您的当前位置：首页 Optical imaging of intrinsic signals

Optical imaging of intrinsic signals

来源：图艺博知识网

JournalofNeuroscienceMethods136(2004)1–21

Invitedreview

Opticalimagingofintrinsicsignals:recentdevelopmentsinthe

methodologyanditsapplications

AngelicaZepedaa,ClorindaAriasa,FrankSengpielb,∗

DepartamentodeBiolog´ıaCelularyFisiolog´ıa,InstitutodeInvestigacionesBiomédicas,

UniversidadNacionalAutónomadeMéxico,México,DF,Mexico

CardiffSchoolofBiosciences,CardiffUniversity,MuseumAvenue,CardiffCF103US,UK

Accepted16February2004

Abstract

Sinceopticalimaging(OI)ofintrinsicsignalswasﬁrstdevelopedinthe1980s,signiﬁcantadvanceshavebeenmaderegardingourunderstandingoftheoriginsoftherecordedsignals.Thetechniquehasbeenreﬁnedandtherangeofitsapplicationshasbeenbroadenedconsiderably.Herewereviewrecentdevelopmentsinmethodologyanddataanalysisaswellasthelatestﬁndingsonhowintrinsicsignalsarerelatedtometaboliccostandelectrophysiologicalactivityinthebrain.Wegiveanoverviewofwhatopticalimaginghascontributedtoourknowledgeofthefunctionalarchitectureofsensorycortices,theirdevelopmentandplasticity.Finally,wediscusstheutilityofOIforfunctionalstudiesofthehumanbrainaswellasinanimalmodelsofneuropathology.©2004ElsevierB.V.Allrightsreserved.

Keywords:Opticalimaging;Intrinsicsignals;Event-related;Fourieranalysis;Functionalarchitecture;Development;Plasticity

1.Introduction

Inrecentyears,functionalbrainimagingtechniqueshavetakenovermoreandmorefromclassicalphysiologicalap-proaches,suchasextracellularsingle-cellrecordingsinstud-iesofthemammalianbraininvivo.Functionalimagingtechniquesrelyonasimilarfundamentalapproachtoun-derstandingbrainorganization;however,theygreatlydifferintheirspatialandtemporalresolutions.Theirprincipalad-vantagesarereducedinvasiveness,theabilitytofunction-allycharacterizelargeareasofthebraininresponsetoasetofstimuli,andtherelativeeaseoflongitudinalstudiesbyrepeatedimagingofanindividualsubject.

Inthelastdecade,imagingstudieshaveshedlightonthefunctionalorganizationofthenormalbrainandmorerecently,onthereorganizationoftheinjuredcortex.Thesestudieshavebeenperformedusingfunctionalmagneticres-onanceimaging(fMRI),near-infraredspectroscopy(NIRS)andopticalimaging(OI)ofintrinsicsignals,allofwhicharebasedonchangesinbloodoxygenationandinopticalor

Correspondingauthor.Tel.:+44-29-2087-5698;fax:+44-29-2087-4094.

E-mailaddress:sengpielf@cf.ac.uk(F.Sengpiel).

∗

magneticpropertiesofneuraltissuecausedbyphysiologicalactivity.

Theprimaryuseofopticalimaginginthepast12yearsorsohasbeenthevisualizationoffunctionalcorticalmapsandtheirarchitecture.PriortotheadventofOI,thefunc-tionalcorticalarchitecturehadbeenassessedmainlywithelectrophysiologicaltechniques(extracellularsingle-andmulti-unitrecordings)andthrough2-deoxyglucose(2-DG)labeling.However,despitetheiruses,thesetechniqueshavemajorlimitations.Electrophysiologicalmappingofalargeareaofcortexisinvasive,time-consumingandsubjecttosamplingbias,whereas2-DGmappingcangenerallyonlybeperformedforoneparticularstimulus(veryrarelyfortwo)andmapscanonlybeanalyzedpostmortem.Thus,chronicexperimentsarenotfeasible.

Opticalimagingofintrinsicsignals,atechniquedevel-opedbyGrinvaldandco-workers(BonhoefferandGrinvald,1996;Frostigetal.,1990;Grinvaldetal.,1986,1999;Ts’oetal.,1990)hasbeenusedverysuccessfullytostudybothacutelyandchronicallytheprinciplesunderlyingorganiza-tionandfunctionalarchitectureofdifferentcorticalregionsinseveralspecies,includinghumans;corticaldevelopmentandsensoryinformationprocessinginvivo.Thetechniqueemploysappropriatesensorystimulitoobtainhighresolu-

2A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–21

tionfunctionalmapsfromarelativelylargearea.Anumberofmapsinresponsetoasetofstimulicanbeobtainedfromthesamecorticalarea,whichcanbeimagedrepeatedlyoveraperiodofweeksorevenmonths.Opticalimagingisproba-blythetechniquethatbestcombinesspatialresolution,cov-erageandspeedforfunctionalmappingofthemammaliancortex.

Theaimofthisreviewistoprovideanoverviewofre-centdevelopmentsinthemethodologyofopticalimagingofintrinsicsignalsandtointroducesomeofitsrecentappli-cations.Inthefollowingsection,wewillﬁrstdescribethemainprinciplesofthetechniqueandimportantexperimen-talaspects.WewillthendiscussthelatestadvancesinOIdataanalysisandfocusonitscontributionstovariousﬁeldsofneuroscience.AmorecomprehensivedescriptionofthetechniquecanbefoundinBonhoefferandGrinvald(1996)andGrinvaldetal.(1999).

2.Methodologicalandtechnicalaspectsofopticalimagingbasedonintrinsicsignals2.1.Sourcesofintrinsicsignals

OIofintrinsicsignalsisthevisualizationofchangesofin-trinsicopticalpropertiesofneuraltissues,inparticularlightreﬂection,duetoneuronalactivity.Thesurfaceofthebrainisilluminatedandimagesarerecordedwithacharge-coupleddevice(CCD)camerawhilethesubjectisbeingstimulated.Thesourcesoftheintrinsicsignalincludereﬂectancechangesfromseveralopticallyactiveprocesses(Cohen,1973),whichcorrelateindirectlywithneuronalﬁring.Atleastthreecharacteristicphysiologicalparametersaffectthedegreetowhichincidentlightisreﬂectedbytheactivecortex.Theseare:(a)changesinthebloodvolume;(b)chromophoreredox,includingtheoxy/deoxy-hemoglobinratio(oxymetry,seebelow);intracellularcytochromeox-idaseandelectroncarriersand;(c)lightscattering(seebelow)(Frostigetal.,1990;Narayanetal.,1994a,1994b).Theﬁrsttwofactorsrelyprincipallyontheincreasedmetabolicdemandofactivecerebraltissue(i.e.ofneu-rons)andonsubsequentdeoxygenationofhemoglobininthemicrocapillaries.NeuronalactivitycauseshydrolysisofATP(seebelow)andtheregenerationofATPbyglucosemetabolismrequiresoxygen(1molofO2per6molofATP).Oxy-hemoglobinmoleculesinthecapillarieswithinanac-tivecorticalareaaretheprimarysourceofoxygen.There-foreduringmetabolicdemand,aﬂuxofoxygenfromthecapillariestoadepletedregioncausesahighlylocalizedincreaseindeoxy-hemoglobinconcentration.ThattheveryﬁrsteventfollowingasensorystimulusisindeedalocaldecreaseinoxygenconcentrationwasrecentlyshownbyVanzettaandGrinvald(1999),whodirectlyassessedmicro-capillaryoxygenconcentrationbymeasuringthequench-ingofaphosphorescentprobe.Opticalimagingmakesuseofthedifferentabsorptionspectraofoxy-hemoglobinand

deoxy-hemoglobin,thelatterhavingahigherabsorptionco-efﬁcientatwavelengthsof600nmandabove.Activecorti-calregionscanthereforebedistinguishedfromlessactiveareassincetheformerreﬂectlessredlightthanthelatter(Frostigetal.,1990;Grinvaldetal.,1986).Thedifferenceinreﬂectancechangebetweenactiveandinactiveregionsisknownasthe“mappingsignal”(seebelow).

Localriseofdeoxy-hemoglobin,ordepletionofoxy-hemoglobin,isfollowedwithin1–2sbylocalcapillaryrecruitmentanddilationofadjoiningarterioles(Maloneketal.,1997).Theresultingincreaseinlocalbloodﬂowandvolumeofoxygenatedbloodcausesadecreaseindeoxy-hemoglobinandanincreaseinoxy-hemoglobin,al-beitlesswellco-localizedwiththeareaofinitialoxygenconsumption.Thus,acloserelationshipexistsbetweenlocallyincreasedneuronalactivityandthehemodynamicresponse.Thisso-calledneurovascularcouplingprovidesalinkbetweenlocalneuronalactivityandcerebralmicrocir-culation(VillringerandDirnagl,1995).

Variousfunctionalimagingtechniquesutilizedifferentas-pectsofthislink.Whiletheearlydecreaseinbloodoxygen,orincreaseindeoxy-hemoglobin,formsamajorsignalcom-ponentinopticalimagingofintrinsicsignals(seeabove),thebloodoxygenationlevel-dependent(BOLD)signalmea-suredinfMRIisattributabletothedelayedandprolongedin-creaseinbloodoxygenation.Duetotherecruitmentofarte-riolesinthevicinityoftheoriginalsiteofoxygenconsump-tion,thespatialresolutionofthisdelayedsignalissomewhatlimitedanddoesnotallowvisualizationofindividualfunc-tionaldomainsinthecortex.However,high-ﬁeldfMRImea-surements(atupto9.4T)haveconﬁrmedthepresenceofan“initialdip”,thatisashortlatencydecreaseinbloodoxy-genationcorrespondingtoanincreaseindeoxy-hemoglobinconcentration(Kimetal.,2000).Asthisisconﬁnedtothesiteofneuronalactivity,itallowsfunctionalimagingwithamuchhigherspatialresolutionthanimagingbasedonlaterincreasesinbloodﬂowandvolume.Indeed,high-ﬁeldfMRIthatutilizesonlytheinitialdipiscapableofresolvingindi-vidualcorticalmodules,suchasorientationcolumnsincatprimaryvisualcortex(Kimetal.,2000),similartoopticalimagingofintrinsicsignals.Arecentstudyinhumanscom-paringspatiotemporalpatternsoffMRIsignalsandintrinsicopticalsignals(measuredintra-operatively)alsosupportedtheconclusionthattheinitialfMRIdipandtheintrinsicOIsignalresultfromthesamephysiologicalevents(Cannestraetal.,2001).

Thethirdfactordeterminingcorticalsurfacereﬂectance,lightscattering,wasﬁrstdiscoveredinthecrablegnervebyHillandKeynes(1949)andhasproventobeaparticu-larlyusefulsignalforfunctionalmappingbecauseofitsrel-ativelytightspatialandtemporalcouplingwithneuralactiv-ity.Inopticalimagingofthelivingbrain,theincidentlightisscatteredtosomeextentasitpenetratesandisreﬂectedthroughtheneuraltissue.Lightscatteringincreasesasacon-sequenceofincreasedactivityandmayresultfromionandwatermovement,expansionandcontractionofextracellu-

A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–213

larspaces,capillaryexpansionorneurotransmitterrelease(forreviewseeCohen,1973).Activity-relatedlightscatter-inghasbeenassociatedwithchangesinmembranepotential(Stepnoskietal.,1991)andglialswelling(MacVicarandHochman,1991).

Howdoeseachofthesignalcomponentscontributetothe“mappingsignal”visualizedinthefunctionalmaps,i.e.thestimulus-speciﬁcdifferentialactivationpattern?Thedifferentcomponentsoftheintrinsicsignalhavedifferenttime-coursesandtheirrelativecontributiondependsonthewavelengthusedforillumination.Theincreaseinlightscat-teringreachesitsmaximumwithin2–3sofstimulusonset,whilethedeoxy-hemoglobin(oxymetry)componentpeaksafter4–6s.Theblood-volumerelatedoxy-hemoglobinsig-nalrisesevenmoreslowly,afteraninitialdip,andfollowscloselytheglobalsignal,beginningtodecrease1–3safterstimulusoffset(BonhoefferandGrinvald,1996).Byin-jectingﬂuorescentdyesintothebloodstream,Frostigetal.(1990)demonstratedthatbloodvolumechangesalonecanyieldafunctionalmap.However,themappingsignalisdominatedbyothermechanisms,includingchangesinthecytochromeoxidaseredoxstateandmoreimportantly,ade-creaseinoxygensaturationofhemoglobinduetoincreasedoxygenconsumptionaswellasincreasedlightscatter.Theoxymetrysignalandevenmoresothelightscatteringcom-ponent,whichismoredirectlyrelatedtoelectricalactivity,haveahigherspatialresolution.Therefore,opticalimagingusingnear-infraredwavelengths(700nmandabove)usu-allyprovidesbetterfunctionalmaps(butseeSection3.1.4forauditorycortex)withreducedbloodvesselartifactsde-spitealowerabsolutesignalmagnitude(McLoughlinandBlasdel,1998).Inarecentstudy,Shtoyermanetal.(2000)estimatedtheindividualcontributionsoftheoxy-hemoglobinandthedeoxy-hemoglobinconcentra-tionstofunctionalmapsinawakemonkeysandfoundthatoculardominancedomainsappearsharperinthedeoxy-hemoglobinmaps,conﬁrmingthatthesignalsfromthechangesinconcentrationofdeoxy-hemoglobinco-localizebetterwithelectricalactivitythanthesignalsfromchangesinoxy-hemoglobinconcentration.2.2.Correlationofintrinsicsignalswithmetaboliccostandphysiologicalactivity

Regardlessofthesourcesoftheintrinsicsignals,asec-ondimportantquestioniswhichaspectsofneuronalactivitycontributetothem,andtowhichdegree?Activitycomprisesnotonlythegenerationandpropagationofactionpoten-tialsbutalsothesynaptictransmission,postsynapticpoten-tials,vesicleandreceptorrecycling,etc.Arecenttheoret-icalpaperbyAttwellandLaughlin(2001)hasshedsomelightonthisissue.TheauthorsconsideredmetaboliccostsintermsofATPofglutamatergicsynaptictransmissionandactionpotentialpropagationintherodentcerebralcortexonthebasisofavailableneuroanatomicalandbiophysicaldata.

Fig.1.Relativeenergybudgetofthecerebralcortex.Metaboliccost(inmoleculesofATP)ofrestingpotentials,synaptictransmissionandaction-potentialpropagationisshown(modiﬁedfromAttwellandLaughlin,2001).Ataﬁringrateof4spikes/s,just13%oftheATPcon-sumptionofneuraltissue(neuronsandglialcellsinapproximatelyequalnumbers)isduetomaintainingrestingpotential,whilealmosthalfoftheATPconsumptionisexpendedonaction-potentialpropagation.Theseﬁguresrefertoexcitatory,glutamatergicneurons,whichconstituteabout80%ofallneuronsinthecortex.

