說(shuō)明分析教案0210

fMRIpost-stimulusundershootsinvisualcortexareneuronalinKarenJMullinger1,2,MatthewCherukara1,SusanTFrancis1,andStephenD1SPMIC,SchoolofPhysicsandAstronomy,UniversityofNottingham,Nottingham,Nottinghamshire,UnitedKingdom,2BUIC,SchoolofPsychology,UniversityofBirmingham,Birmingham,WestMidlands,UnitedKingdomIntroduction:Thepost-stimulusundershootisawellrecognisedcomponentofthe response[1].However,theBOLDresponseoriginatesfromacomplexinctionbetweencerebralbloodflow(CBF),cerebralbloodvolume(CBV)andthemetabolicrateofoxygenconsumption(CMRO2),makingthephysiologicaloriginofthepost-stimulusundershootunclear.InarecenthumanEEG-fMRIstudy[2],wehaveprovidedevidencethatthepost-stimulusBOLDundershootis,atleastinpart,neuronalinorigin.Thisworkshowedthat

AmplitudeAmplitude

2SignalSignalChange0

80CSignalSignalChange0

post-stimulusevent-relatedsynchronisation(PERS)powerofEEGactivityinthe8-13Hz(alpha/mu)frequencybandwascorrelatedwiththenaturalvariabilityintheamplitudeofthepost-stimulusBOLDandCBFundershootsinthesensorimotorcortex.Here,werecordEEG-fMRIresponsestoflickeringandstaticcheckerboardvisualstimulipreviouslyshowntodrivedifferencesinthepost-stimulusBOLDundershoot(flicker=strongundershoot,

