fMRIbrainvoyager 2文档格式.docx

fMRIbrainvoyager 2文档格式.docx

《fMRIbrainvoyager 2文档格式.docx》由会员分享，可在线阅读，更多相关《fMRIbrainvoyager 2文档格式.docx（11页珍藏版）》请在冰豆网上搜索。

Howcangowrong:

alterbloodflow（drgus,hormonalstates,age,gender）.Sometimestherearesomeneuralactivitydifferencesindifferentconditions,perhapstherearesomedifferencesinneuralresponesesbutitmaynotbemeasuredbyfirm.Onewaythatcanhappyisthatitisnotabulkchangeorstablechange.Sincefmrishowanaverageresponseofthetimespace.

Voxel,60,000perhapsbeanimage.

Howtomakeaclaimfromawellmakebraintomentalthoughtsofaman.

thebrainareafor“love”.Howtomakeinferencementaloperation.

∙Brainactivityandbehaviour:

Ø

Forwardinference:

determinewhichbrainregionisassociatedwithanisolatedbehaviour.Parametricstudy:

FocalReverseinference:

uselocalbrainactivityto（determinewhichmentalstatesareevokedbyacomplexbehaviour）identifymentaloperation.Nootherbrainareaactivated.Justforpain.Goodespecifition.Areyoulying?

Mindreadingispossible?

Lookdifferencesbetween

Multi-voxelclassification:

usedistributedpatternsofbrainactivitytopredictwhichmentalstateisbeingexperienced.

Signalandnoise

Fouriertransform:

analysesignals,powerspectrum（功率谱）。

Onlychangesthatarequiteslowlycanbedetectablebythefmri,changesthatquiterapidlyupsanddowncannotdetectedbythefmri（notpasswell）.Likeevery3/4seconds.Designexperimentsthatwerereallychangesslowly.5minsfaces,5minshouseswouldbebest,butthatisnotthewholestorybecauseofthenoise.

Becauseofthestuckbetweenthehemodynamic（人的血液动力反应）responsesuggestthelowfrequencyoftheexperimentorstimulus,whileintrinsicnoise（固有噪声）suggestthehighfrequencyofthestimulus.Sohowtobalancebetweenthesetwo.

20swouldbeabetterchoice.Blockdesign.20s-30sseconds.Frequencyrepretations.Whatisthenoise?

couldbeheartrate,breathing,brainareas/cortexcommunicated/connectedwitheachother.

Frequencies.

Experimentaldesignhaslowfrequency.Smooththedata.wenormallysmooth6mm.however,thereareotherpeoplewhosmoothdatain8-12mm.donotuseGLM,ifyoudoit,youwantto.

3*3*3brainmodule.Youranotomicaldataishighrevolutionthanyourfunctiondata.

.

fMRI:

lecturesfromFrank:

Presentation:

tellthepresentationtostart----presentationstart—tellthescanstart.

Whatthefinalanalysisaregoingtobe.Plannedeverything,knowinadvancewhatanalysissoftwareyouareGonneuse.

Fmrisafetycourse,youshouldtakethesafetycourse,theywillshowalltheworstthingthatcouldgowrong.Videofrequencyinthescanner,thatisquitedangerous.Minimizethebadthings.

Projector,highresolution,eyetrakler.

Programmetorunthecomputer:

Computerwillcontrolthedeliveryofthestimuli,potentialprogramesthatcouldrunthecomputeris:

presentation（whichistheonewegenerallyuse）,eprime,andmatlab,eachoneofthoseisok.Showthingsintothescanner,needthePCandscantalktoeachother.

Oncewegointothebrainvoyager,brianvoyageralwaysgiveanoptiontoskipseveralvolumes.Itisgoodtogetridit.Thereasonis:

Ifyoulookatthestatisticanalysis,allthedatathatinclusesomeconditionsthatsometimesgoestoyourbaseline,andifyourbaselineincludesthesenoiseanditannoysyourbaseline,sojustskipsomevolumesthatdoesnotcontributeatall.Butitisimportanttokeeptrackofthedataifmultipeopleworkonthesamedata.

Presentationlogfile:

itisgoodtohavethelogfilefromthestimulusdelivery.Sothecomputersisshowingthestimulu,itisreallyhandyifkeepingtrackofthelogofwhatisgoingon.Whenitisshowing,whatisupthescanningandwhatishearingfromthescanner.

Youneedhavethelogfilefromthescanningsession.ifyouhavethelogfileandhaveitlogged:

logtheresponse,loginallinformation,youcangobackandcheckeverythingfindeverything（wedidnotlogoneofourparticipantsresponses,andthenwewanttoanalyse,butwedidnothavetheseresponsessincewedidnotlogin）,oneword,ifyoudonothavethesestuffatthebeginning,thenyoudonothavetheseatall.

Twowaystodotimingwithbrainvoyager:

uniformtime,2000msTR=2.Ifyouwanttoseethebraininms,youreallyneedtodoitbyyourself,otherwise,thescannerjusttakethebrainevery2s.