AttwellandLaughlin(2001)estimated8therestingcon-sumptionofATPbyneuronsat3.4×10s−1andthatofgliacellsat1.0×108s−1.Atanaverageﬁringrateof4spikes/s,−1apyramidalneuronconsumesanadditional2.8×109ATPs.Justoverhalfofthiscostisattributabletothepropagationoftheactionpotentials,theremaindertosynaptictrans-mission.Thecostofthelatterisdominatedbytheenergyrequiredtorestorepostsynapticiongradients,whichout-weighspresynapticglutamaterecyclingandCa2+pumpingbymorethan10:1(Fig.1).Theseﬁguressuggestthatonlyjustover10%ofneuronalenergyconsumptionareduetomaintainingtherestingstate,andtheoverallcorticalenergyconsumptionismoreorlessproportionaltotheaverageﬁr-ingrate.ATPconsumptioninthecortexmeasuredinvivo(ClarkeandSokoloff,1999)wouldthencorrespondtoanaveragespikerateofabout5–6s−1.Thatthisisclosetoob-servedvaluesconﬁrmsthevalidityofthecalculations.Es-timatesofenergyconsumptionfornon-pyramidalneurons,suchasinhibitory(GABAergic)interneurons,arenotyetavailable.However,assumingthatvaluesaresimilartothoseforpyramidalcells,roughlyhalfofthemetabolizedATPisrequiredatthesynapsesandtheotherhalfforspikepropaga-tion.Thissuggeststhatprocesseswhichdonotresultinac-tionpotentialsbeinggenerated,i.e.subthresholdexcitatoryaswellasinhibitoryinputs,contributeverysigniﬁcantlytooverallmetaboliccostsandthereforepresumablytothein-trinsicsignalsthatformthesubstrateofOI.Sincetheaboveconsiderationspertainprimarilytotheoxymetrycomponentofintrinsicsignals,itisimpossibletoquantifythecontribu-tionofsynapticevents,whenlightscatteringchangesareamajoraspectoftheoverallsignal.

Adirectconsequenceoftherebeingcontributorsotherthanactionpotentialstotheintrinsicsignalisthefactthatthespatialextentofintrinsicsignalselicitedbyasensory

4A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–21

stimulusislargerthantheareaofcortexwhereneuronsre-spondtothatstimuluswithactionpotentials.Inmostcases,theso-called“point-spread”oftheimagingsignal,i.e.theareaofcortexactivatedbyaverysmall(nearpoint-sized)stimulus,islargerthanthepoint-spreadforactionpotentials.Incatvisualcortex,DasandGilbert(1995)foundthattheopticalresponsetoaverynarrowbarstimulus(width,0.1◦)extendedoveracorticalregionofonaverage3.9mmdiame-ter,correspondingto2.25–6◦ofvisualspace(dependingoneccentricity),while◦receptiveﬁelddiametersofsingleneu-ronswere0.3–1wideandspikeswererecordedfromanaveragecorticalarea0.74mmindiameter.DasandGilbert(1995)concludedthattheopticalpoint-spreadisupto20timeslargerinareathanthesuprathresholdneuronalactiv-ity.Boskingandco-workers(2002),inastudyofretinotopyandorientationselectivityintreeshrewvisualcortex,foundasimilarpoint-spreadoftheopticalsignal,as0.25◦widebarspanningover1◦ofvisualspaceelicitedresponsesfromanareaofV1correspondingto4.9◦ofvisualspace.Incon-trasttoresultsreportedbyDasandGilbert(1995)forthecat,however,thiswasonlymarginallywiderthanthewidthofpositionaltuningasdeterminedbymulti-unitrecording,suggestingspeciesdifferences.

2.3.Protocolforanopticalimagingexperiment

ThissectionwillbrieﬂydescribethestandardprotocolforOIexperiments(formoredetails,seeBonhoefferandGrinvald,1996;Grinvaldetal.,1999),andwewillpointoutrecentmethodologicalandtechnicaladvances.

2.3.1.Surgeryandanimalpreparation

Anesthesiaincat,monkeyandferretisgenerallyinducedwithamixtureofketamineandxylazine(BonhoefferandGrinvald,1996;ChapmanandBonhoeffer,1998);pento-barbitoneandurethanehavebeenusedinratexperiments(MasinoandFrostig,1996;MeisterandBonhoeffer,2001;Polleyetal.,1999).Afterinitialanesthesia,theanimalisintubated(chronicexperiments)ortracheotomized(acuteexperiments).ItisthenmountedonastereotaxicapparatusandconnectedtoarespiratorwhichdeliversamixtureofN2OandO2supplementedbyhalothane(orisoﬂurane)asnecessarytomaintainadequateanesthesia.Inimagingstud-iesofauditorycortex,bothpentobarbitoneandthesteroidsaffan(alphaxalone/alphadolone)havebeenusedsuccess-fully,sincehalothanewasfoundtodepressresponses(Versneletal.,2002).Acombinationofketamineandurethanehasbeenusedinmice,whichwillbreathesponta-neouslyandneednotbeartiﬁciallyrespirated(Schuettetal.,2002).

Todate,thequestionwhetherdifferentanestheticsresultinqualitativelyorquantitativelydifferentfunctionalimageshasnotbeenaddressedsystematically.However,subtledif-ferencesbetweenhalothaneandisoﬂuraneintheireffectsonvisualcorticaladaptationhaverecentlybeendescribed(SengpielandBonhoeffer,2002).CO2,ECG,temperatureand,whenneuromuscularblock-ersareused,EEGarecontinuouslymonitoredtoensuread-equateanesthesia.Accesstothecortexcanbeachievedbyopening(e.g.BonhoefferandGrinvald,1996)orthinningtheskull(Boskingetal.,1997;Masinoetal.,1993;MasinoandFrostig,1996;Polleyetal.,1999)abovetheregionofcortextobestudied.Inmice,itisevenpossibletoimagethroughtheexposedbutintactskull(KalatskyandStryker,2003;Schuettetal.,2002).Underfavorableconditions,op-ticalimagingcanprovideactivitymapswithaspatialresolu-tionofupto80–100␮m.Inordertoachievethisresolution,itisimportanttominimizemovementofthebrain,whichnormallyoccursduetoheartbeatandrespiration-relatedpul-sations.Craniotomyisusuallyperformedinlargespecies(i.e.monkey,cat)anditisoftenrequiredtoopentheratheropaquedurainordertogetgoodqualityimages.Oneofthedisadvantagesofperformingadurotomyinchronicexperi-mentsisthepossiblegrowthofopaquetissueontopofthecorticalsurface,whichmakesimagingdifﬁcult.Inaddition,capillaryproliferationmayoccurinthegrowingmembrane,thusincreasingtheriskforhemorrhagewhenresectingit.Anothermajorproblemofdurotomyisthattheexposedcortexbecomesmoresusceptibletoinfectionsevenwhenworkingundersterileconditionsandwhenapplyingtopi-calanti-inﬂammatory(e.g.dexamethasone)andantibioticdrugs.Recently,twodifferentgroups(Arielietal.,2002;Chenetal.,2002)developedatransparentduralsubstituteforlong-termimagingexperiments,whichallowedcorticalimagingforupto1yearafterimplantationwithoutcom-plications.Theartiﬁcialduraiseithermadeoutofsilicone(Arielietal.,2002)orpolyurethane(Chenetal.,2002)andisabout0.1–0.2mmthick.Themainadvantagesofusingtheduralsubstituteinchronicexperimentsare:(1)protectionofthecerebrumagainstinﬂammation,(2)preventionofleak-ageofthecerebrospinalﬂuid,and(3)transparency,whichallowsmaintainingthecortexinagoodopticalconditionforlongperiodsbypreventinggrowthovertheexposedcortex.Inaddition,anelasticduralsubstitutehasprovenusefultoallowmicroelectrodestopassthroughwithoutsufferinganydamage(Arielietal.,2002).

Differenttypesofchambersystemorcranialwindowhavenowbeendevelopedtobothprotectthebrainandminimizemovement.Inlargeranimals(cats,adultferrets,monkeys)achambermadeoftitaniumisused,whichhasaninletandanoutlettowhichtubingisattachedinordertoﬁllthechamberwithsiliconeoil(BonhoefferandGrinvald,1996).Itisthensealedwithaglasscoverslipwhichispressedontoasiliconegasketwithathreadedring.Itismountedontotheskullwithdentalcementandinternalgapsbetweentheskullandthechamberaresealedwithmelteddentalwax.

Recently,ArieliandGrinvald(2002)designedaskull-mounting‘sliding-topcranialwindow’tofacili-tatethecombinationofopticalimagingwithvariousmicroelectrode-basedtechniquesinchronicandacuteex-perimentsinthecerebralcorticesofcatsandmonkeys.Thisassemblyhasallowedgaininginsightsintherelationship

A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–215

betweenneuronalmorphologyofsingleneuronsandfunc-tionalcorticalarchitectureaswellasbetweenthedynamicstateofthecorticalnetworksandthefunctionalresponsetoastimulus.

Alternativelytothechambersystem,ithasbeenpossibletoobtaincorticalfunctionalmapsinferretsandratsthroughalayerofagaroseandaglasscoverslipplacedovertheexposedcortex(ChapmanandBonhoeffer,1998;MeisterandBonhoeffer,2001;Schuettetal.,2001;SchwartzandBonhoeffer,2001)andthroughsalinecontainedinawallofvaselineorthroughagaroseandacoversliptocoverthethinnedbone(Boskingetal.,1997;Polleyetal.,1999).Inmice,thecortexcanbeimagedthroughtheintactskull,oncetheskinhasbeenretracted;transparencyoftheboneismaintainedbyapplyingsiliconeoildirectlytotheskull(Schuettetal.,2002).Thismethodisthereforeideallysuitedforchronicimaging.2.3.2.Dataacquisition

2.3.2.1.Thecamera.Differenttypesofcameras,suchasphotodiodearrays(Grinvaldetal.,1986)andvideocameras(BlasdelandSalama,1986)havebeenusedforfunctionalbrainimaging.Nowadays,mostimagingsystemscontaincharge-coupleddevicetypesensors.PhotonsreﬂectedfromthecortexstriketheCCDfaceplateliberatingelectronsthataccumulateinSiO2“wells”,atarateproportionaltoincidentphotonintensity.Slow-scandigitalCCDcamerashavebeenwidelyusedforintrinsicsignalimaging(Ts’oetal.,1990).Theyprovidegoodsignal-to-noiseratiosatahighspatialresolution,andtheirmaindisadvantage,thelowimageac-quisitionorframerate(<10Hz),isnotcriticalforimagingoftheratherslowintrinsicsignals.Incontrast,videocam-eraswithCCD-typesensorsaremuchfaster(25Hz)andhaveanevenbettersignal-to-noiseratioatthelightlevelstypicalofanOIexperiment.Inthepast,theywereham-peredbyeight-bitframegrabbers,whichcouldnotdigitizeintensitychangesof<1/256(withthetypicalsignalampli-tudeinOIbeingonly∼1/1000).However,thisproblemcanbeovercomebydifferentialsubtractionofastored(analog)referenceimage,resultinginaneffective10-to12-bitdigi-tization.Thisimageenhancementisnolongernecessary,asprecisionvideocameraswith10-bitdigitizationhavebeendeveloped,allowingopticalimagingatupto60Hz.How-ever,forimagingofvoltage-sensitivedyes,muchhigherframeratesaredesirable;camerasofferingupto1700Hzarenowavailable(Shohametal.,1999).

2.3.2.2.Illuminationandﬁlters.Optimalilluminationoftheareaofinterestiscrucialforthequalityofthemaps.Theproperwavelengthsoftheilluminatinglightdependonthesourcesofintrinsicsignalsofinterest(seebelow).Oneshouldalsobearinmindthatlightoflongerwavelengthswillpenetratedeeperintothetissue.

Evenilluminationisbestachievedbyusingatleasttwoﬁber-opticlightguidesdirectedattheregionofinterest,

whereasahighqualityregulateddcpowersupplyisessentialforguaranteeingastablelightintensity.

Band-passinterferenceﬁltersareusedtolimitthewave-lengthoftheilluminatinglight.Themostfrequentlyusedﬁltersare:(1)greenﬁlter,546nm(30nmwide)—bestforobtainingthebloodvessel/surfacepicture;(2)orangeﬁlter,605nm(5–15nmwide)—atthiswavelengththeoxymetrycomponentdominatesthesignal;(3)redﬁlter,630nm(30nmwide)—atthiswavelengththeintrinsicsignalisdominatedbychangesinbloodvolumeandtheoxygena-tionsaturationlevelofhemoglobin;(4)nearinfraredﬁlters,700–850nm(30nmwide)—atthesewavelengths,thelightscatteringcomponentdominatestheintrinsicsignal,whilethecontributionofhemoglobinsignalsismuchreduced(Bloodetal.,1995;Frostigetal.,1990;Narayanetal.,1995).Opticalimagingcanthusbeusedtomapdifferentphysiologicalprocessesdependingonthespeciﬁcwave-lengthchosenforillumination.

Analternativetotheuseoflightguides(incombina-tionwithband-passﬁlters)istheilluminationbyaringoflight-emittingdiodes(LEDs)ofspeciﬁcwavelengths(e.g.Mayhewetal.,1996).

2.3.2.3.Timingofdataacquisition.AnumberofmajorbiologicalsignalsourcesinOIarenotassociatedwithneu-ralactivitybutwithrespiratoryandcardiovasculareventsandaretherefore“noise”.Themostprominentonesaretheso-calledvasomotionsignalaswellasheartbeatandventila-tionartifactsattheirrespectivefundamentalfrequenciesandharmonics(seeKalatskyandStryker,2003).Thevasomo-tionsignalrepresentsalowfrequency(peakingnear0.1Hz)oscillationresultingfrommodulationsofregionalcerebralbloodﬂow(Mayhewetal.,1996).Althoughheartbeatandrespirationartifactsoccuratfrequencieshigherthanvaso-motion,theyallfallwithinthecategoryof“slow”noiseofaperiodicitythatiswithinoneortwoordersofmagnitudeofthetime-courseoftheneuralsignals.

Inordertominimizetheireffectsonfunctionalmaps,itisbeneﬁcialtosynchronizeheartbeatandrespirationwithimageacquisition.Thisisachievedbytriggeringboththerespiratorandtheimageacquisitionofftheheartbeat,amea-surethatcanreduceslownoisebyuptoafactorof1.5(Grinvaldetal.,1991).Ofcourse,accumulationofalargenumberoftrialsshouldaverageoutperiodicartifacts.Al-ternatively,temporalﬁlteringoftheintrinsicsignalmaybeemployedtoremoveperiodicartifacts(seeSection2.3.3).Underoptimalconditions,albeitnoisyfunctionalmapscanbeobtainedinasingletrial.Generally,between20and100responsestoanystimulusconditionareaveragedinor-dertoimprovethesignal-to-noiseratio.Sincethespatiallo-calizationtothesiteofneuralactivityisbetterfortheearlyintrinsicsignalcomponents(oxymetry,lightscattering)thanthelatercomponents(bloodﬂow),stimuliaretypicallylim-itedto3–4sduration,whiledataacquisitionmaybeslightlylonger,dependingonthetime-courseofthemappingsignal.Inter-stimulusintervalsmayalsovary.However,aminimum

6A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–21

ofabout7–8sisrequiredfor“metabolicrelaxation”,i.e.recoverytonearbaseline(BonhoefferandGrinvald,1996).Ontheotherhand,inter-stimulusintervalshouldnotbetoolonginordertomaximizethenumberofimagescollectedandtoavoidsystematicerrorsresultingfromslowdriftsinthebaselinestateofthecortex.Frequently,theendoftheinter-stimulusintervalisutilizedtorecordoneormore“ﬁrstframes”,themeanofwhichmaybesubtractedfromsubse-quentframescollectedduringstimuluspresentationinor-dertocorrectforrelativelytime-invariantbiologicalnoise(BonhoefferandGrinvald,1996).