Time Time TimeFigure1:GroupEEGalphapower(A),BOLD(B)andCBF(C)responsesfromcontralalV1totheflicker(green)andstatic(purple)visualstimuli.Significantdifferences(pairedT-tests)intheresponseamplitudesbetweenflickerandstaticstimuliatcertaintime-pointsaredenotedby**=p<0.001,*=p<0.05.Hypothesis:Previouslyobservedpost-stimulusEEG-fMRIcorrelations[2]canbeelicitedinothersensorymodalities;adifferenceinthemeanpost-stimulusBOLDundershootbetweenflickerandstaticstimuliwillbeassociatedwithacorrespondingdifferenceinPERSalphapower,whilstcorrelationsbetweenthenaturaltrial-by-trialvariabilityofPERSalphapowerandBOLDpost-stimulusundershootwillalsobeobserved.Ifproventhiswouldprovidenewevidencethatpost-stimulusundershootsareneuronalinoriginthroughoutthebrain.Methods:fMRIandEEGdatawereacquiredsimultaneouslyusingaPhilipsAchieva3TMRscanneranda64-channelEEGsystem(BrainProducts).AFAIRDABS[2]sequencewasusedtosimultaneouslyacquirebackgroundsuppressedASLandBOLDdatafromprimaryvisualcortex(V1)(TR=2.6s,TE=9.2/40.2ms[ASL/BOLD],labeldelay=1400ms,3x3x5mm3voxels,12slices,212mmFOV,SENSEfactor2.3,backgroundsuppressionatTIBGS1/TIBGS2=339ms/560ms).Datawereacquiredon14subjects(age=27±6yrs)whofixatedonacentrallydisyedcrossthroughoutwhilstpassivelyviewingstatic(100%contrast)andflicker(3Hz,33%contrast)left-hemifieldcheckerboardstimuli.StimulusintensitieswerematchedtoinducecomparableprimaryBOLDresponseamplitudes[3].Datawererecordedover4runswith8pseudo-randomisedtrialsofeachstimulusperrun(10s/30sstimulation/rest)anda98.8srestperiodatthestartofeachruntomeasurefMRIbaselinesignal.ysis:1subjectwacludedfromfurtherysisbecausetheyfellasleep.Intheremaining13subjectsEEGandfMRIdatawerepre-processedusingconventionalmethodsinBrainVisionyzer2andFSL[2].EEG:Pre-processeddatawerefiltered8-13Hz.Aregularised,scalarbeamformerwasusedtolocalisethealpharesponsetovisualstimuli[2,4].VirtualelectrodetimecoursesofelectricalactivitywereextractedfromthepeakpseudoT-statlocationincontralalV1(active/passivewindow:0-9.5s/30-39.5s)derivedusingallstimuluspresentations.Averageresponsesforthestaticandflickerstimuliwerefound.Fortheaverageresponsesandindividualtrialsinthetwostimulusconditions,themeanstimulusresponse(0-9.5s),PERS(10.5-20s),andcontrol(35-39.5s)windowalphapowervalueswerecalculated.Foreachsubjectandstimuluscondition,trialsweresortedintolower(0-25%),median(37.5-62.5%)andupper(75-100%)quartilesbasedon:stimulusresponse,PERS,orcontrolwindowalphapower[2].fMRI:GLMysis(FEAT6.01)identifiedsignificantpositiveandnegativeresponsesofBOLDandCBFsignalscombinedacrossbothvisualstimuli.Thegroup-levelconjunctionofBOLDandCBFresponsestobothstimuliwasusedtodefinesubjectspecific,cubicregionsofinterest(ROIs,3x3x3voxels).Thesewerecentredonthesubject’speakvoxelinthepositive(contralal)andnegative(ipsilal)V1BOLDregions[2].BOLDandCBFsingle-trialhaemodynamicresponses(HRs)wereextractedforeachtrial;allowingdirectcomparisonbetweenCBFandBOLDresponsesandstimulusconditions.HRswereconvertedtopercentagechangerelativetotheinitialbaselineperrun.HRsweresortedintoquartilesaccordingtoeitherthePERS,stimulusresponseorcontrolwindowalphapowerseparayforthestaticandflickerstimuliandtheagedoversubjects.Results:EEG:Figure1Ashowstheaveragealphapowerresponsestobothstimuli.Pairedstudentt-testsrevealedsignificantlyhigheralphapowerinflicker(green)thanstatic(purple)trialsduringthePERS(p<0.001)windowbutnosignificantdifferenceinthestimulusorcontrolwindows.NosignificantcorrelationbetweenPERSalphapowerandeitherstimulusresponseorcontrolwindowalphapowerwasobserved.fMRI:TheprimaryBOLD/CBFsignalresponse(at~10s)tovisualstimulationwaspositiveontralalV1(Fig.2pink)andnegativeinipsilalV1(Fig.2green).ThegroupaveragecontralalBOLDandCBFresponses(Figure1B&C)bothshowedsignificantlylargerundershootsforflickerthanstaticstimuli,agreeingwith[3].InbothpositiveandnegativeBOLD/CBFregions,thepost-stimulusundershootamplitudewasnegativelycorrelatedwithPERSalphapower(Fig.2).Foreachtimepoint,aone-wayRMANOVAwasusedtotestforsignificantdifferenceinHRamplitudweenPERSalphapowerquartiles.Intheundershootperiod,asignificanteffectofPERSpowerwasobservedontralalV1BOLD(18.5-25sflicker&16.7-24.5sstatic)andCBF(17.9-22.5sflicker&20-22.5sstatic)amplitudes(Fig2A,C,E&G).ForipsilalV1(Fig2B,D,F&H),asignificantdifferenceinundershootbetweenquartileswasonlyobservedforthenegativeBOLDtimecoursefortheflickeringstimulus(20.4-24.5s)(Fig2B).NosignificantsortingeffectontheprimaryfMRIresponseswasobservedforeitherstimulus.NosignificanteffectsonBOLDorCBFresponseswerefoundwhensortingaccordingtoeitherstimulusorcontrolalphapower.Discussion:Here,weprovidefurtherevidencethattheBOLDpost-stimulusundershoot(>17s)islinkedtopost-stimuluschangesinelectricaloscillatoryactivity(10-20s)inhumans.Weadvancepreviousworkintwoways.Firstly,byshowingthatcorrelationsbetweennaturalvariabilityinpost-stimulusEEGandfMRIresponsesseeninsensorimotorcortex[2]extendtothevisualsystem;trialswithhigherpost-stimulusalphapowerexhibitedmorenegativeBOLD/CBFundershootsinbothpositiveandnegativeBOLDregions,agreeingwith[2].ThelackofrelationshipbetweenfMRIundershootandeitherstimulusorcontrolalphapower,providesfurtherevidencethatourfindingsspecificallyreflectcorrelationofthepost-stimulusneuronalandfMRIsignals.Secondly,byshowingaconcurrentsignificantdifferenceinthemeanpost-stimulusalpha,BOLDandCBFresponsesinthecontralalhemispherebetweenflickerandstaticstimuli,withhigherPERSalphapowerandlargerundershootsobservedfortheflickerstimulus,complementing[3].SealphaERSiscommonlybelievedtoreflectinhibitoryneuronalactivity[5],wehypothesizethatachangeinthebalanceofexcitatory/inhibitoryactivityoccursuponterminationofstimulationcausingachangeMRO2.Wesuggestthislonglastingchangeinneuronalactivitymaybeamechanismbywhichthebrainreturnstoarestingstate.Giventhegrowingevidencethatpost-stimulusundershootsareneuronalinorigin,andrelativelyslowresponses,futureworkwillfurtherinvestigatethefunctionalroleofthesebrainresponses.Proc.Intl.Soc.Mag.Reson.Med.23 fMRIKarenJMullinger1,2、MatthewCherukara1、 Mayhew21SPMIC，物理與天文學(xué)學(xué)院，諾丁漢大學(xué)，諾丁漢，諾丁漢郡，英國(guó)，2BUIC，翰大學(xué)心理學(xué)院，翰，西茲，英國(guó)簡(jiǎn)介：刺激后下沖是公認(rèn)的大膽的回應(yīng)[1]。然而，BOLD反應(yīng)源于腦血流EEG-fMRI[2BOLD關(guān)同步(PERS)功率與刺激后BOLD和CBF下沖幅度的自然變異相關(guān)。感覺(jué)運(yùn)動(dòng)皮層。在這里，我們記錄了對(duì)閃爍和靜態(tài)棋盤(pán)視覺(jué)刺激的EEG-fMRI反應(yīng)，先前顯示這些刺激會(huì)驅(qū)動(dòng)刺激后BOLD下沖的差異（閃爍=強(qiáng)下沖，靜態(tài)=無(wú)下沖[3]），以進(jìn)一步研究刺激后下沖。EEG-fMRI[2]BOLDPERSαPERSα方法：使用PhilipsAchieva3TMR掃描儀和64通道EEG系統(tǒng)（BrainProducts）同時(shí)采集fMRIEEGFAIRDABS[2(V1)ASLBOLD數(shù)據(jù)（TR=2.6s，TE=9.2/40.2ms[ASL/BOLD]，標(biāo)簽延遲=1400ms，3x3x5mm3體素，12個(gè)切片，212mmFOV，SENSE因子2.3，TIBGS1/TIBGS2=339ms/560ms的背景抑制）。數(shù)據(jù)采集自14名受試者（年齡=27±6），他們始終注視著中央顯示的十字，同時(shí)被動(dòng)地觀看靜態(tài)（100%度）和閃爍（3Hz，33%對(duì)比度）左半視野棋盤(pán)刺激。刺激強(qiáng)度相匹配以誘導(dǎo)可比較的初級(jí)BOLD反應(yīng)幅度[3]。數(shù)據(jù)記錄在4次運(yùn)行中，每次運(yùn)行每次刺激進(jìn)行8次偽隨機(jī)試驗(yàn)（10秒/30秒刺激/休息），每次運(yùn)行開(kāi)始時(shí)有98.8秒的休息時(shí)間來(lái)測(cè)量fMRI基線信號(hào)。分析：113像數(shù)據(jù)使用BrainVisionyzer2和FSL中的常規(guī)方法進(jìn)行預(yù)處理[2]。EEG：預(yù)處理數(shù)據(jù)經(jīng)8-13Hzalpha[2,4]。電活動(dòng)的V1T（：0-9.5秒/30-39.5秒）中提取的。發(fā)現(xiàn)了靜態(tài)和閃爍刺激的平均反應(yīng)。對(duì)于兩種刺激條件下的αPERSalpha2ofMRI：GLM分析(FEAT6.01)確定了兩種視覺(jué)刺激中BOLD和CBF信號(hào)組合的顯著陽(yáng)性和陰BOLDCBF（ROI，3x3x3體素）。這些以受試者在正（對(duì)側(cè)）和負(fù)（同側(cè)）V1BOLD區(qū)域的峰值體素為中心[2]。為每個(gè)試驗(yàn)提取BOLD和