Whenthesubjectwasinthescanner,whathappenedtothescannerorcomputer.Attheendofthesessionoratthebeginning（itisuptoyou）,werequiretheanatomicdata.Itwilltake4-8minstocollecttheanatomicaldata,thedifferencesbetween4-8minsisthelongertime,thehighercontrastbetweenthewhitematterandthegreymatter.

InFriday,wedid2functionalrunsand1aunotomicrun.Everytime,thereisarunalwaysproduceonefile,butthefunctionalrunsalwaysproduce2files,oneisrawdata,theotherismotioncorrected（moco）data.Wegeneralusetherawdata（frankspreferstousetherawdata,butinsomeothermanuals,someofthemusethemotioncorrecteddata,whichisthesecondoneofeachrun）.

Theprocess:

（thesestepsareubtertwined）

Planning,stimuluspresentation,measurement,pre-processingandstatisticanalysis

∙Synchronisations:

scannerwaitsfor7s,untilsignalstobestable,pressbuttonofstimuluscomputerandscanneratthesametime.Soskipfirst1or2volumesthissynchronisationissueneedtobedecidedsincelaterinpreprocessingBVQxwillaskyouifyouwanttoskip.Logfile:

formatistext,every2seconds.#

∙Anatomicalrun:

192sclies,1*1*1mm（thewholehead）

Functionalrun:

32sclices,3*3*3mmandawholebrainvolumeacquiresevery2seconds.

Preprocessing:

∙Anatomicalpreprocessing:

talariakithebrain,filterthebraininthesameway（covertthebraintothetalariakespace）

∙Functionalpreprocessing:

∙Corregistration

CreatPRTfilewhenyouwanttocontrastdifferentconditions.

Brainnormalizationistoputbraininthetalariakispace.Thecontrastofthebrainactivitywhenyouseescrambledpersonandintactperson.

Weusethewizardtoputinthedata.

32slices,ifyouhavesomeareathatyoureallyinterestedininparticular,youmaychangethefunctionalruntothespecificbrain,butwewouldbealittlebitopentoseethewholebrain,toseewhatotherbrainareasmaybeinvolved.Wetrytomeasurethewholebrain.Typically,youcando32sclices,3*3*3voxelsandacquirethewholebrainevery2sceonds.Ifpeoplehavesmallbrains,thatisenough,butifpeoplehavelargerbrain,maybethereissomethinghasmissed.Youhavetheoptionofeithermissingthesopsuperiorfrontalcorxetonthetoporthetemporalcortex.Alltheanotomicdata,everyonealmostgetthesamething.Nomatterthebrainsize.Wherethesekindsofissuesbecomeimportant,youwanttoseewhathappensonaverageabout20brains,hereyouneedtousethenormalization.thatall20brainsatsomehowshouldbealigned.therearetechniquestoallowyou.presentonebrainatatime,yougivepreliminarytheregionyouareinterested..

Butweusuallydoistoaverageallthebraintogether,thensomehowyouneedaspacethatcombineallthesebrains.Allthebrainshavedifferentshapesandsize,evenwetalarikithesecorrectly,cortexbasedalignment..thereisstillalittlechancetohavebrains.

CreatetheVMRanatomicaldata.

CreatetheVMRinthestep9,page41,checkthenumberofsclices（192）-next,checktheresolutionofanatomicalscan/images（256*256）.Thenchecktheseslicesandresolutions.---checkthecontrastandbrightness（generallythedefaultisok,sokeeptheoptionsremainthereandthenpressOK）.

AftercreatingtheVMR,youcanjustchangetheintensityofthebrain.Whichisinhomogeneitycorrection,V16Tools.OriginalVMR–volumes—inhomogeneitycorrection-loadv6—justclicktheV16file（wecanalsodothisstepbeforetalairach）.

CreatetheFMR,Createnamethefirstvmr（therawdata）asR1（here,itisimportanttousethe_underscoretoavoidtheconfusionoffilenames）,choosethenumberofvolumesis243,（keepeverythingasdefault）.AndthencreateFMRR2,heretheoriginaldesignfortheexperimentis243（thefirstone-first）,butfrancesalwaysrunthescanlongerthantheexperimentdesign,andthenjustpressstopwhenitisok.Inordertobeequalastoourdesign,letusgofor243.Thenfornumberofvolumes,letusgofor243.Then,westillnotquitetrustthefirsttwovolumes,sowegoskipthefirsttwovolumes.

Toseethetimecoursemovie:

gototheoptions—timecoursemovies（givethequalitiesofthemovie）.

AttachPRTfile:

openthefmrwejustcreated,gotoFMRproperties,choosethefmrrun1—choosethePRTfile（choosetherun1PRTfile）–close

Thenopenthefmrrun2,gotoFMRproperties,choosethefmrrun2—choosethePRTfile（choosetherun2PRTfile）–close

Thus,ourPRTfileisattachedtooutfmr.

Preprocessingthefunctionaldata,closeotherfileandonlykeepthefmropen,gotooptions—preprocessingFmri—clickpreprocessingslicescantimecorrection,3Dmotioncorrection–spatialsmoothing--temporalfiltering（themeanintensityadjustment,whileisnotneedinmostcases,sodon’tchoosethis）---spatialsmoothingistheonegeneralcanbedowna