2.3.2.4.Basicexperimentalsetup.Oncetheanimalisanesthetized,itisheldinastereotaxicframe.Thebrainisilluminatedwithlightoftheappropriatewavelengthandimagesareacquiredbythecamerapositionedabovethecor-tex.Thecameramustbeﬁrmlymountedinavibration-freedevice.Thebestarrangementshouldhavethecameraat-tachedtoaholderthatallowstiltingthecameratoanydesiredangle.Thecameraholdershouldhave,preferably,anxyz-translatorforﬁnepositioningandfocusing.Itisadvisabletofocusonthecorticalsurfaceﬁrstinordertochoosearegionofinterestandcaptureapictureofthebloodvessels,whichcanlaterbeusedtorelateactivitymapswithanatomicallandmarks.Lensapertureshouldbereducedduringtherecordingofthebloodvesselpatterntoavoidblurringalongtheedgesoftheimageduetothecurvatureofthecortex.Forrecordingactivitymaps,thecamerashouldbefocused300–700␮mbelowthecorti-calsurface,andthelensaperturesshouldbefullyopen;a“macroscope”assemblyprovidinghighnumericalapertureisidealformaximallightyieldandashallowdepthoffocus(RatzlaffandGrinvald,1991).

2.3.3.Dataanalysis

Inthelivingbrain,intrinsicsignalsareverysmall.Changeinlightintensityat605nmduetoneuronalactivityisatbestabout0.5%ofthetotalintensityofthereﬂectedlight

(andtypicallyunder0.1%).Thus,intrinsicsignalsarenotapparentbuthavetobeextractedfromtheimageswiththeappropriateanalysisprocedures(BonhoefferandGrinvald,1996).

Likeallfunctionalimagingtechniques,OImapsthedifferencesfoundinacertainbrainregionbetweenthebasalactivitylevelunderarestingorcontrolconditionandanactivatedstatefollowingaspeciﬁcstimulus.Thechoiceofbaselineconditionandthemethodofextract-ingthestimulus-inducedsignalfrombiologicalaswellasshotnoise(thestochasticﬂuctuationsoflightemission)arethereforecriticalformeaningfuldatainterpretation,sincethemeasurementofabsolutesignalstrengthisnotpossible.2.3.3.1.Event-relatedimaging.Thestandardformofstimuluspresentationandimageanalysisisoneofevent-relatedimaging.Inotherwords,foreachstimulus,thechangeofthereﬂectancesignaloftheindividualpixelsintheimageisrecordedduringand/orafterthepresenta-tionofthestimulus,andthesignalsareaveragedoveranumberoftrials.Thisprocedureisequivalenttothewayperi-stimulustimehistogramsareobtainedforsingle-cellresponses.Typically,absorptionbeginstoincreaseabout0.5safterstimulusonsetandreachesamaximumafter3–4s(seeFig.2F).

Whereasinsingle-cellrecordingsthespontaneousactiv-ityintheabsenceofexplicitstimulationprovidesasimplebaseline,asimilarcontrolconditionforintrinsic-signalim-agesisfarhardertodeﬁnebecauseoftheratherindirectrela-tionbetweenneuronalactivityandthemappingsignal.Theequivalentof“spontaneousactivity”isthe“blankimage”obtainedfromtheunstimulatedcortex.Thedifﬁcultywithusingthe“blank”asacontrolarisesfromthefactthatmostadequatestimuli(e.g.visualstimuliincaseoftheprimaryvisualcortex)willcauseanoverallelevationinabsorption.Thisglobalresponseincludesregionswhereatthelevelofneuronal(spiking)activitythereisnoresponseatalltothestimulus(seeFig.2F).

Fig.2.Orientationpreferencemapsobtainedwithevent-relatedopticalimaging.(A)Iso-orientationmapsofcatarea17obtainedbydividingsingle-orientationresponsesbyblank-screenresponses.Stimulusorientationisindicatedbythecoloredbarinthetopleftcornerofeachimage.Theim-ageshavenotbeenﬁltered.Theyhavebeenrange-ﬁttedidentically,withapixelvalueof1.0(responsetogratingequaltoresponsetoblank)representedbyagray-scalevalueof255(white)andavalueof0.988(orlower)representedbyagray-scalevalueof0(black).Theoverallsignalstrengthisthereforeapproximately1.2%.Notethateachorientationcausesaglobalresponsefromtheentireimagedregionofvisualcortex.(B)Iso-orientationmapsobtainedbydividingsingle-orientationresponsesbythe“cocktailblank”(seetext).Theimageshavebeenhigh-passﬁltered(ﬁlterwidth,80pixels=1.7mm)andrangeﬁtted(pixelvaluesof0and255,respectively,signify±0.2%signalchangecomparedwiththecocktailblank).(C)Bloodvesselpatternoftheimagedregionofcortex,imagedwithgreenilluminatinglightthroughtheintactdura.Area17ofbothhemispheresisvisible;thearrowsindicatetheorientation(a,anterior;p,posterior).Scalebar,1mm.(D)Anglemapoforientationpreference,obtainedbyvectorialadditionofthemapsshownin(B)andsubsequentlow-passﬁltering.Thevectorangle(preferredorientation)ofeachpixelisencodedashue,asindicatedbythecolorlegendbelow.(E)Polarmapoforientationpreference,obtainedbyvectorialadditionofthemapsshownin(B)andsubsequentlow-passﬁltering.Thevectorangle(preferredorientation)ofeachpixelisencodedashue,asindicatedbythecolorlegendbelow(D),whilethevectorlengthisencodedasbrightness.(F)Timecourseofreﬂectionsignalmeasuredbeforeandduringpresentationofahorizontal(0◦)grating(stimulusduration=4.2s,asindicatedbyhorizontalbarbelowtimeaxis).Sixblocksoffourtrialseachwererecorded,usingImager2001(OpticalImagingInc.)with2×2pixelbinning.Eachtrialcontainedfourstimuliofeachofthefourorientations,0,45,90and135◦.Withintheactivatedregionofcortex,weﬁrstselectedthose25%ofpixelsrespondingmoststronglytogratingsof0◦orientation.Weaveragedtherawvaluesonﬁleforthesepixelsacrossthefour0◦stimuliandthenacrossthesixdatablocks;thevalueobtainedfortheﬁrstframehasbeenarbitrarilysettozero(blackcurve).Thesamecalculationwasrepeatedforthose25%ofpixelsrespondingmoststronglytogratingsof90◦(redcurve).ErrorbarsrepresentS.E.M.sacrossblocks.Notethemagnitudeofthenon-orientationselectiveresponsecomponent.(Actualsignalsrepresentreﬂectanceandchangesarenegative,butfordisplaypurposeschangesareshownaspositive.)

A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–217

Inordertoremovethedcresponsefromtheimagesandtoextractonlythestimulusselectiveresponsecomponents,imagescanalternativelybedividedbythe“cocktailblank”,whichisthesumoftheresponsestoallstimuliinaset(BonhoefferandGrinvald,1993).Forthisproceduretohavevalidresults,twoimportantconditionsmustbemet.First,thestimuliusedmustcoverthestimulusspaceevenly.Forexample,inasetoforientedgratings,theorientationsmustcovertherangeof180◦instepsofequalsize.Theuseofonlyasubsetofstimulicansigniﬁcantlyaffectthemapob-tained(Issaetal.,2000).Second,thesumoftheresponsestoallstimulimustbeuniformacrosstheimagedregion.Ifitispatchy,divisionbythe“cocktailblank”willresultinapatchymapevenforastimuluswhichitselfelicitednoresponseatall(BonhoefferandGrinvald,1993,1996).Inthecaseof“orthogonal”stimulithatelicitresponsesfromlargelynon-overlappingpopulationsofneurons,differenceimagespresentanalternativetothecocktail-blankproce-dure.Examplesarethesubtraction(ordivision:forverysmalldifferences,asisthecaseinOI,theresultsarein-terchangeable,seeBonhoefferandGrinvald,1996)ofre-sponsestohorizontalversusverticalgratingsorleft-eyever-

8A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–21

susright-eyestimulationinordertoobtainorientationandoculardominancemaps,respectively.

Inadditiontodivisionbythe“blank”controlimage(orasanalternativetoit),theso-called“ﬁrst-frame”subtractionhasproveduseful.Inthisapproach,thereﬂectanceimageisrecordedforafewframespriortotheactualstimulusonset,andissubtractedfromthepoststimulusframes(BonhoefferandGrinvald,1996).Thismethodverysuccessfullyremovesartifactsthataremoreorlesstime-invariantwithinthedu-rationofthestimuli.

Fig.2showsorientationpreferencemapsofcatprimaryvisualcortex.Fig.2Arepresentsingleconditionactivitymapsforfourdifferentorientationsinbothcorticalhemi-spheres(B).Theseiso-orientationmapseachrepresentthesummedresponsesto96presentationsofthesamestimu-lusorientation;theyweredividedbythe“blank”response(theresponsetoauniformgreyscreen).InFig.2C,thesamesummedresponsesweredividedbythe“cocktailblank”(BonhoefferandGrinvald,1993).Thedifferencesbetweenthemoreglobalizedresponsesin(A)andthemoreorientation-speciﬁcresponsesin(C)areevident.Theycanbequantiﬁedbycalculatingtwo-dimensionalcorrelationcoefﬁcientsbetweenmapsobtainedwithorthogonalorien-tations:thesetendtobepositiveincaseofblankdividedimagesandnegativeincaseofcocktail-blankdividedim-ages.Fig.2Drepresentstheorientationpreference“anglemap”obtainedbypixel-by-pixelvectorialaddition(BlasdelandSalama,1986)ofthesingleconditionmapsshownin(C).Thecolorsintheimagecodefortheangleofthepreferredgratingorientation.Additionalinformationmaybeprovidedbydisplayingthemagnitudeoftheresultingvectorasbrightness.Theresulting“polarmap”(Ts’oetal.,1990)showsthepreferredorientationascolorhueandthemagnitudeofthevectorasbrightness(Fig.2E).

Analternativemethodtoderivesecondaryparametermaps(suchasthe“anglemap”oforientationpreferencementionedabove)ispixel-by-pixelanalysisofresponsestoasetofstimuli.Heretheresponsesofeachimagepixelaretreatedmuchinthesamewayasresponsesofasingleneurontoasetofstimuli,andtuningfunctionspreviouslydescribedforsingleneuronscanbeﬁttedtothosepixelresponses.Forexample,Swindaleetal.(2003)ﬁttedcircu-larnormalfunctionstopixelresponsesforasetofgratingstimulivaryinginorientationanddirectionofdriftinordertocalculatemapsoforientationanddirectionpreferenceaswellastuningwidth,revealingspatialrelationshipsbe-tweentheseparametersnotevidentwithtraditionalvectoraveraging.Similarly,wehaverecentlydetermined,onapixel-by-pixelbasis,contrast-responsefunctionsandorien-tationtuningcurvesincatV1,usinghyperbolicratioandGaussianfunctions,respectively(CarandiniandSengpiel,2004).Wefoundthattheﬁtparametersdescribingcontrastresponsesweremoreorlessuniformoverthecorticalsur-face.Moreover,stimuluscontrasthadnoimpactonmapsoforientationpreference,theorientationselectivityofeachpixel,justasthatofsingleneurons,wascontrast-invariant.

2.3.3.2.Principalcomponentanalysis(PCA)andrelatedmethods.Asdescribedearlier,intrinsicopticalsignalscon-sistofanumberofcomponents,bothstimulus-relatedandstimulus-independent,thatexhibitdistinctspatialandtem-poralpatterns.Principalcomponentanalysis(PCA)isusedtodecorrelatesignalsofdifferentoriginsinalinearmix-tureofsignals,whichareassumedtobeorthogonal,andtoﬁnddirectionsofexternalvarianceinthedataspace.SignalrecoveryfromOIdatabymeansofspatialPCAwasdevel-opedbySirovichandEverson(1992)andimprovedfurtherbycombiningitwiththestandarddifferenceimagemethod(Gabbayetal.,2000).PCAovertime(ratherthanspace)ofcorticalimagesacquiredintheabsenceofanystimulusandsubsequentselectionofcomponentsthataremoststronglycorrelatedwiththesurfacevasculaturepatternallowsre-movalofbloodvesselartifactsfromimagesbylinearextrac-tion(Schuettetal.,2000).AnexampleoftheeffectivenessofthisprocedureisillustratedinFig.3.

Blindsourceseparation(BSS)describesagroupofsignal-processingtechniquesthatcanberegardedasex-tensionsofPCA,makingadditionalassumptionsaboutthestatisticalstructureofthesignalsourcesinordertorecoverthemfromthemixtures.Independentcomponentanaly-sis(ICA;BellandSejnowski,1995)andextendedspatialdecorrelation(ESD;Stetteretal.,2000)assumethatthedifferentsourcesarestatisticallyindependentandmutuallyuncorrelated,respectively.Thesemethodsimprovetheex-tractionofthestimulus-relatedspatialsignalfromOIdata.2.3.3.3.Periodicstimulationimaging.Thisrecentlyde-velopedapproach(KalatskyandStryker,2003)iscompa-rabletothestandardmodeofimageacquisitioninfMRI(Boyntonetal.,1996;Engeletal.,1994).Insteadofmea-suringaresponsefollowingeachindividualstimulus,stim-uliarepresentedinperiodicfashionoveralongerperiodoftime.Thiskindofstimulationresultsinaperiodicallymod-ulatedreﬂectancesignalforeachpixelintheimage,whichcanbedecomposedintosinewavesofdifferentfrequenciesusingFourieranalysis(Fig.4).Theonlyfrequencyofinter-estisthatcorrespondingtothestimuluspresentation,whilethoserelatinge.g.totheheartandrespirationrateandtova-somotioncanbeﬁlteredout.Intheory,then,theamplitudeandphaseofthepixelresponseovertimeatthestimulusfrequencycanbeusedtodetermineresponsestrengthandstimuluspreference,respectively.Oneobviousadvantageofthisparadigmisthefactthatabsoluteresponselevelsplaynopartintheanalysis,asonlyrelativeresponsemodulationisassessed.Moreover,responsecomponentswhoseperiod-icitydoesnotmatchthatofthestimulus(suchasheartbeatandrespirationartifacts)canberemovedeasily.Finally,datacaninprinciplebeacquiredinamuchshorterperiodoftimethanispossibleinevent-relatedimaging(KalatskyandStryker,2003).However,therearesomecaveatstoo.First,signalcomponentswhosefrequencyisveryclosetothatofthestimuluscannotbeﬁlteredoutbutwillcontaminatetheresults.Thefrequencyofstimulationshouldthereforebedif-

A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–219

Fig.3.Removalofbloodvesselartifactsbymeansoflinearextractioncombinedwithprincipalcomponentanalysis:(A)oculardominancemapobtainedfromkittenV1byrecordingresponsestodriftinggratingsof0,45,90and135◦throughleftandrighteyeseparatelyanddividingthesummedresponses.Ahugebloodvesselartifactcanbeseenintheoccipitalpartoftheimage.(B)Surfacebloodvesselpatternrecordedwithgreen(546nm)illuminatinglight,showingclearlytheveinthatcausedthelargeartifactin(A).(C)Followingprincipal-componentanalysisofimagesobtainedwhiletheanimalviewedablankscreen,componentsthatshowedthehighestspatialcorrelationwiththebloodvesselpatternwereselectedandextractedfromimagesobtainedinthepresenceofgratingstimuli(Schuettetal.,2000).

