradiollevelsduringthestresscondition (Fig.2andSupp.Table3).Thisfindingindicatesthatwomen in thehigh estradiol phaseof themenstrual cycle have a greaterlefthippocampalactivityduringpsychosocialstress thanwomeninthelowestradiolphase.

Bilateralhippocampalactivityduringpsychosocialstress was directly associated with estradiol levels in an analy- sis of the 20 participants included in the estradiol phase group analysis (uncorr. p=0.005, cluster size=58, corr.

p<0.001)withgreateractivityinthebilateralhippocampal ROIbeingassociatedwith higherestradiollevels.Examin- ing the individual right or left hippocampal ROIs showed thatthecorrelationbetweenestradiollevelandpercentsig- nalchangeineitherrightorlefthippocampusdidnotmeet statisticalsignificance(righthippocampus:r=−0.14,n=28,

Figure2 EstradioleffectsfMRIstresscondition—–controlcondition:highestradiol(n=10)—–lowestradiol(n=10).Greateractivity inlefthippocampusduringthestressconditioninthehighestradiolgroupthanthelowestradiolgroup(cluster-levelcorrectedto p<0.005).

p=0.45;lefthippocampus:r=−0.21,n=28,p=0.37),indi- catingthattheeffectofestradiolwasnotlateralized.

In an analysis that included all participants grouped accordingtotheirassignedmenstrualphase(withoutcon- firmationbyestradiollevel),therewerenobrainareasthat showed greater activity in the periovulatory group com- paredtotheearlyfollicularphasegroupateithercorrected p<0.01orcorrectedp<0.05.

3.2.2. Subjectivedistresseffects

To compute a group contrast for the effects of subjec- tive distressonbrainactivation changes,we performed a mediansplitofsubjectivedistressscores;differenceinpre vs.postMISTstressscoreontheStressandArousalCheck- list(highdistressn=14,lowdistressn=14).Thesubsequent fMRIanalysiswiththehighandlowdistressgroupsshowed anassociatedofdistressandbrainactivityduringthepsy- chosocial stress task. Women with higher distress scores showedsignificantlylessactivityinbilateralhippocampus, midbrain,leftparietal,andleftfrontalregionsthanwomen with higher distressscores (Fig. 3 andSupp. Table 4).By contrast,greater subgenualcingulateactivity wasseenin womenwithhigherdistressscorescomparedtowomenwith lower distress scores (Fig. 3). These differences indicate thatwomenwhorespondtotheMISTwithhighsubjective distresshaveagreaterchangeinbrainactivityduringpsy- chosocial stress than women with low subjective distress followingtheMIST.

Bilateralhippocampalactivityduringpsychosocialstress wasinverselyassociatedwith distressscoresinananalysis ofallparticipants(p<0.001)withloweractivityinthehip- pocampalROIbeingassociatedwithgreaterdistress.These correlations were statistically significant bilaterally (left hippocampus r=0.51, n=28, p=0.006, right hippocampus r=0.49, n=28, p=0.008). We re-analyzedthe correlation

after step-wiseremovalofpercentsignal change extreme values and association between left hippocampalpercent signal change and distress remained significant (r=0.39, n=28,p=0.04)buttherighthippocampusdidnot(r=0.31, n=28, p=0.10). A two way factorial ANOVA for E2 group and Distress group effects onhippocampal percent signal change did notshow a significant interaction of estradiol phasegroupand distressgroupfor right (p=0.592)orleft (p=0.977)hippocampusROIs.

There was no significant difference in brain activity betweenparticipantsrunatthetwostudylocationsinanyof thecontrastsofinterestatanylevelofsignificanceincluding uncorrectedp<0.05.

3.3. Behavioralandmoodmeasures

3.3.1. Estradioleffects

There wasnoeffectofmenstrualcycle phaseor estradiol phasegrouponpre-MISTStressandArousalChecklistscores, TraitAnxietyInventoryscores,orStateAnxietyscores,and noeffectonanyoftheProfileofMoodscores,Arousalscores, oranyofthescalesoftheStressTaskVisualAnalogueScale following theMIST(Supp.Table5).Distress scores(differ- enceinSACLStressscorebeforeandaftertheMIST)were lower in women with high estradiol than in women with low estradiol levels (Table 1). A multiple regression was run topredict stress change fromestradiol level, proges- terone level, and age, F(3, 24)=4.2, p<0.05, R²=0.344.