ferentfromthatthemajorhemodynamiccomponents.Still,evenafterremovalofslowchangesinimageintensity,ananalysisofresponseamplitudeversusphasewilloftenshowthatresponsephasesarenotrepresentedequally(aswouldbeexpectedforastimulusset,suchasorientation).Sec-ond,anunknownlagtime(hemodynamicdelay)betweenstimulusandintrinsicsignalresponsemeansthatresponsephasescannotbetranslateddirectlyintoanabsolutestimu-luspreferencemap.Thiscanbeovercome(attheexpenseofdoublingthedataacquisitiontime)bycyclingthroughthestimulussetbothinascendinganddescendingorderorbymeasuringthehemodynamicdelayseparatelyusingjustasinglestimulus(KalatskyandStryker,2003).Itisimpor-tanttokeepinmindthattheformermethodwillonlyyieldvalidresultsiftheorderofstimuluspresentationdoesnotitselfaffectresponses.Third,ifthestimuliinaperiodicsetarenotequallyefﬁcaciousatdrivingcorticalresponses,thestimulus-phaserelationshipmaynotbestraightforward,asthephaselagmaynotbethesameforallstimuli.Forexam-ple,followingabriefperiodofmonoculardeprivation,wefoundthatresponsesthroughthetwoeyesinkittenV1toalternatingstimulation,usingacontrast-reversingchecker-board,werenotpreciselyinanti-phase,aswouldhavebeenexpected(F.Sengpiel,unpublishedobservation),presum-ablybecauseoflatencydifferencesbetweenthedeprivedandthenon-deprivedeye.TheadditionalpossibilityofresponsedelaydifferencesatmapedgesisdiscussedbyMrsic-Flogelandco-workers(Mrsic-Flogeletal.,2003).

catandmonkeyvisualcortex(Frostigetal.,1990;Grinvaldetal.,1986;Ts’oetal.,1986,1990).However,nowadays,opticalimaginghasbecomeanimportanttoolfor(i)study-ingthefunctionalarchitectureofmotor,somatosensory,au-ditorycorticesandtheolfactorybulb,(ii)assessingcorti-calmapsinawakeanimals,and(iii)investigatingfunctionalcorticaldevelopmentandplasticityundernormalandpatho-logicalconditionsandfollowingenvironmentalmanipula-tions.Lately,thetechniquehasalsobeenusedtovisualizethespreadoffocalepilepticseizuresandthereorganizationoffunctionalcorticalmapsinthesurroundingofafocalis-chemicinjury,andithasbeenadaptedtoimagethehumancortexintra-operatively.Inthissectionwewilldiscusssomeofthemorerecentinsightsintothefunctionalorganizationofthebraingainedbymeansofopticalimagingofintrinsicsignals.

3.1.Acuteexperimentsinsensorycortices

3.1.1.StudiesonfunctionalarchitectureofvisualcortexThefunctionalarchitectureofvisualcortexhadbeenex-tensivelystudiedlongbeforeopticalimagingwasdeveloped.Usingelectrophysiologicaltechniques,HubelandWiesel(1962)ﬁrstreportedtheexistenceoforientationpreferenceandoculardominancecolumns,whichwereconﬁrmedus-ingtransneuronallabelingand2-deoxyglucose(2DG)map-pingtechniques(Singer,1981;Singeretal.,1981).Payneetal.(1981)andTolhurstetal.(1981)describedclusteringofcellsaccordingtopreferreddirectionofmotion;andus-ingthe2DGtechnique,Tootelletal.(1981)describedspa-tialfrequencycolumns.

Opticalimagingofintrinsicsignalshasprovidedapow-erfultoolforestablishingthepreciselayoutandtheinter-relationshipoftheaforementionedcorticalfeaturerepre-

3.ApplicationsofopticalimagingofintrinsicsignalsWhenintrinsicopticalimagingwasﬁrstdeveloped,ithelpedunderstandingthedetailedfunctionalarchitectureof

10A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–21

sentations(BonhoefferandGrinvald,1991;Grinvaldetal.,1986,1991;Hübeneretal.,1997;Ts’oetal.,1990;Whiteetal.,2001)whichhadremainedelusiveusingclassicaltechniques.Themainadvantageofopticalimaginginthatrespectisthatitallowstovisualizemapsforalltheabovementionedmapsinthesamesubjectatthesameimagingtime,whichinturnfacilitatesestablishinggeometricrela-

tionshipsbetweendifferentcolumnarsystems(BartfeldandGrinvald,1992;Boskingetal.,2002;Hübeneretal.,1997;Kimetal.,1999;LandismanandTs’o,2002;ShmuelandGrinvald,1996;Welikyetal.,1996).Importantly,Hübenerandco-workers(1997)reportedthatmostcolumnarsystemstendtointersecteachotheratrightanglesmorefrequentlythanwouldbeexpectedinarandomarrangement.

Knowledgeoftheserelationshipshasenabledresearcherstotestandvalidatethehypothesisofcoverageoptimiza-tion(Swindale,2000;Swindaleetal.,2000),whichisattheheartofthe“ice-cubemodel”(HubelandWiesel,1977)oftheprimaryvisualcortex.Interestingly,nolocalspatialrelationshipappearstoexistbetweenretinotopicandorien-tationpreferencemaps,atleastnotintreeshrewV1,whilecoverageuniformityismaintained(Boskingetal.,2002).Analternativeinterpretationofhowmultiplefeaturesmayberepresentedinthevisualcortexhasrecentlybeenputforward.Basoleandco-workers(2003)suggestthatratherthanreﬂectingtheintersectionofmultiplemaps,populationactivitymaybebetterdescribedbyasingle,spatiotemporalenergymap.

Thedevelopmentofopticalimaginghasmadeitpossibletovisualizenotonlythelayoutoforientationpreferenceandoculardominancemapswithinarea17and18inanum-berofspeciesincludingcat,ferret,macaque,treeshrew,barnowlandmarmosetmonkey(Bonhoefferetal.,1995;

Fig.4.Orientationselectiveresponsesobtainedwithperiodicstimulation.(A)Standardiso-orientationmapsobtainedfromtreeshrewarea17,us-inghorizontalandverticalgratingsrespectively(seeiconsbelowmaps).Scalebar,1mm.(B)Timecourseofrawsignalduringperiodicstimula-tion.Thestimuluswasadriftinggratingwhoseorientationadvancedby22.5◦every0.5s,suchthata180◦cyclewascompletedevery4s.Thereﬂectancesignalwassummedupover25%ofpixelsthatrespondedbesttohorizontalgratingsinaregionofinterestdeﬁnedonthebasisofresponsestostandardstimulation(seeA).Thebarabovetheabscissaindicateswhen,duringthecontinuousperiodicstimulation,ahorizontalgratingwaspresent.(C)Timecourseoftemporallyandspatiallysmoothedsignal.Temporalsmoothingwasbysubtractionofboxcaraverageofsig-nalacrossonestimulusperiod,spatialsmoothingbysubtractionofboxcarsignalaverageovera200-by-200-pixelarea(pixelwidth,21.2␮m).Notethatactivityofpixelsrespondingbesttohorizontalgratings(blackcurve)isinanti-phasewithactivityofthosepixelsrespondingbesttoverticalgratings(redcurve).Forgreaterclarity,onlytheﬁrst50sareshownatanexpandedtime-scalecomparedwith(B).Thebarsabovetheabscissain-dicatewhen,duringtheperiodicstimulation,ahorizontalgrating(black)oraverticalgrating(red)waspresent.(D)Powerspectrumofsmoothedsignal(black,pixelsrespondingbesttohorizontalgratings;red,pixelsrespondingbesttoverticalgratings).Notethepeakofeachspectrumat0.25Hz,correspondingtothestimuluscyclingperiodof4s.(E)Orienta-tionpreferencemapsobtainedwith2hofstandardevent-relatedimaging(left)and20minofperiodicstimulation(right).Thestandardorientationpreferencemapiscalculatedbyvectorialadditionofiso-orientationmapsinresponsetogratingsof0,45,90and135◦;theresultingvectorangleisplotted(seecolorcode).Theperiodicstimulationmapplotsthephaseangleofeachpixel’sresponseatthefrequencyofstimulation(0.25Hz).Theapparentdifferenceinpreferredorientationbetweenthetwomapsofabout45◦correspondstothehemodynamicdelayofcorticalresponsesduringperiodicstimulation.

A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–2111

Boskingetal.,1997;Grinvaldetal.,1991;Issaetal.,1999;LiuandPettigrew,2003;ShmuelandGrinvald,2000;Shtoyermanetal.,2000)butalsomapsofdirectionofmo-tionpreference(ShmuelandGrinvald,1996;Welikyetal.,1996),spatialfrequencypreference(Shohametal.,1997;Issaetal.,2000),formprocessingmodulesinmacaqueareaV4(GhoseandTs’o,1997),clustersofcolorselectiveneu-ronsinV1(LandismanandTs’o,2002)andahue-selectiveorganizationwithinthethincytochromeoxidasestripesofV2(Xiaoetal.,2003)havebeendescribed.Morere-cently,retinotopicmapsinmacaquemonkeyV1(BlasdelandCampbell,2001),owlmonkeyV3(Lyonetal.,2002),mouse(KalatskyandStryker,2003;Schuettetal.,2002),treeshrew(Boskingetal.,2000)andcat(Zepedaetal.,2003)havealsobeenreported.Furthermore,opticalimag-inghasbeenemployedtorevealthefunctionalarchitectureofowlmonkeyareaMT(Maloneketal.,1994),toresolvetherelationshipbetweencolumnarsystemsinareaV2ofthesquirrelmonkey(Malachetal.,1994)andtodemonstratethedistributedprocessingofobjectfeaturesinmacaqueinferotemporalcortex(Tsunodaetal.,2001;Wangetal.,1996).

Bycombiningopticalimagingwithothertechniques,ithasbeenpossibletorevealpropertiesofindividualneuronsatidentiﬁedlocationswithinthemapsandtodescribecorti-calcharacteristicsatanevenﬁnerspatialscale(e.g.Malachetal.,1993;Maldonadoetal.,1997;Ts’oetal.,2001;Yousefetal.,1999).Togetherwithanterogradeandretrogradela-belingtechniques,opticalimaginghasprovideddetailedin-formationregardingtheanatomicalunderpinningsoffunc-tionalmaps,thussheddinglightonhowexcitatoryhorizon-tal(Boskingetal.,1997;Malachetal.,1997)andlateralinhibitory(Kisvardayetal.,1994,1997)aswellascallosalconnections(Boskingetal.,2000)contributetoneuronalresponseproperties.

3.1.2.Studiesonfunctionalarchitectureofsomatosensorycortex

Amongsomatosensoryareas,opticalimaginghasbeenappliedmainlytomappingoftheprimarysomatosensorycortexofrodents(S1)andmonkeys(S-I).

Intherat,S1isdominatedbytherepresentationoffacialwhiskersindiscretecytoarchitectonicunitsknownasbar-rels,ﬁrstdescribedbyWoolseyandVanderLoos(1970).Throughopticalimagingofintrinsicsignals,ithasbeenpossibletoshowthefunctionalrepresentationofindivid-ualwhiskersinratandgerbilbarrelcortex(Bloodetal.,1995;Brett-Greenetal.,2001;Grinvaldetal.,1986;Hessetal.,2000;Masinoetal.,1993;Narayanetal.,1994a,1994b,1995;PetersonandGoldreich,1998;Polleyetal.,1999),andithasbeenpossibletoresolvethearealex-tentandpoint-spreadofsinglewhiskerrepresentationsinprimarysomatosensorycortexofrats(Brett-Greenetal.,2001;Chen-Bee,1996;Chen-Beeetal.,1996;MasinoandFrostig,1996;Shethetal.,2003).Opticalimagingresultsareingoodagreementwithfunctionalmapsobtainedus-ingvoltage-sensitivedyes(Takashimaetal.,2001).How-ever,signalsfromwhiskerstimulationobtainedthroughop-ticalimagingareoftenlargerthanexpectedwhencomparedtoelectrophysiologicalmapping(Brett-Greenetal.,2001;Narayanetal.,1994b).Thisdivergenceofactivitymaybeabasicfunctionalfeatureofthewhisker-to-barrelprojec-tion(Brett-Greenetal.,2001).Thelargearealextentofthefunctionalrepresentationofsinglewhiskersobtainedthroughoptimalimagingmayresultfromhorizontalactivityspreadthroughexcitatoryconnectionsinlayers2/3,whichincreasesinextentwiththedegreeofwhiskerdeﬂection,asrevealedbyvoltage-sensitivedyeimaging(Petersenetal.,2003).However,theapparentlyverylargesingle-whiskeractivationareasobtainedintheaboveOIstudiesmayalsobeaconsequenceofthefactthatimagesoftheunstimu-latedstateofthebarrelcortexweresubtractedfromimagesinresponsetosingle-whiskerstimulation,ratherthanim-agesfrom“orthogonal”stimulusconditions,asiscommonlydoneinimagingstudiesofthevisualcortex(e.g.horizontalversusverticalgratingsorleft-eyeversusrighteyestimu-lation)inordertoenhancemapcontrastanddomaindelin-eation(Frostigetal.,1990).Itiscertainlyproblematicthatresultsvarysigniﬁcantlywiththemethodofanalysisused(SchulzeandFox,2000).

Insomatosensorycortex,thetopographicmapofthebodysurfacehasbeenwellestablishedusingelectrophysiologi-caltechniques(Nelsonetal.,1980;Ponsetal.,1985,1987;Woolseyetal.,1942).Anumberofgroupshavestudiedthecorticalsomatosensorytopographicmapofrat,cat,squirrelmonkey,macaqueandhumanusingopticalimaging(Chenetal.,2001;Gochinetal.,1992;Goddeetal.,1995;ShohamandGrinvald,2001;Tommerdahletal.,1993,1996,1998,1999a,1999b).Resultshaverevealedthat,consistentwithelectrophysiologicalobservations,somatotopicrepresenta-tionoftheﬁngerpadsexhibitsanorderlymedialtolat-eralprogressionfromD5toD1ﬁngerpads(Nelsonetal.,1980;Ponsetal.,1987;Suretal.,1982).However,asforthevisualcortex,electrophysiologicalmethodsdonotallowtoresolvewhetherdifferenttactilefeaturesformmultiplefunctionaldomainswithintheprimarysensorycortex.Inanattempttorevealtheorganizationofresponseofdifferentsensorystimuliinsomatosensorycortexofcatandsquir-relmonkey,Tommerdahletal.(1993,1996,1998,1999a,1999b)haveaddressedS-Icorticalresponsestocutaneousﬂutter,vibration,tapping,andskinheatingwhileChenetal.(2001)haveadditionallyaddressedthecorticalrepresenta-tionofpressure.Resultsfromthesegroupshaveshownthatarea3aintheanteriorparietalcortexhasaleadingroleintheprocessingofskin-heatingstimuli(Tommerdahletal.,1996),andthathigh-frequency(200Hz)vibrotactilestimuliactivateneuronsincorticalregionsotherthanareas3band1(Tommerdahletal.,1999a)whereasinarea3bthesensationofpressure,ﬂutterandvibrationpreferentiallyactivateonereceptorpopulationevenwhenfunctionalcorticalrepresen-tationsforeachsensationoverlap(Chenetal.,2001).Thus,inaccordancewithapreviousstudybyTommerdahletal.

12A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–21

(1993),whichproposedthatactivationofamini-columninS-Iencodesinformationaboutthephysicalpropertiesoftactilestimuli,theauthorssuggestthatinitialcorticalpro-cessingcouldinvolvetheseparationofsensoryinforma-tionintodistinctfunctionalmaps.Interestingly,theyreportthatunderbarbiturateanesthesia,thefunctionalactivationoftheﬁngerpadsforallsensationswerediscreteandexhib-itedminimaloverlapbetweenthem.However,underisoﬂu-raneanesthesia,therepresentationofﬁngerpadsonadja-centﬁngershadahigherdegreeofoverlapthanwithPen-tothalanesthesiaeventhoughthegeneraltopographywasstillmaintained(Chenetal.,2001).Asforbarrelcortexstudies,noorthogonalstimulusconditionsareavailableforanalysis(seeabove).Thus,alternativeanalyses,suchastheﬁrst-frameorblanksubtractionandsubtractingthesumofimagesobtainedforonestimuluscondition(e.g.pressure)fromthatobtainedunderanotherhavebeenusedfordataanalysis.

3.1.3.StudiesonfunctionalarchitectureofolfactorybulbOlfactorysensoryneuronsthatexpressagivenodorantreceptorarewidelydistributedwithintheolfactoryepithe-lium.Theolfactoryepitheliumprojectstotheolfactorybulbintheforebrain,whereaxonsfromolfactorysensoryneu-ronsexpressingthesameodorantreceptorconvergeontosingleglomeruli(Mombaerts,1999).Ithasbeensuggestedthatglomeruliarefunctionalunitsinolfactoryprocessing(HildebrandandShepherd,1997).Inrecentyears,thecen-tralorganizationofodorantrepresentationhasreceivedpar-ticularattention.However,untilthedevelopmentofimagingtechniques,therelationshipbetweenthemolecularbiologyofodorantreceptorsandthefunctionalorganizationoftheolfactorysystemremainedpoorlyunderstood(forreview,seeBozzaandMombaerts,2001).

Usingelectrophysiologicaltechniques,2-deoxygluocoseautoradiography,c-fosexpressionandfunctionalmagneticresonanceimaging,anumberofgroupshaveprovidedin-sightsintotheorganizationofgroupsofglomeruliwithre-specttoodormolecules(Guthrieetal.,1993;Imamuraetal.,1992;Johnsonetal.,1998;Sharpetal.,1975,1977;Yangetal.,1998).However,thespatialresolutionofthesetech-niques,didnotallowassessmentofresponsesofindividualglomerulitodifferentodors.

Spatiotemporalactivitypatternsintheolfactorybulbwereﬁrststudiedinsalamandersusingvoltage-sensitivedyes(Kauer,1988).Inarecentstudyinrats,RubinandKatz(1999)usedopticalimagingofintrinsicsignalstovisualizethepatternsofactivationofdifferentglomeruliinresponsetoawiderangeofodorants.Thestudyprovidedreﬁnedinformationofodorantorganization;odorantsarerepre-sentedbydistributedpatternsofactivatedglomerulithatarebilaterallysymmetric,anddistinctpatternsofglomerularactivitycorrelatewithdifferencesinodorantconcentrationandodorantidentity.Furtherstudiesshowedthatodorantswithdifferentfunctionalgroupsactivatedistinctdomainsintheolfactorybulbandthatsubtlechangesinodorant

structure,suchaslengthorconﬁgurationofcarbonchainselicitdistinctactivitypatterns(MeisterandBonhoeffer,2001;Uchidaetal.,2000).Forexample,shortchainlengthsofaliphaticaldehydesaremappedinthemiddleofeacholfactorybulb,whereasglomerulirespondingtothelongestaldehydesarefoundnearthelateraledgeofthebulb.Thus,anorderedrepresentationormapofodorsexistsintheolfactorybulb,basedonthealiphaticchainlength.More-over,fromthedynamicsoftheresponses,MeisterandBonhoeffer(2001)concludedthatthesignalsprobablyde-rivedfromafferentsoftheolfactorysensoryneuronsratherthanfromsecond-orderneurons.(Notethatinthisrespect,theolfactorybulbdiffersfromsensoryneocortex,whereintrinsicsignalsaredominatedbypostsynapticevents.)Histochemicalanalysisusingcytochromeoxidaseshowedthatfunctionalglomerulimatchedinsizeanddistributionanatomicallydeﬁnedglomeruli(BelluscioandKatz,2001;MeisterandBonhoeffer,2001).Thus,thesestudieshavepro-videdimportantinformationregardingthemolecularbasisofodorantrepresentationsandthefunctionalarchitectureoftheolfactorybulb.

3.1.4.StudiesonfunctionalarchitectureofauditorycortexClassicalelectrophysiologicalstudiesinthecatauditorycortex(Merzenichetal.,1973,1975;Rose,1949)describedacoreprimaryauditoryarea(AI)surroundedbyananteriorauditoryﬁeld(AAF),aventralsecondaryarea(AII),andaposteriorectosylvianﬁeldformingabeltaroundthecore.Basedonextensiveelectrophysiologicalmappingstud-ies,aprecisecochleotopicmapoftonerepresentationwasﬁrstdescribedincatauditorycortexbyMerzenichandco-workers(1975).Inpioneeringopticalimagingstud-ies,Bakinetal.(1996)reportedasuprathresholdtono-topicorganizationofratandguinea-pigauditorycortex,whileHessandScheich(1996)describedfrequency-andintensity-dependentspatiotemporalactivitypatternsinAIofawakegerbils.

Inalaterstudy,Harrisonandco-workers(1998)assessedsoundfrequencyandintensityresponsesinprimaryaudi-torycortexoftheanesthetizedchinchillaanddetectedin-trinsicactivityinanareacorrespondingtotheelectrophysi-ologicallydeﬁnedAIcortex.Inagreementwithelectrophys-iologicaldata,theauthorsfoundalow-tohigh-frequencytonotopic(orcochleotopic)organizationalongtheantero-posterioraxisofAI.Morerecently,Hareletal.(2000)de-ﬁnedauditoryareasAIIandAAFinchinchillaonthebasisofintrinsicactivity,andtheywereabletoshow,withinAI,AII,andAAF,atonotopicorganizationbasedonpuretonesatoctave-spacedfrequenciesfrom500Hzto16kHz.TheyfoundthemapsinAIandAIItobearrangedorthogonaltoeachother.

Inadditiontothestudyofcochleotopicmaps,opticalimaginghasbeenemployedtoaddresstheeffectsofacuteelectricalcochlearstimulationonthetopographyofthecatauditorycortex(Dinseetal.,1997).Theauthorsreportthatsystematicvariationofthecochlearfrequencysitesevoked

A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–2113

acorrespondingshiftoftheresponseareasthatmatchedtheunderlyingfrequencyorganization,thussuggestingtheutilityofopticalimaginginmappingresponseareasevokedbyelectricalcochlearstimulation.

Eventhoughprogresshasbeenmadeinthestudyoftheauditorycortexusingopticalimagingofintrinsicsignals,reportsaresparsecomparedtothoseonothersensorycor-ticalareas.Theremaybeseveralreasonsforthis.First,op-timalstimuliforevokingsustainedactivityinprimaryandsecondaryareashavebeendifﬁculttodetermine;typicallyacousticstimuliproduceshortburstsofrelativelyfewspikes,whicharelikelyassociatedwithonlyamoderateincreaseinmetabolicdemand.Theresultinglowsignal-to-noiseratiomayberesponsibleforthelowspatialresolution(400␮m)andtheconsiderableoverlapinintrinsicsignalevokedbytonesofdifferentfrequencies(e.g.Spitzeretal.,2001).Second,inareasoutsideofAI,anesthesiamayinducetheenhancementofinhibitorymechanismsthusleadingtoareductionintonicresponses(Hareletal.,2000;Zuritaetal.,1994).Neuralactivityinsecondaryareasmaybesoweakthattheassociatedmetabolicdemandisnotsufﬁcienttoinitiateameasurablehemodynamicresponsetoacousticstimuli.

Tworecentapproachespromiseanimprovedsignalqual-ityinintrinsicsignalimagingofauditorycortex.First,Versneletal.(2002)carriedoutasystematicstudycom-paringtheefﬁcacyofdifferenttypesofsoundstimulitoevokeintrinsiccorticalsignalsinAIofanesthetizedferrets.Theyfoundtone-piptrainsaswellasfrequency-modulatedtonestobeoptimal.Furthermore,greenilluminatinglight(546nm)appearedtoyieldmoreconsistentresultsthanthewavelengthsof600–700nmusedinimagingofvisualareas,despitestronglyincreasedvascularartifactsandtheaddi-tionaldrawbackthatthespatialcorrelationoftheopticalactivitywithneuralactivityissmallerandtheactivatedareaislarger.Another,perhapsevenmorepromisinginnova-tionistheuseofcontinuousperiodicstimulationoriginallydescribedbyKalatskyandStryker(2003)inanimagingstudyofthevisualcortex(seeabove).BothKalatskyandStryker(2002)andMrsic-Flogel,GrotheandHübener(per-sonalcommunication)havebeenabletorapidlyobtainhigh-qualitytonotopicmapsfromratandgerbilAI,respec-tively,usingascendinganddescendingtone-pipstimuli.3.2.Chronicexperimentsinintactanimals

Oneofthegreatestadvantagesofopticalimagingisthatitallowstherepeatedrecordingofmultipleactivitymapsinsingleanimalsandenablesstudyingthefunctionalarchi-tectureofparticularcorticalareasoveraperiodofweeksorevenmonths.Chronicexperimentsusingopticalimaginghavebeendesignedtostudytheontogeneticdevelopmentofcorticalmapsaswellastoexplorefunctionalmapsinbehavinganimalsandtofollowupfunctionalmapsinex-perimentalmodelsofmonoculardeprivation,ischemiaandepilepsy.

3.2.1.Developmentalstudies

Opticalimagingofthebrain,especiallyinyounganimals,isarelativelynon-invasiveprocedure,neverthelesschronicimagingrequiressomemethodologicalchangesandspecialcaremustbetakeninordertominimizetheriskofinfection.Ideally,thedurashouldbeleftintact,soasnottoexposethebrainitself.Thisisusuallypossibleinstudiesofyoungcatsorferrets,wheretheduraistranslucentenoughunlessexcessivegrowthoccursfollowingtheinitialexposure.Chronicopticalimagingstudieshaveelucidatedthede-velopmentoforientationselectivityinthevisualcortexofcatsandferrets(Chapmanetal.,1996;ChapmanandBonhoeffer,1998;Gödeckeetal.,1997).Orientationpref-erencemapsappearveryearlyindevelopment,ataroundthetimeofeye-opening,andalthoughithastobeborneinmindthatanormallyrearedanimalwillhavesomepat-ternedvisualexperiencedthroughtheclosedeyelids(Krugetal.,2001),orientationmapshaveevenbeenobservedintheabsenceofanyvisualexperiencefollowingdark-rearing(Whiteetal.,2001).However,normalvisualinputandnor-malpatternsofneuronalactivityarenecessaryforthemat-urationandmaintenanceoforientationmaps(Crairetal.,1998;ChapmanandGödecke,2000;Whiteetal.,2001).Animportantissueincorticaldevelopmentistheques-tionofmapstabilityovertime.Chapmanetal.(1996),Gödeckeetal.(1997)andShtoyermanetal.(2000)havealldemonstratedthestabilityoforientationpreferencemapsinthedevelopingvisualcortexofyoungcatsandferretsaswellasinV1ofadultmacaquemonkeys.Moreover,KimandBonhoeffer(1994)andGödeckeandBonhoeffer(1996)showedthatevenintheabsenceofanycommonvisualex-periencemoreorlessidenticalorientationmapsdevelopin-dependentlythroughleft-andright-eyestimulation(seealsoSection3.2.2).

3.2.2.Visualandsomatosensorycortexplasticity

Opticalimagingisanexcellenttoolforassessingplasticityofvisualcorticalmapsinresponsetovariousmanipulationsofthevisualinput(Dragoietal.,2000;Schuettetal.,2001;Sengpieletal.,1998,1999).

Acuteaswellaschronicexperimentshavefocusedontheplasticityoforientationpreferenceand,toalesserex-tent,oculardominancemaps.Awiderangeofexperimen-talmanipulationsincludingstrabismus(Engelmannetal.,2002;Löweletal.,1998;Sengpieletal.,1998),monocularandbinoculardeprivation(Crairetal.,1997,1998;Gödeckeetal.,1997;Issaetal.,1999),reverselid-suture(GödeckeandBonhoeffer,1996),stripe-rearing(Sengpieletal.,1999),dark-rearing(Whiteetal.,2001),patternadaptation(Dragoietal.,2000)aswellasacombinationofvisualandelectricalstimulation(Schuettetal.,2001)havegenerallyledtotheconclusionthatwhilee.g.thesizeofindividualfunctionaldomainscanbeaffecteddramatically,theoveralllayoutofthemaps(e.g.periodicityofmodules)appearstoberemark-ablystable,althoughitcanreorganizetosomeextentafterfocalischemicinjury(Zepedaetal.,2003).

14A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–21

Opticalimaginghasalsoservedtoelucidatetheroleofneurotrophinsandtheirreceptorsinvisualcorticalplastic-ity.Inarecentstudy,Gillespieandcoworkers(2000)ex-aminedthefunctionaleffectsofinfusionofNT-4/5,NGF,andneurotrophin-3(NT-3)onoculardominanceplasticitycausedbymonocularvisualdeprivationduringthecriticalperiodinkittens.Thisstudyrevealedthatvisualcortexre-ceivinganNT-4/5(butnotNT-3)infusionfor2daysatthepeakofthecriticalperiod,showedenhancedcorticalre-sponsestothedeprived-eye,butthatorientationpreferencemapswerelostwithintheinfusedregion.

Beforestudyingplasticityofsomatosensorycortexrep-resentationsthroughopticalimagingitwasimportanttoes-tablishthestabilityanddynamicsofopticalsignalsfromsomatosensorycortexovertime.MasinoandFrostig(1996)foundthatstimulationofasinglewhiskerconsistentlyacti-vatedasurprisinglylargeareaofbarrelcortex.Whileloca-tionofthefunctionalrepresentationandtimecourseofthestimulus-relatedintrinsicsignalresponsewereconsistent,non-systematicchangesbothintheshapeandthearealex-tentofthewhiskerrepresentation,aswellastheamplitudeoftheintrinsicsignalwereobserved.Therefore,quantitativeimagingresultsfrombarrelcortexmustbeinterpretedwithcare,sinceoptimalmethodsfordataanalysishavenotyetbeenestablished.Thepossiblereasonshavebeendiscussedearlier(seeSection3.1.2).Despitethesecaveats,severalgroupshaveattemptedtostudyplasticityofthesomatosen-sorysystemusingopticalimaging.