Only estradioladded statistically significantly to the pre- diction,p<0.05.

3.3.2. Distresseffects

TherewasnosignificantdifferencebetweengroupsonStress andArousalChecklistscores,TraitAnxietyInventoryscores, or StateAnxietyscores beforecompletingtheMIST(Supp.

Figure3 SubjectivedistresseffectsfMRIstresscondition—–controlcondition:highdistress(n=14)—–lowdistress(n=14).Greater activityinbilateralhippocampusandlesssubgenualcingulateactivationduringthestressconditioninthelowdistressgroupthan thehighdistressgroup(cluster-levelcorrectedtop<0.001).

Table1Behavioralmeasuresatstudyday—–postMIST. HighE2SDLowE2SDE2Phase GroupsHighDistressSDLowDistressSDDistressGroups (n=10)(n=10)p(n=14)(n=14)p MeanMeanMeanMean PostSACL—–stress6.505.5213.905.000.015.504.4314.642.73<0.001 PostSACL—–arousal8.203.018.802.700.646.863.449.002.250.60 POMStension8.807.2815.107.460.076.573.9616.937.15<0.001 POMSdepression2.402.226.907.370.082.211.769.8612.290.03 POMSanger4.103.638.209.920.242.713.248.218.680.04 POMSfatigue5.905.045.303.500.765.364.736.074.950.70 POMSconfusion6.902.699.105.300.266.003.5910.295.770.03 POMSvigor11.705.6610.907.000.7810.436.2010.296.270.95 POMStotal—–’TMD’16.3014.0433.2029.670.1212.0013.9741.0732.950.01 STVAS—–personallyinvolved83.3017.2981.4022.760.8477.2122.9987.9310.470.12 STVAS—–stressful86.208.9987.8011.320.7383.439.8190.298.700.06 STVAS—–new65.2031.8871.0024.830.6665.7938.4371.5023.450.63 STVAS—–uncontrollable69.5010.3273.9016.110.4866.2914.7772.8624.650.40 STVAS—–unpredictable68.2028.7573.4013.970.6167.5023.1272.0717.750.56 STVAS—–negativeconsequences32.2024.8160.5037.410.0641.1426.7552.0736.300.37 Distress4.606.9511.705.380.022.933.8514.002.45<0.001 E2(pg/mL)103.5234.0924.919.17<0.00174.4940.9343.4735.260.04 E2=estradiol,STAI=stateandtraitanxietyinventory,SACL=stressandarousalchecklist,POMS=profileofmoodstates,STVAS=stresstaskvisualanaloguescaledistress=post,SACL stress—–preSACLstress.

Figure4 Estradiolandprogesteronelevelsfor groupanaly- ses.Lowestradiolgroup(n=10)estradiolmean=24.91pg/mL, SD=9.17, high estradiol group (n=10) estradiol mean=103.52pg/mL,SD=34.09(t (18)=7.04,p<0.001).Low estradiol group progesterone mean=48.90pg/mL, SD=36.10, high estradiol group progesterone mean=50.54pg/mL, SD=36.91(t(18)=0.099,p=0.92).Lowdistressgroup(n=14) estradiol mean=74.49pg/mL, SD=40.93, high distress group (n=14)mean=43.47pg/mL,SD=35.26,(t(26)=2.15,p<0.05).

Low distress progesterone mean=41.04pg/mL, SD=35.49, highdistressprogesteronemean=58.96pg/mL,SD=36.84−(t (26)=−1.31,p=0.20).

Table 5). Women with higher distress scores had higher scores (worse mood) on all negative scales of the post- taskProfile ofMood States (includingtension, depression, anger,fatigue,confusion,andtotalmooddisturbance)com- paredtowomenwithlowerdistressscores,indicatingworse mood following the MIST (Table 1). Additionally, women with higher distress scores had significantly lower estra- diol levels (mean=43.47pg/mL, SD=35.26) than women withlowerdistressscores(mean=74.49pg/mL,SD=40.93) (t (26)=2.15, p<0.05), and there was an inverse corre- lationbetween estradioland distresswith lowerestradiol levels associated with greater distress (r=−0.50, n=29, p=<0.01). There was no difference in progesterone lev- els between distress groups—–low distress progesterone mean=41.04pg/mL,SD=35.49,high distressprogesterone mean=58.96pg/mL, SD=36.84−(t (26)=−1.31, p=0.20) (Fig.4).