Prakashetal.(1996)exploredtheeffectsofthetopicalapplicationofdifferentneurotrophinsonthebarrelcortex.TheyshowedthattopicalapplicationofBDNFresultedinarapidandlong-lastingdecreaseinthesizeofawhiskerrepresentation,andadecreaseintheamplitudeoftheactivity-dependentintrinsicsignal.Incontrast,NGFappli-cationprovokedarapidbuttransientincreaseinthesizeofawhiskerrepresentation,accompaniedbyanincreaseintheamplitudeoftheactivity-dependentintrinsicsignal.Thus,neurotrophinsexertdifferentialeffectsontheactivityandfunctionalorganizationofwhiskerrepresentations.Studiesonreorganizationafterperipheraldeafferentation(Polleyetal.,1999)haveshownthatafterwhiskerremoval,plasticchangesareexpressedeitherasanexpansionoracontractionofthesparedwhisker’sfunctionalrepresenta-tiondependingontheanimal’susageofitswhiskersduringtheperiodofsensorydeprivation.Thereasonastowhypro-vidingtheanimalswithanopportunitytousetheirsparedwhiskerinactiveexplorationresultsinadecreaseinitsfunc-tionalrepresentationremainsunknown.

3.2.3.Studiesinawakeanimals

Despitethefactthatmostimagingstudiesaimedatex-ploringthefunctionalmodularityinsensoryneocortex,thestudyofsuchcorticalmodulesinrelationtoperceptualandcognitivebehaviorinawakeanimalsandhumansisrecent(Cannestraetal.,2000;Grinvaldetal.,1991;Haglundetal.,1992;Satoetal.,2002;Shtoyermanetal.,2000;Siegeletal.,2003;Vneketal.,1999).Anesthetizedsubjectsareunsuit-ableformanytypesofstudies,suchasmotivation,attentionorarousalaffectingsensoryprocessingandperception,mo-torfunction,consciousness,andothercognitivefunctions.Inaddition,long-termplasticchangesrelatedtomemoryandlearningorrecoveryoffunctionaftertraumaorstrokearedifﬁculttopinpointwithoutimaging.Studiesinhumansub-jectsarelimitedtonon-invasiveapproaches(EEG,fMRI);electricalrecordingoranatomicaltechniquesarenotanop-tion.Therefore,fortheforeseeablefuture,theawakemon-keymodelislikelytoremainthepreparationofchoicetounderstandbetterthefunctionalorganizationoftheprimatebrain.

Experimentsinawakeandbehavingmonkeyrequiredanumberofadditionalissuestobesolved.Amongthemprob-ablythemostimportantconcernedtheeliminationofthenoiseresultingfrommovementandthelargeopticalnoiseproducedbycardiacandrespiratorypulsations.Anotherim-portantissuethathadtobeaddressedwastheeffectofanes-thesiaonthecharacteristicsoftheintrinsicopticalsignal.Inordertoresolvetheseproblemsandtoinvestigateifimag-inginawakeanimalswasfeasible,theﬁrststudiesaimedatcomparingcorticalactivityobtainedintheawakever-sustheanesthetizedanimal.Inapioneeringstudy,Grinvaldetal.(1991)performedopticalimagingofoculardominancecolumnsinV1ofuntrainedmonkeys,bytakingimagesoftheexposedcortexwhiletheanimalwasviewingvideomoviesalternativelywitheacheye.TheyusedachamberlikethatdescribedinSection2.3.1,whichdiminishedmovementrelatedtopulsationandmaintainedthecortexinaclosedenvironmentwheneverthesubjectwasnotintherecordingapparatus.Theyalsodescribedthatmovement-relatednoisecouldbealmostcompletelyavoidedby:(i)mountingthemonkey-chairtoaheavyanti-vibrationtable;(ii)holdingtheheadofthemonkeybyrestrictingitsmovementand;(iii)usingrigidbarstoanchortheheadholder,themonkey-chairandthelensofthecameratoeachother.Bythisproceduretheywereabletoobtainhighresolutionmapsofoculardom-inancecolumns,andobservedthatitwasnotnecessarytosynchronizerespiratoryandcardiaccyclestoimageacqui-sition.

Thisstudyrevealedthatalthoughtheintrinsicopticalsig-nalsintheawakeanimalweresimilartotheanesthetizedanimalinwavelengthdependencyandtimecourse,thelatterwasslightlyslowerintheawakeanimal.Therefore,alongerframedurationforimageacquisitionwasusedtoimprovethequalityofthemaps.However,ShohamandGrinvald(2001)haveoptimizedtheimagingprocedures,andhavebeenabletoobtainfunctionalmapsmorerapidly.More-over,Vneketal.(1999)showedthatwelltrainedanimalsrequiredconsiderablyfewertrialsinordertoobtainagoodsignal-to-noiseratio.

Opticalimaginghasalsocontributedtotheunderstandingofthefunctionalarchitectureinassociationcortices,partic-ularlytheinferiorparietallobeareas,whoseneuronscom-binevisualinformationwitheyepositionsignals.Inarecent

A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–2115

study,Siegeletal.(2003)reportedanoveltopographicalmapoftheeffectofeyepositiononvisualresponsestoop-ticﬂowintheinferiorparietallobuleofmacaquemonkeys.Theauthorsproposedthatthisfunctionalarchitecturemayserveasthescaffoldingonwhichothersensory,attentional,andintentionalmapsmaybeembeddedatﬁnercolumnarscales.

3.3.Studiesinhumans

Duetotheopacityofthehumanskull,opticalimagingofintrinsicsignals,ofthekinddescribedinthisreviewsofar,inhumanshasbeenlimitedtointra-operativeprocedures.However,recentadvancesinopticalimagingtechnologyandmethodology,reviewedbyPouratianetal.(2003),haveal-lowedsigniﬁcantlyimprovingopticalimagingstudiesofthehumancortexandevenmakingthestudyofhighercognitiveprocessesfeasible.

ThepioneeringstudiesbyCannestraetal.(1998),Haglundetal.(1992)andTogaetal.(1995)usingopticalimaginginsurgicalprocedures,wereaimedatdelineatingfunctionalbordersthathelpedminimizingthepotentialdamagetohealthytissuethatcanoccurduringresectionoftumors,epilepticfociorarteriovenousmalformations.

Oneprominentareaofstudyhasbeenthesomatosensoryandmotorcortex.Togaetal.(1995)analyzedthetempo-ralandspatialevolutionofopticalsignalscombinedwithevokedpotentialinresponsetotranscutaneouselectricalstimulationofthemedianandulnarnervesinpatientsunder-goingsurgicalresectionofbraintumors.Theobtainedop-ticalsignalcolocalizedwiththelargestevokedpotentialsinbothmotorandsensoryregions,illustratingtherelationshipbetweenneuronalﬁringandvascularandmetabolicfunc-tion.Morerecently,Cannestraetal.(1998)demonstratedthatthepeakopticalresponsesgeneratedafterthestimu-lationofdifferentﬁngersdonotoverlapbutthenon-peaksignalsdo.Thisresultcouldbedueeithertoapartialover-lappingofdigitrepresentationsinthehumancortexortothelargepoint-spreadgenerallyobservedinsomatosensorycor-tex(seeabove).Inanotherstudyofhumansomatosensorycortex,Satoetal.(2002)identiﬁedaneuronalresponseareainthepostcentralgyrusdifferentiallyactivateddependingontheﬁngerthatwasstimulated.Whileﬁrstandﬁfthdigitstimulationelicitedopticalresponsesindifferentareasnearthecentralsulcus,theirstimulationresultedinoverlappingactivityclosertothepostcentralsulcus,suggestingahierar-chicalorganizationoftheprimarysomatosensorycortex.Anotherapplicationofopticalimaginginhumanshasbeenthefunctionalcharacterizationofcorticallanguagear-easinawakepatientsundergoinganeurosurgicalproce-dure(Cannestraetal.,2000;Haglundetal.,1992;Pouratianetal.,2000).ThepioneeringstudybyHaglundetal.(1992)demonstratedcognitivelyevokedactivityinlanguageareas.Inthisreport,theauthorscorrelatedtheopticalchangeswithcorticalstimulationmappingandobservedthatfunctionalimagingyieldedsigniﬁcantactivationinbothessentialand

secondarylanguagesregions,incontrastwithelectrocorticalstimulation(ESM),whichonlyidentiﬁedtheessentiallan-guagecortex(i.e.Wernicke’sandBroca’sareas).Cannestraetal.(2000)usedimagingcoupledwithESMandstudiedcorticalactivationinresponsetodifferentlanguagetasksinawakepatients.Distinctspatialandtemporalresponsepat-terns,dependentontaskandperformance,werecharacter-izedbothwithinBroca’sandWernicke’sareas,consistentwiththeexistenceoftask-speciﬁcsemanticandphonologicregionswithintheseareas;thedifferingtemporalpatternswereproposedtoreﬂectuniqueprocessingperformedbyreceptive(Wernicke’s)andproductive(Broca’s)languagecenters.

Theﬁrstnon-invasiveopticalimagingstudiesinhumanswereperformedwithNIRS(VillringerandChance,1997).Usinglightof700–1000nmwavelengthforillumination,thereﬂectancesignalinNIRSprimarilyrepresentstheincreaseinoxy-hemoglobin(anddecreaseindeoxy-hemoglobin)associatedwiththedelayedincreaseinbloodﬂowandvolumefollowingcorticalactivationandisthereforeofoppositepolaritytotheintrinsicsignalsthatthisreviewisconcernedwith.TheNIRSsignalhasasimilartime-courseandspatialpatternandresolutionlimitasmostfMRIstud-ies.Morerecently,Grattonandco-workers(Grattonetal.,1997;GrattonandFabiani,2001)havedevelopedanopticalimagingmethod,inwhichthesignalisderivedprimarilyfromactivity-inducedchangesinlightscatteringandwhichisthereforemuchclosertothetechniquereviewedhere.This“event-relatedopticalsignal”(EROS)allowsaspa-tialresolutionofbetterthan1cm3andhasalatencyofaround100ms.Illuminationatwavelengthsof690–850nmistypicallyprovidedbylaserdiodes,which(incontrastwithhalogenlamps)permithighfrequencymodulation(110–220MHz)oflightintensity;aslightdetectors,photo-multipliertubesorCCDcamerascanbeused.Thevariationoftheincidentlightintensityallowsforprecisemeasure-mentsofthetimerequiredbyphotonstotravelfromthesourcetothedetector;thedifferentdelaysintroducedbycorticaltissuesconstituteamoresensitivemeasureofneu-ralactivitythanthetotalnumberofphotonsabsorbed.Themaximumdepthofpenetrationwiththistechniqueiscurrentlylimitedtoabout3cm(GrattonandFabiani,2001).3.4.Clinicallyrelevantstudies

Opticalimaginghasnotonlyprovedtobeabreakthroughintheunderstandingoffunctionalorganizationandphysi-ologyofthecerebralcortex,buthasalsoallowedgaininginsightsintopathophysiologicalprocesses,suchasepilepsyandstroke.

3.4.1.Epilepsy

Epilepsyisaneurologicalconditioncharacterizedbyre-currentseizureswhichcomprisecomplexelectricalﬁringofapopulationofneurons.Moststudiesinepilepsyhavebeendoneusingelectrophysiologicalrecordingsfromsur-

16A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–21

faceﬁeldandextracellularsingle-unitelectrodes.Howeverthesemethodologieshavesigniﬁcantlimitationsintheacutelocalizationofthegenerationandspreadofneuronalactiv-itymainlyduetotemporalandspatialsamplinglimitations.Othertechniquesbasedonthefocalcouplingofalterationofbloodﬂowandmetabolismwithneuronalactivity,suchasfMRI,positronemissiontomography(PET),singlephotonemissioncomputedtomography(SPECT),donothavethetemporalorspatialresolutiontoresolvebriefparoxysmalorinterictalspikesandtolocalizeproximalareasofearlyelectricalactivityspread.

Togaininsightsinthestudyoftheepileptogenicpatho-physiology,somemodelsofepilepsy,suchascorticalslices(Hochmanetal.,1995),isolatedguinea-pigwholebrain(Federicoetal.,1994),penicillin-inducedseizuresinrat(Chenetal.,2000),inducedepilepticfociinferretcerebralcortex(SchwartzandBonhoeffer,2001)andevenintra-operativeproceduresinhumans(Haglundetal.,1992)havebeenanalyzedusingOI.

SchwartzandBonhoeffer(2001)mappedspontaneousepilepticevents,suchasinterictal,ictalandsecondaryho-motopicfociaswellasadecreasedneuronalactivitysur-roundingtheepilepticfocusinvivo.Inthisstudy,interictaleventswereinducedbythefocalapplicationoftheGABAAreceptorantagonist,bicucullineandictalactivitybyinject-ing4-aminopyridineintocorticallayersII/III.OIallowedthegenerationofhigh-resolutionmapsofthespreadepilep-tiformactivityanditsrelationwiththefunctionalcorticalarchitectureinrealtime.

Opticalmappingoffersthepotentialofbeingusedintra-operativelyduringtheresectionofanepilepticfocusinhumans,whereitcouldcontributetoamuchhigher

Fig.5.TemporalevolutionofchangesinorientationpreferencemapsandlesionsizefollowingafocalischemiclesioninkittenV1(modiﬁedfromZepedaetal.,2003).Highmagniﬁcationofpolarorientationmapsoftheimagedcorticalareainonekitten(A–D):(A)intheintactcortex(pre-lesion);(B)immediatelypostlesion;(C)13dayspostlesion(dPL)and,(D)33dayspostlesion.Arrowheadspointatdomainswhichrecoveredafter13dPLandenlargedbyday33PL.Orientationofimagedcorticalareaisshown(a,anterior;p,posterior;m,medial;l,lateral).Scalebar:1mm.(E)Reductionoffunctionallysilentareaobtainedfrompolarmapsduringupto5weeksfollowingthelesion.Eachdatapointrepresentsthemean±S.D.ofthesilentareapertime-groupasassessedthroughimaging.

A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–2117

degreeofneurosurgicalprecision(Haglundetal.,1992;SchwartzandBonhoeffer,2001).

3.4.2.Ischemiaandstroke

Anumberofexperimentshaveshownthatcerebrovas-culardiseasecansigniﬁcantlyinﬂuencethecerebralbloodﬂowandoxygenresponsetofunctionalactivation.However,untilrecentlyeffectsonfunctionalcorticalmapsofanis-chemiceventremainedunknown.Inarecentstudy,Zepedaetal.(2003)inducedasmallphotochemicallesioninpri-maryvisualcortexofkittensandanalyzedthesubsequentreorganizationofcorticalmapsusingopticalimaging.Giventhatphotochemicallesionsareinducedwithoutmanipulat-ingthebrainandarehighlyreproducible,theyprovideavery“clean”methodforstudyingtheconsequencesofafo-calischemicevent.

Zepedaetal.(2003)observedanareaofcapillaryocclu-sionthatwasco-extensivewithanareavoidoffunctionalactivityimmediatelyafterthelesion(Fig.5).Somerevascu-larizationstartedwithinthefunctionallysilentareaasearlyas2weeksafterthelesion;thisprocesscoincidedintimewiththereductioninsizeoftheinactiveregion.Moreover,nearthelesionboththeretinotopicandtheorientationpref-erencemapswerefoundtoreorganizeoveraperiodof5weeksafterthelesion.