Therewasnosignificantdifferenceinanyofthescreen- ingorstudydaysubjectiveorbehavioralmeasuresbetween participantsrunatthetwostudylocationsatp<0.05.

3.4. Cortisol

All subjectsshowed an increase in cortisolrelatedto the MISTandcortisolAUCwasassociatedwithdecreasedmedial prefrontalactivityduringpsychosocialstress(Supp.Fig.3);

howevertherewasnosignificantcorrelationbetweenestra- diollevelordistressscoregroupwithcortisolresponsetothe

MIST.CortisolAUCshowedsubstantialinter-subjectvariabil- itywithanaverageAUCfortheentiresampleof0.8anda standarddeviationof2.8.

4. Discussion

Overall, women showed brain activity patterns during psychosocialstressthatweresimilartothepatternsprevi- ouslyseen in studiesthatincluded both menandwomen;

deactivation of limbic regions (Pruessner et al., 2008).

However, this study showed that higher estradiol levels at periovulation were associated with greater hippocam- pal activity during psychosocial stress in normallycycling premenopausal women. Menstrual cycle phase and cor- responding estradiol levels were directly correlated with hippocampalactivityduringthestressconditionoftheMIST;

activity inthe hippocampusduring stresswassignificantly lower in women in the low estradiol levels compared to women with higher estradiollevels.These results suggest thatlowestradiollevelsduringtheearlyfollicularphaseof the menstrualcycleexaggerate theeffectofpsychosocial stress onbrain activity. Womenwith higher periovulatory estradiol levels also had lower distress scores following the psychosocial stress task.Group analysisbased ondis- tressfollowingtheMISTconfirmedtherelationshipbetween estradiol and both brain and mood response to stress:

women with higher distress had lower left hippocampal activityduringtheMIST,morenegativemoodfollowingthe MIST,andlowerestradiollevels.

PreviousworkusingtheMISThasrevealeddeactivations duringthestressconditionofthetask,inbrainareasthatare partofthelimbiccircuit(includinghippocampus,hypotha- lamus, medio-orbitofrontal cortex and anterior cingulate cortex) (Pruessner et al., 2008). These studies have pro- posed that reduced limbic circuit function is associated, perhaps causally, with the stress response. Additionally, hippocampal deactivation during the stress condition of the MISThasbeen directly relatedtocortisolreleaseand inversely correlated with measures of self-esteem, indi- cating that hippocampal activity is a marker of both the peripheralendocrineresponsetostressandrelatedtopsy- chological vulnerability to psychosocial stress (Pruessner et al., 2008). These findings are consistent with previ- ous studiesshowingthatthehippocampusisan important inhibitor oftheHPAsystem(JacobsonandSapolsky,1991).

Hippocampalfunctionisalsoimportantinprocessingemo- tionallyvalencedinformation(Canlietal.,2002),especially in women,andinevaluatingthecontextofsocialinterac- tions(Berthozetal.,2002).Ourfindingssuggestthathigher circulating estradiol may support continued hippocampal activityduringpsychosocialstress.

Althoughwedidnotseeadifferenceincortisolresponse betweenestradiolphaseordistressgroups,ourfindingthat cortisolresponsetotheMISTwasassociatedwithprefrontal deactivationagreewithanimalmodelsshowingworseper- formanceontasksthatrelyonprefrontalcortex following acutestress(Shanskyetal.,2004).Thisstudydidnotshow an effectofestradiolonprefrontalactivity,unlikeanimal models whichshow thathigh estrogenlevelsduring proe- strousinrodentsenhancethenegativeeffectsofstresson prefrontal function.It may that we were precluded from

seeingan interaction ofestradioland stressonprefrontal activitybythesamplesizeincludedinthisstudy,andfuture studies shouldexaminethis hypothesis throughprefrontal activityandperformancemeasuresinwomen.

The brainregions thatarecommonlyindicated instud- iesofmajordepressionandinrecentneuroimagingstudies ofpsychosocialstressareimportanttargetsfor investigat- ingvulnerabilitytomooddisorderandtheinteractionwith stressful life events.Thatthese sameregions arerespon- sivetoestradiolmanipulation(AndreanoandCahill,2010;

Goldsteinet al.,2005)is interesting in light of themood effectsof estradiolinboth healthy women(Gondaetal., 2008)andmooddisordersrelatedtoovarianhormonefluc- tuations.Estradiolhasbeenshowntomodulateactivityon brain circuits important for emotional processing; estra- diollevelchangesacrossthemenstrualcycleareassociated withchangesin brainactivityrelatedtoarousalfor nega- tivevalencedimages(Goldstein etal.,2005), andin fear conditioning (Zeidan et al.,2011). These findings provide evidence that estradiol may affect emotional responding throughincreased top-down modulation of emotional cir- cuitry,includingbrainareasinvolvedinthestressresponse, andmaybeprotectiveagainstfearandanxiety.Ourfindings suggest that greater estradiol levels during the periovu- latory phase of the menstrual cycle decrease the brain activity change in response to psychosocial stress, and reduce the acute behavioral and mood consequences of the stress experience. This interpretation is further sup- portedbythelocationofestradiolreceptorsinthecentral nervous system; the hippocampus is rich in both estra- dioland glucocorticoid receptors, makingit an important area of interaction between these hormones. In animal models,estrogenappearstosupportneuroplasticityatthe hippocampus (Gould et al., 1990; Woolley et al., 1990) andtobeprotectiveagainstthenegativeeffectsofstress (BredemannandMcMahon,2014).Highestradiollevelsdur- ing the periovulatory phase of the menstrual cycle may supporthippocampalfunctioninpartthroughtheseprotec- tivemechanisms.

Estradiol receptors are located in a number of brain areas,includingregionsimportantfortheautonomic,hor- monal, and cognitive-emotional response to psychosocial stress(Loveetal.,2010).Therelationofstresstodepres- siononset(Franketal.,1994;Kendleretal.,1999;Kendler etal.,2000)andthealteredfunctionofthestresssystem inmajordepression(Burkeetal.,2005;Lupienetal.,2009) suggestthatmodulationofthepsychosocialstressresponse may be a mechanism through which estradiol fluctuation maycontribute toMDD and PTSDrisk. The resultsof this studysuggestthatestradiollevelsmaymodulateactivityin brainareasimportantforprocessingemotionalinformation duringpsychosocialstress.Increasedemotional processing andphysiologicalresponsetopsychosocialstress,duringlow estrogen menstrual phases, may contribute to depressed moodinwomenwithvulnerabilitytoMDD.Indeed,women withMDDhavegreaterHPAaxisdysregulationthanmenwith MDD(Youngetal.,1994;YoungandRibeiro,2006),suggest- ing that the stress system may be particularly important todepression in women. Estradiol may attenuate sympa- theticandHPAaxisactivitytostress(KajantieandPhillips, 2006;Rocaetal.,2005).Althoughourstudydidnotinclude womenwithmooddisorders,theeffectofestradiolonbrain

activity andmood response topsychosocialstress suggest that periods of low estradiol are associated with height- enednegativeemotionalresponding.Thesephasesmaythus presentwindowsofincreasedvulnerabilitytothedepresso- geniceffectsofpsychosocialstressinwomen.Furtherwork isneededtodeterminewhethertheeffectsofestradiolon stressrespondingdifferinwomenwithvulnerabilitytomood and anxiety disorders, and ifthere is a relation between theoccurrence ofstressful events duringdifferentphases ofthemenstrualcycleandsubsequentMDDorPTSDonset invulnerablewomen.