4.Conclusion

Opticalimaginghasemergedasapotenttooltoanalyzethespatialdistributionofneuronalactivityinvivooverlargeareasofthebrainsurface,withhighspatialresolution.OIstudiescontributetoourunderstandingoftheneuronalin-tegrationofdifferentstimulusfeaturesatapopulationlevelandallowtoobservethefunctionaldevelopmentofthebrainandtostudyitsplasticityunderdifferentexperimentalma-nipulationsandinexperimentalmodelsofneuropathology.Moreover,OImaybeappliedinhumanintra-operativepro-cedures,providingatoolfordelineatingthefunctionalbor-dersofepilepticfociorduringatumorresection.Ifin-trinsicsignalimagingmethods,suchasEROS,whichal-lowimagingthroughtheintacthumanskull,couldbeim-provedfurtherintermsofspatialresolutionandacquisi-tiontimes,thenwemightevenbeabletostudythefunc-tionalarchitectureofthehumancerebralcortexandmon-itorpatternsofactivityduringtheexecutionofdifferenttasks.

Acknowledgements

WethankThomasMrsic-Flogelforhelpfulcommentsonthemanuscript.A.Z.andC.A.weresupportedbyCONA-CYT36250M.F.S.wassupportedbytheMedicalResearchCouncilandtheHumanFrontiersScienceProgramOrgani-zation.

References

ArieliA,GrinvaldA.Opticalimagingcombinedwithtargetedelectrical

recordings,microstimulation,ortracerinjections.JNeurosciMethods2002;116:15–28.

ArieliA,GrinvaldA,SlovinH.Duralsubstituteforlong-termimaging

ofcorticalactivityinbehavingmonkeysanditsclinicalimplications.JNeurosciMethods2002;114:119–33.

AttwellD,LaughlinSB.Anenergybudgetforsignalinginthegrey

matterofthebrain.JCerebBloodFlowMetab2001;21:1133–45.BakinJS,KwonMC,MasinoSA,WeinbergerNM,FrostigRD.

Suprathresholdauditorycortexactivationvisualizedbyintrinsicsignalopticalimaging.CerebCortex1996;6:120–30.

BartfeldE,GrinvaldA.Relationshipsbetweenorientation-preferencepin-wheels,cytochromeoxidaseblobs,andocular-dominancecolumnsinprimatestriatecortex.ProcNatlAcadSciUSA1992;:11905–9.BasoleA,WhiteLE,FitzpatrickD.Mappingmultiplefeaturesinthe

populationresponseofvisualcortex.Nature2003;423:986–90.

BellAJ,SejnowskiTJ.Aninformation-maximizationapproachtoblind

separationandblinddeconvolution.NeuralComput1995;7:1129–59.BelluscioL,KatzLC.Symmetry,stereotypy,andtopographyofodorant

representationsinmouseolfactorybulbs.JNeurosci2001;21:2113–22.BlasdelG,CampbellD.Functionalretinotopyofmonkeyvisualcortex.

JNeurosci2001;21:8286–301.

BlasdelGG,SalamaG.Voltage-sensitivedyesrevealamodularorgani-zationinmonkeystriatecortex.Nature1986;321:579–85.

BloodAJ,NarayanSM,TogaAW.Stimulusparametersinﬂuencecharac-teristicsofopticalintrinsicsignalresponsesinsomatosensorycortex.JCerebBloodFlowMetab1995;15:1109–21.

BonhoefferT,GrinvaldA.Iso-orientationdomainsincatvisualcortex

arearrangedinpinwheel-likepatterns.Nature1991;353:429–31.BonhoefferT,GrinvaldA.Thelayoutofiso-orientationdomainsinarea

18ofcatvisualcortex:opticalimagingrevealsapinwheel-likeorga-nization.JNeurosci1993;13:4157–80.

BonhoefferT,GrinvaldA.Opticalimagingbasedonintrinsicsignals.

Themethodology.In:TogaA,MazziotaJ,editors.Brainmapping:themethods.London:AcademicPress;1996.p.55–97.

BonhoefferT,KimDS,MalonekD,ShohamD,GrinvaldA.Optical

imagingofthelayoutoffunctionaldomainsinarea17andacrossthearea17/18borderincatvisualcortex.EurJNeurosci1995;7:1973–88.BoskingWH,CrowleyJC,FitzpatrickD.Spatialcodingofpositionand

orientationinprimaryvisualcortex.NatNeurosci2002;5:874–82.BoskingWH,ZhangY,SchoﬁeldB,FitzpatrickD.Orientationselectivity

andthearrangementofhorizontalconnectionsintreeshrewstriatecortex.JNeurosci1997;17:2112–27.

BoskingWH,KretzR,PucakML,FitzpatrickD.Functionalspeci-ﬁcityofcallosalconnectionsintreeshrewstriatecortex.JNeurosci2000;20:2346–59.

BoyntonGM,EngelSA,GloverGH,HeegerDJ.Linearsystemsanalysis

offunctionalmagneticresonanceimaginginhumanV1.JNeurosci1996;16:4207–21.

BozzaTC,MombaertsP.Olfactorycoding:revealingintrinsicrepresen-tationsofodors.CurrBiol2001;11:R687–90.

Brett-GreenBA,Chen-BeeCH,FrostigRD.Comparingthefunctional

representationsofcentralandborderwhiskersinratprimaryso-matosensorycortex.JNeurosci2001;21:9944–54.

CarandiniM,SengpielF.Contrastinvarianceoffunctionalmapsincat

primaryvisualcortex.JVision2004;4:130–43.

CannestraAF,PouratianN,BookheimerSY,MartinNA,BeckerandDP,

TogaAW.TemporalspatialdifferencesobservedbyfunctionalMRIandhumanintraoperativeopticalimaging.CerebCortex2001;11:773–82.

CannestraAF,BlackKL,MartinNA,CloughesyT,BurtonJS,Rubin-steinE,WoodsRP,TogaAW.Topographicalandtemporalspeci-ﬁcityofhumanintraoperativeopticalintrinsicsignals.NeuroReport1998;9:2557–63.

18A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–21

CannestraAF,BookheimerSY,PouratianN,O’FarrellA,SicotteN,Mar-tinNA,BeckerD,RubinoG,TogaAW.Temporalandtopographicalcharacterizationoflanguagecorticesusingintraoperativeopticalin-trinsicsignals.NeuroImage2000;12:41–54.

ChapmanB,BonhoefferT.Overrepresentationofhorizontalandvertical

orientationpreferencesindevelopingferretarea17.ProcNatlAcadSciUSA1998;95:2609–14.

ChapmanB,GödeckeI.Corticalcellorientationselectivityfailstodevelop

intheabsenceofON-centerretinalganglioncellactivity.JNeurosci2000;20:1922–30.

ChapmanB,StrykerMP,BonhoefferT.Developmentoforientationpref-erencemapsinferretprimaryvisualcortex.JNeurosci1996;16:42–53.

ChenJW,O’FarrellAM,TogaAW.Opticalintrinsicsignalimagingina

rodentseizuremodel.Neurology2000;55:312–5.

ChenLM,FriedmanRM,RamsdenBM,LaMotteRH,RoeAW.Fine-scale

organizationofSI(area3b)inthesquirrelmonkeyrevealedwithintrinsicopticalimaging.JNeurophysiol2001;86:3011–29.

ChenLM,HeiderB,WilliamsGV,HealyFL,RamsdenBM,RoeAW.

Achamberandartiﬁcialduramethodforlong-termopticalimaginginthemonkey.JNeurosciMethods2002;113:41–9.

Chen-BeeCH.Variabilityandinterhemisphericasymmetryofsingle-whiskerfunctionalrepresentationsinratbarrelcortex.JNeurophysiol1996;76:884–94.

Chen-BeeCH,KwonMC,MasinoSA,FrostigRD.Arealextentquan-tiﬁcationoffunctionalrepresentationsusingintrinsicsignalopticalimaging.JNeurosciMethods1996;68:27–37.

ClarkeDD,SokoloffL.Circulationandenergymetabolismofthebrain.

In:SiegelG,AgranoffB,AlbersR,FisherS,UhlerM,editors.Basicneurochemistry:molecular,cellularandmedicalaspects.Philadelphia:Lippincott-Raven;1999.p.637–69.

CohenLB.Changesinneuronstructureduringactionpotentialpropagation

andsynaptictransmission.PhysiolRev1973;53:373–418.

CrairMC,GillespieDC,StrykerMP.Theroleofvisualexperienceinthe

developmentofcolumnsincatvisualcortex.Science1998;279:566–70.

CrairMC,RuthazerES,GillespieDC,StrykerMP.Relationshipbetween

theoculardominanceandorientationmapsinvisualcortexofmonoc-ularlydeprivedcats.Neuron1997;19:307–18.

DasA,GilbertCD.Long-rangehorizontalconnectionsandtheirrolein

corticalreorganizationrevealedbyopticalrecordingofcatprimaryvisualcortex.Nature1995;375:780–4.

DinseHR,GoddeB,HilgerT,ReuterG,CordsSM,LenarzT,vonSeelen

W.Opticalimagingofcatauditorycortexcochleotopicselectivityevokedbyacuteelectricalstimulationofamulti-channelcochlearimplant.EurJNeurosci1997;9:113–9.

DragoiV,SharmaJ,SurM.Adaptation-inducedplasticityoforientation

tuninginadultvisualcortex.Neuron2000;28:287–98.

EngelSA,RumelhartDE,WandellBA,LeeAT,GloverGH,Chichilnisky

EJ,ShadlenMN.fMRIofhumanvisualcortex.Nature1994;369:525.EngelmannR,CrookJM,LöwelS.Opticalimagingoforientationand

oculardominancemapsinarea17ofcatswithconvergentstrabismus.VisNeurosci2002;19:39–49.

FedericoP,BorgSG,SalkauskusAG,MacVicarBA.Mappingpatternsof

neuronalactivityandseizurepropagationbyimagingintrinsicopticalsignalsintheisolatedwholebrainoftheguinea-pig.Neuroscience1994;58:461–80.

FrostigRD,LiekeEE,Ts’oDY,GrinvaldA.Corticalfunctionalarchitec-tureandlocalcouplingbetweenneuronalactivityandthemicrocircu-lationrevealedbyinvivohigh-resolutionopticalimagingofintrinsicsignals.ProcNatlAcadSciUSA1990;87:6082–6.

GabbayM,BrennanC,KaplanE,SirovichL.Aprincipalcomponents-basedmethodforthedetectionofneuronalactivitymaps:applicationtoopticalimaging.NeuroImage2000;11:313–25.

GhoseGM,Ts’oDY.FormprocessingmodulesinprimateareaV4.J

Neurophysiol1997;77:2191–6.

GillespieDC,CrairMC,StrykerMP.Neurotrophin-4/5altersresponses

andblockstheeffectofmonoculardeprivationincatvisualcortexduringthecriticalperiod.JNeurosci2000;20:9174–86.

GochinPM,BedenbaughP,GelfandJJ,GrossCG,GersteinGL.Intrinsic

signalopticalimagingintheforepawareaofratsomatosensorycortex.ProcNatlAcadSciUSA1992;:8381–3.

GoddeB,HilgerT,vonSeelenW,BerkefeldT,DinseHR.Optical

imagingofratsomatosensorycortexrevealsrepresentationaloverlapastopographicprinciple.NeuroReport1995;7:24–8.

GödeckeI,BonhoefferT.Developmentofidenticalorientationmapsfor

twoeyeswithoutcommonvisualexperience.Nature1996;379:251–4.GödeckeI,KimDS,BonhoefferT,SingerW.Developmentoforientation

preferencemapsinarea18ofkittenvisualcortex.EurJNeurosci1997;9:1754–62.

GrattonG,FabianiM.Theevent-relatedopticalsignal:anewtoolfor

studyingbrainfunction.IntJPsychophysiol2001;42:109–21.

GrattonG,FabianiM,CorballisPM,HoodDC,Goodman-WoodMR,

HirschJ,KimK,FriedmanD,GrattonE.Fastandlocalizedevent-relatedopticalsignals(EROS)inthehumanoccipitalcortex:comparisonswiththevisualevokedpotentialandfMRI.NeuroImage1997;6:168–80.

GrinvaldA,FrostigRD,SiegelRM,BartfeldE.High-resolutionoptical

imagingoffunctionalbrainarchitectureintheawakemonkey.ProcNatlAcadSciUSA1991;88:11559–63.

GrinvaldA,LiekeE,FrostigRD,GilbertCD,WieselTN.Functional

architectureofcortexrevealedbyopticalimagingofintrinsicsignals.Nature1986;324:361–4.

GrinvaldA,ShohamD,ShmuelA,GlaserDE,VanzettaI,ShtoyermanE,

SlovinH,WijnbergenC,HildesheimR,SterkinA,ArieliA.Invivoopticalimagingofcorticalarchitectureanddynamics.In:WindhorstU,JohanssonH,editors.Moderntechniquesinneuroscienceresearch.Heidelberg:SpringerVerlag;1999.p.3–969.

GuthrieKM,AndersonAJ,LeonM,GallC.Odor-inducedincreasesin

c-fosmRNAexpressionrevealananatomical“unit”forodorprocess-inginolfactorybulb.ProcNatlAcadSciUSA1993;90:3329–33.HaglundMM,OjemannGA,HochmanDW.Opticalimagingofepilep-tiformandfunctionalactivityinhumancerebralcortex.Nature1992;358:668–71.

HarelN,MoriN,SawadaS,MountRJ,HarrisonRV.Threedistinct

auditoryareasofcortex(AI,AII,andAAF)deﬁnedbyopticalimagingofintrinsicsignals.NeuroImage2000;11:302–12.

HarrisonRV,HarelN,KakigiA,RavehE,MountRJ.Opticalimaging

ofintrinsicsignalsinchinchillaauditorycortex.AudiolNeurootol1998;3:214–23.

HessA,ScheichH.OpticalandFDGmappingoffrequency-speciﬁc

activityinauditorycortex.NeuroReport1996;7:23–7.

HessA,StillerD,KaulischT,HeilP,ScheichH.Newinsightsintothe

hemodynamicbloodoxygenationlevel-dependentresponsethroughcombinationoffunctionalmagneticresonanceimagingandopticalrecordingingerbilbarrelcortex.JNeurosci2000;20:3328–38.

HildebrandJG,ShepherdGM.Mechanismsofolfactorydiscrimination:

convergingevidenceforcommonprinciplesacrossphyla.AnnuRevNeurosci1997;20:595–631.

HillDK,KeynesRD.Opacitychangesinstimulatednerve.JPhysiol

1949;108:278–81.

HochmanDW,BarabanSC,OwensJW,SchwartzkroinPA.Dissociation

ofsynchronizationandexcitabilityinfurosemideblockadeofepilep-tiformactivity.Science1995;270:99–102.

HubelDH,WieselTN.Receptiveﬁelds,binocularinteractionandfunc-tionalarchitectureinthecat’svisualcortex.JPhysiol1962;160:106–54.

HubelDH,WieselTN.Functionalarchitectureofmacaquemonkeyvisual

cortex.ProcRSocLondB1977;198:1–59.

HübenerM,ShohamD,GrinvaldA,BonhoefferT.Spatialrelation-shipsamongthreecolumnarsystemsincatarea17.JNeurosci1997;17:9270–84.