Werecognizethatmenstrualcycleeffectsarelikelynot the only, or even the main, determinant of psychosocial stress responding inwomen; futurestudiesareneeded to examine the effects of personality factors, and lifetime traumaor chronicstress.Limitationsalsoincludethatthe separate roles and mechanisms of estradiol and proges- teroneinemotional processingandresponseremaintobe delineated. Although we did not directly investigate the effectofdifferingprogesteronelevelsonstressresponding, weattemptedtoisolatetheeffectsofestradiolbyhaving womenexperiencethe psychosocialstresstaskduring the early follicular phase or the periovulatory phases when progesterone levels are low. We excluded women whose estradiollevelswerenotinconcordancewiththeirassigned menstrualphasefromourmenstrualcyclephase/estradiol fMRIanalysis,asthesewomenwerelikelyinthemidluteal phase, when both estradiol and progesterone levels are increasing.Theperiovulatoryphaseisaveryshortduration oftime(2—4days)anditispossiblethatsomewomeninour study were studiedafter ovulation.Indeed, brainactivity duringthepsychosocialstresstaskdidnotshowadifference between groups basedonly onassigned menstrual phase, suggesting that the women who were omitted from the estradiol level- basedanalysis maynot have been in the correctphaseorthatestradiollevelratherthanphasehasa strongereffectonneuralactivityinresponsetopsychosocial stress.Althoughwebelievethisapproachwasappropriate toinvestigatetheeffectsofestradiollevel,itresultedina reducedsamplesizeforthemenstrualcyclephase/estradiol level analysis. As progesterone levels did not appear to have an effect on distress in this study, we included all participants in the analyses by subjective distress. This studywasnotdesignedtoinvestigatetheseparateeffects ofestradiolandprogesteronebymenstrualphase,however itdidrevealalinearinverserelationshipbetweenestradiol levelsandsubjectivedistressthatdidnotexistforproges- terone.Additionally,thisstudydidnotincludewomenwho experience mood disruptiondirectly related tomenstrual phaseorovarianhormones(suchasPremenstrualDysphoric DisorderorPolycysticOvarianSyndrome),nordidthisstudy model these disorders. The effect of ovarian hormone level/menstrual phase on brain activity during emotional processesinwomenwiththesedisorderslikelydiffersfrom healthy women, and the effects on stress responding in thesepopulationsshouldbeexamined separately.Another potentiallimitationofthisstudyisthedifference inblock length between stressand control conditionsof theMIST.

Thestressconditionsofthetaskneededtobeofsufficient lengthtoallowthestressresponsetodevelop,howeverthe limitationsintasktime(andefficientblocklength)didnot allowforthecontrolconditionstobeofsimilarlength.We

believe thecontrol condition wasofsufficient length and had enough repetitions to use asa comparison toisolate brainactivityrelatedtothestressresponse.

This study is unique in that that we used a moderate psychosocial stressor in the MRI environment at different menstrual phases, which allowed us to examineboth the subjectivemoodresponseandbrainactivityresponsetothe stress task.Alsoourstress taskincluded socialevaluative threat—–a typeof stressthat maybeparticularlyrelevant formoodinwomen(Kendleretal.,2001).Ourdatasuggest thatestradiolbuffersthebrainactivitychangesandnega- tivemoodresponsetopsychosocialstressinnormallycycling women.Thisworkhasimportantimplicationforanunder- standingoftherelationshipbetweenestradiollevelsandthe responsetostressfullifeevents.Whetherthishasimplica- tions for the developmentof psychopathology remains to bestudied.Futureworkshouldextendtheinvestigationof ovarian hormoneeffectsonpsychosocialstressresponding towomenwithvulnerabilitymoodoranxietydisordersand furtherexaminetherelationoftheseeffectstoknownrisk factorsformooddisorders.

Role of funding sources

This work is supported by NIAR01AG021476, GCRC MO1RR00109, Vanderbilt CTSA grant UL1 TR000445 from NCRR/NIH.

Conflicts of interest statement

Nonedeclared.

Acknowledgments

WewouldliketothankDr.TerryRabinowitzandDr.Warren TaylorfortheirhelpinadministeringtheMISTandsugges- tions inpreparing themanuscript.Thanksarealsodue to VioletGauforresearchassistance,andtheresearchnursing staffattheUniversityofVermontandVanderbiltUniversity CRCsfortheirsupportofthisstudy.Wealsothankourpar- ticipantvolunteersfortheirdedicationtoclinicalresearch.

References

Andreano, J.M., Cahill, L.,2010. Menstrual cycle modulation of medialtemporalactivityevokedbynegativeemotion.NeuroIm- age53,1286—1293.

Baca-Garcia, E., Diaz-Sastre, C., Ceverino, A., Garcia Resa, E., Oquendo,M.A.,Saiz-Ruiz,J.,deLeon,J.,2004.Premenstrual symptoms and luteal suicide attempts. Eur. Arch. Psychiatry Clin.Neurosci.254,326—329.

Beck,A.T.,Ward,C.H.,Mendelson,M.,Mock,J.,Erbaugh,J.,1961.

Aninventoryformeasuringdepression.Arch.Gen.Psychiatry4, 53—63.

Berthoz,S.,Armony,J.L.,Blair,R.J.,Dolan, R.J.,2002.AnfMRI studyofintentionalandunintentional(embarrassing)violations ofsocialnorms.Brain125,1696—1708.

Bredemann,T.M.,McMahon,L.L.,2014.17betaEstradiolincreases resilienceandimproveshippocampalsynapticfunctioninhelp- lessovariectomizedrats.Psychoneuroendocrinology42,77—88.

Bromet,E.,Andrade,L.H.,Hwang,I.,Sampson,N.A.,Alonso,J.,de Girolamo,G.,deGraaf,R.,Demyttenaere,K.,Hu,C.,Iwata,N., Karam,A.N.,Kaur,J.,Kostyuchenko,S.,Lepine,J.P.,Levinson, D.,Matschinger,H.,Mora,M.E.,Browne,M.O.,Posada-Villa,J., Viana,M.C.,Williams,D.R.,Kessler,R.C.,2011.Cross-national epidemiologyofDSM-IVmajordepressiveepisode.BMCMed.9, 90.

Burke,H.M.,Fernald,L.C.,Gertler,P.J.,Adler,N.E.,2005.Depres- sive symptoms are associated with blunted cortisol stress responses in very low-income women. Psychosom. Med. 67, 211—216.

Canli, T.,Desmond,J.E., Zhao,Z.,Gabrieli,J.D.,2002. Sexdif- ferencesintheneuralbasisofemotionalmemories.Proc.Natl.

Acad.Sci.U.S.A.99,10789—10794.

Chatterton,R.T.,Mateo,E.T.,Hou,N.,Rademaker,A.W.,Acharya, S., Jordan, V.C., Morrow, M., 2005. Characteristics of sali- varyprofilesofoestradiolandprogesteroneinpremenopausal women.J.Endocrinol.186,77—84.

CostaJr.,P.T.,McCrae,R.R.,2000.Overview:innovationsinassess- mentusingtherevisedNEOpersonalityinventory.Assessment7, 325—327.

Davydov,D.M.,Shapiro,D.,Goldstein,I.B.,Chicz-DeMet,A.,2005.

Moodsineverydaysituations:effectsofmenstrualcycle,work, andstresshormones.J.Psychosom.Res.58,343—349.

First, M.B., R.L., S.,Gibbon, M., Williams, J.B.W., 2001. Struc- tured ClinicalInterview forDSM-IV-TRAxis IDisorders-Patient Edition, SCID-I/P,2/2001ed.American PsychiatricPressInc., Washington,DC.

Frank, E.,Anderson,B.,Reynolds3rd,C.F., Ritenour,A.,Kupfer, D.J., 1994. Life events and the research diagnostic crite- ria endogenous subtype. A confirmation of the distinction usingtheBedfordCollegemethods.Arch.Gen.Psychiatry51, 519—524.

Goldstein,M.J.,Poldrack,Russell,Ahern,Todd,Kennedy,DavidN., Seidman, LarryJ., Makris,Nikos,2005. Hormonalcyclemod- ulates arousal circuitry in women using functional magnetic resonanceimaging.J.Neurosci.25,9309—9316.

Gonda, X., Telek, T., Juhasz, G., Lazary,J., Vargha, A.,Bagdy, G.,2008.Patternsofmoodchangesthroughoutthereproduc- tive cyclein healthy women without premenstrual dysphoric disorders. Prog. Neuropsychopharmacol. Biol. Psychiatry 32, 1782—1788.

Gould, E., Woolley, C.S., Frankfurt, M., McEwen, B.S., 1990.

Gonadalsteroidsregulatedendriticspinedensityinhippocam- pal pyramidal cellsin adulthood.J.Neurosci. 10,1286—1291 (theofficialjournaloftheSocietyforNeuroscience).

Hankin,B.L.,Mermelstein,R.,Roesch,L.,2007.Sexdifferencesin adolescentdepression: stressexposureandreactivity models.

Child.Dev.78,279—295.

Jacobson, L., Sapolsky, R., 1991. The role of the hippocampus in feedback regulation of the hypothalamic—pituitary—adrenocortical axis. Endocr. Rev.

12,118—134.

Kajantie,E.P.D.,2006.Theeffectsofsexandhormonalstatusonthe physiologicalresponsetoacutepsychosocialstress.Psychoneu- roendocrinology31,151—178.

Kajantie,E.,Phillips,D.I.,2006.Theeffectsofsexandhormonal statusonthephysiologicalresponsetoacutepsychosocialstress.

Psychoneuroendocrinology31,151—178.

Kendler,K.S.,Karkowski,L.M.,Prescott,C.A.,1999.Causalrela- tionship betweenstressfullifeeventsand theonsetofmajor depression.Am.J.Psychiatry156,837—841.