A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–21

ImamuraK,MatagaN,MoriK.Codingofodormoleculesbymitral/tufted

cellsinrabbitolfactorybulb.PartI.Aliphaticcompounds.JNeuro-physiol1992;68:1986–2002.

IssaNP,TrepelC,StrykerMP.Spatialfrequencymapsincatvisual

cortex.JNeurosci2000;20:8504–14.

IssaNP,TrachtenbergJT,ChapmanB,ZahsKR,StrykerMP.Thecritical

periodforoculardominanceplasticityintheFerret’svisualcortex.JNeurosci1999;19:6965–78.

JohnsonBA,WooCC,LeonM.Spatialcodingofodorantfeatures

intheglomerularlayeroftheratolfactorybulb.JCompNeurol1998;393:457–71.

KalatskyVA,StrykerMP.Newparadigmforopticalimaging:temporally

encodedmapsofintrinsicsignal.Neuron2003;38:529–45.

KalatskyVA,StrykerMP.Suprathesholdorganizationofratauditory

cortexrevealedbynewmethodofintrinsicsignalopticalimaging.SocNeurosciAbstractViewer2002;ProgramNo.354.15.

KauerJS.Real-timeimagingofevokedactivityinlocalcircuitsofthe

salamanderolfactorybulb.Nature1988;331:166–8.

KimDS,BonhoefferT.Reverseocclusionleadstoapreciserestorationof

orientationpreferencemapsinvisualcortex.Nature1994;370:370–2.KimDS,DuongTQ,KimSG.High-resolutionmappingofiso-orientation

columnsbyfMRI.NatNeurosci2000;3:1–9.

KimDS,MatsudaY,OhkiK,AjimaA,TanakaS.Geometricaland

topologicalrelationshipsbetweenmultiplefunctionalmapsincatprimaryvisualcortex.NeuroReport1999;10:2515–22.

KisvardayZF,KimDS,EyselUT,BonhoefferT.Relationshipbetween

lateralinhibitoryconnectionsandthetopographyoftheorientationmapincatvisualcortex.EurJNeurosci1994;6:1619–32.

KisvardayZF,TothE,RauschM,EyselUT.Orientation-speciﬁcrelation-shipbetweenpopulationsofexcitatoryandinhibitorylateralconnec-tionsinthevisualcortexofthecat.CerebCortex1997;7:605–18.KrugK,AkermanCJ,ThompsonID.Responsesofneuronsinneona-talcortexandthalamustopatternedvisualstimulationthroughthenaturallyclosedlids.JNeurophysiol2001;8:1436–43.

LandismanCE,Ts’oDY.Colorprocessinginmacaquestriatecortex:re-lationshipstooculardominance,cytochromeoxidase,andorientation.JNeurophysiol2002;87:3126–37.

LiuGB,PettigrewJD.Orientationmosaicinbarnowl’svisualWulst

revealedbyopticalimaging:comparisonwithcatandmonkeystriateandextra-striateareas.BrainRes2003;961:153–8.

LöwelS,SchmidtKE,KimDS,WolfF,HoffsümmerF,SingerW,

BonhoefferT.Thelayoutoforientationandoculardominancedomainsinarea17ofstrabismiccats.EurJNeurosci1998;10:2629–43.

LyonDC,XuX,CasagrandeVA,StefansicJD,ShimaD,KaasJH.

OpticalimagingrevealsretinotopicorganizationofdorsalV3inNewWorldowlmonkeys.ProcNatlAcadSciUSA2002;99:15735–42.MacVicarBA,HochmanD.Imagingofsynapticallyevokedintrinsic

opticalsignalsinhippocampalslices.JNeurosci1991;11:1458–69.MalachR,TootellRB,MalonekD.Relationshipbetweenorientationdo-mains,cytochromeoxidasestripes,andintrinsichorizontalconnec-tionsinsquirrelmonkeyareaV2.CerebCortex1994;4:151–65.MalachR,AmirY,HarelM,GrinvaldA.Relationshipbetweenintrinsic

connectionsandfunctionalarchitecturerevealedbyopticalimagingandinvivotargetedbiocytininjectionsinprimatestriatecortex.ProcNatlAcadSciUSA1993;90:10469–73.

MalachR,SchirmanTD,HarelM,TootellRB,MalonekD.Organiza-tionofintrinsicconnectionsinowlmonkeyareaMT.CerebCortex1997;7:386–93.

MaldonadoPE,GödeckeI,GrayCM,BonhoefferT.Orientationselectivity

inpinwheelcentersincatstriatecortex.Science1997;276:1551–5.MalonekD,TootellRB,GrinvaldA.Opticalimagingrevealsthefunctional

architectureofneuronsprocessingshapeandmotioninowlmonkeyareaMT.ProcRSocLondBBiolSci1994;258:109–19.

MalonekD,DirnaglU,LindauerU,YamadaK,KannoI,GrinvaldA.Vas-cularimprintsofneuronalactivity:relationshipsbetweenthedynam-icsofcorticalbloodﬂow,oxygenation,andvolumechangesfollowingsensorystimulation.ProcNatlAcadSciUSA1997;94:14826–31.

MasinoSA,FrostigRD.Quantitativelong-termimagingofthefunctional

representationofawhiskerinratbarrelcortex.ProcNatlAcadSciUSA1996;93:4942–7.

MasinoSA,KwonMC,DoryY,FrostigRD.Characterizationoffunc-tionalorganizationwithinratbarrelcortexusingintrinsicsignalop-ticalimagingthroughathinnedskull.ProcNatlAcadSciUSA1993;90:9998–10002.

MayhewJE,AskewS,ZhengY,PorrillJ,WestbyGW,RedgraveP,Rector

DM,HarperR.Cerebralvasomotion:a0.1Hzoscillationinreﬂectedlightimagingofneuralactivity.NeuroImage1996;4:183–93.

McLoughlinNP,BlasdelGG.Wavelength-dependentdifferencesbetween

opticallydeterminedfunctionalmapsfrommacaquestriatecortex.NeuroImage1998;7:326–36.

MeisterM,BonhoefferT.Tuningandtopographyinanodormaponthe

ratolfactorybulb.JNeurosci2001;21:1351–60.

MerzenichMM,KnightPL,RothGL.Cochleotopicorganizationofpri-maryauditorycortexinthecat.BrainRes1973;63:343–6.

MerzenichMM,KnightPL,RothGL.Representationofcochleawithin

primaryauditorycortexinthecat.JNeurophysiol1975;38:231–49.MombaertsP.Molecularbiologyofodorantreceptorsinvertebrates.Annu

RevNeurosci1999;22:487–509.

NarayanSM,SantoriEM,TogaAW.Mappingfunctionalactivityinrodent

cortexusingopticalintrinsicsignals.CerebCortex1994a;4:195–204.NarayanSM,SantoriEM,BloodAJ,BurtonJS,TogaAW.Imagingoptical

reﬂectanceinrodentbarrelandforelimbsensorycortex.NeuroImage1994b;1:181–90.

NarayanSM,EsfahaniP,BloodAJ,SikkensL,TogaAW.Functional

increasesincerebralbloodvolumeoversomatosensorycortex.JCerebBloodFlowMetab1995;15:754–65.

NelsonRJ,SurM,FellemanDJ,KaasJH.Representationofthebody

surfaceinpostcentralparietalcortexofMacacafascicularis.JCompNeurol1980;192:611–43.

PayneBR,BermanN,MurphyEH.Organizationofdirectionpreferences

incatvisualcortex.BrainRes1981;211:445–50.

PetersonBE,GoldreichD.Opticalimagingandelectrophysiologyofrat

barrelcortex.PartI.Responsestosmallsingle-vibrissadeﬂections.CerebCortex1998;8:173–83.

PetersenCCH,GrinvaldA,SakmannB.Spatiotemporaldynamicsof

sensoryresponsesinlayer2/3ofratbarrelcortexmeasuredinvivobyvoltage-sensitivedyeimagingcombinedwithwhole-cellvoltagerecordingsandneuronreconstructions.JNeurosci2003;23:1298–309.PolleyDB,Chen-BeeCH,FrostigRD.Twodirectionsofplasticityinthe

sensory-deprivedadultcortex.Neuron1999;24:623–37.

PonsTP,GarraghtyPE,CusickCG,KaasJH.Asequentialrepresen-tationoftheocciput,arm,forearmandhandacrosstherostrocau-daldimensionofareas1,2and5inmacaquemonkeys.BrainRes1985;335:350–3.

PonsTP,WallJT,GarraghtyPE,CusickCG,KaasJH.Consistentfeatures

oftherepresentationofthehandinarea3bofmacaquemonkeys.SomatosensRes1987;4:309–31.

PouratianN,BookheimerSY,O’FarrellAM,SicotteNL,CannestraAF,

BeckerD,TogaAW.Opticalimagingofbilingualcorticalrepresen-tations.Casereport.JNeurosurg2000;93:676–81.

PouratianN,ShethSA,MartinNA,TogaAW.Sheddinglighton

brainmapping:advancesinhumanopticalimaging.TrendsNeurosci2003;26:277–82.

PrakashN,Cohen-CoryS,FrostigRD.Rapidandoppositeeffectsof

BDNFandNGFonthefunctionalorganizationoftheadultcortexinvivo.Nature1996;381:702–6.

RatzlaffEH,GrinvaldA.Atandem-lensepiﬂuorescencemacroscope:

hundred-foldbrightnessadvantageforwide-ﬁeldimaging.JNeurosciMethods1991;36:127–37.

RoseJE.Thecellularstructureoftheauditorycortexofthecat.JComp

Neurol1949;91:409–40.

RubinBD,KatzLC.Opticalimagingofodorantrepresentationsinthe

mammalianolfactorybulb.Neuron1999;23:499–511.

20A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–21

SatoK,NariaiT,SasakiS,YazawaI,MochidaH,MiyakawaN,

Momose-SatoY,KaminoK,OhtaY,HirakawaK,OhnoK.Intra-operativeintrinsicopticalimagingofneuronalactivityfromsubdi-visionsofthehumanprimarysomatosensorycortex.CerebCortex2002;12:269–80.

SchuettS,BonhoefferT,HübenerM.Mappingofretinotopyinratvisual

cortexbycombinedlinearextractionandprinciplecomponentanalysisofopticalimagingdata.EurJNeurosci2000;12(Suppl11):74.

SchuettS,BonhoefferT,HübenerM.Pairing-inducedchangesoforien-tationmapsincatvisualcortex.Neuron2001;32:325–37.

SchuettS,BonhoefferT,HübenerM.Mappingretinotopicstructurein

mousevisualcortexwithopticalimaging.JNeurosci2002;22:6549–59.

SchulzeS,FoxK.Singlewhiskerrepresentationinthebarrelcortexof

ratsvisualisedwithopticalimagingofintrinsicsignals.SocNeurosciAbstr2000;26:1687.

SchwartzTH,BonhoefferT.Invivoopticalmappingofepilepticfociand

surroundinhibitioninferretcerebralcortex.NatMed2001;7:1063–7.SengpielF,BonhoefferT.Orientationspeciﬁcityofcontrastadaptation

invisualcorticalpinwheelcentresandiso-orientationdomains.EurJNeurosci2002;15:876–86.

SengpielF,StawinskiP,BonhoefferT.Inﬂuenceofexperienceonorien-tationmapsincatvisualcortex.NatNeurosci1999;2:727–32.

SengpielF,GödeckeI,StawinskiP,HübenerM,LöwelS,BonhoefferT.

Intrinsicandenvironmentalfactorsinthedevelopmentoffunctionalmapsincatvisualcortex.Neuropharmacology1998;37:607–21.

SharpFR,KauerJS,ShepherdGM.Localsitesofactivity-relatedglucose

metabolisminratolfactorybulbduringolfactorystimulation.BrainRes1975;98:596–600.

SharpFR,KauerJS,ShepherdGM.Laminaranalysisof2-deoxyglucose

uptakeinolfactorybulbandolfactorycortexofrabbitandrat.JNeurophysiol1977;40:800–13.

ShethS,NemotoM,GuiouM,WalkerM,PouratianN,TogaAW.

Evaluationofcouplingbetweenopticalintrinsicsignalsandneuronalactivityinratsomatosensorycortex.NeuroImage2003;19:884–94.ShmuelA,GrinvaldA.Functionalorganizationfordirectionofmotion

anditsrelationshiptoorientationmapsincatarea18.JNeurosci1996;16:6945–.

ShmuelA,GrinvaldA.Coexistenceoflinearzonesandpinwheelswithin

orientationmapsincatvisualcortex.ProcNatlAcadSciUSA2000;97:5568–73.

ShohamD,GrinvaldA.Thecorticalrepresentationofthehandinmacaque

andhumanareaS-I:highresolutionopticalimaging.JNeurosci2001;21:6820–35.

ShohamD,GlaserDE,ArieliA,KenetT,WijnbergenC,ToledoY,

HildesheimR,GrinvaldA.Imagingcorticaldynamicsathighspatialandtemporalresolutionwithnovelbluevoltage-sensitivedyes.Neuron1999;24:791–802.

ShohamD,HübenerM,SchulzeS,GrinvaldA,BonhoefferT.

Spatio-temporalfrequencydomainsandtheirrelationtocytochromeoxidasestainingincatvisualcortex.Nature1997;385:529–33.

ShtoyermanE,ArieliA,SlovinH,VanzettaI,GrinvaldA.Long-term

opticalimagingandspectroscopyrevealmechanismsunderlyingtheintrinsicsignalandstabilityofcorticalmapsinV1ofbehavingmonkeys.JNeurosci2000;20:8111–21.

SiegelRM,RafﬁM,PhinneyRE,TurnerJA,JandoG.Functionalar-chitectureofeyepositiongainﬁeldsinvisualassociationcortexofbehavingmonkey.JNeurophysiol2003;90:1279–94.

SingerW.Topographicorganizationoforientationcolumnsinthecat

visualcortex.Adeoxyglucosestudy.ExpBrainRes1981;44:431–6.SingerW,FreemanB,RauscheckerJ.Restrictionofvisualexperience

toasingleorientationaffectstheorganizationoforientationcolumnsincatvisualcortex.Astudywithdeoxyglucose.ExpBrainRes1981;41:199–215.

SirovichL,EversonRM.Managementandanalysisoflargescientiﬁc

datasets.IntJSupercompAppl1992;6:50–68.

SpitzerMW,CalfordMB,ClareyJC,PettigrewJD,RoeAW.Spontaneous

andstimulus-evokedintrinsicopticalsignalsinprimaryauditorycortexofthecat.JNeurophysiol2001;85:1283–98.

StetterM,SchiesslI,OttoT,SengpielF,HübenerM,BonhoefferT,Ober-mayerK.Principalcomponentanalysisandblindseparationofsourcesforopticalimagingofintrinsicsignals.NeuroImage2000;11:482–90.StepnoskiRA,LaPortaA,Raccuia-BehlingF,BlonderGE,SlusherRE,

KleinfeldD.Noninvasivedetectionofchangesinmembranepotentialinculturedneuronsbylightscattering.ProcNatlAcadSciUSA1991;88:9382–6.

SurM,NelsonRJ,KaasJH.Representationsofthebodysurfacein

corticalareas3band1ofsquirrelmonkeys:comparisonswithotherprimates.JCompNeurol1982;211:177–92.

SwindaleNV.Howmanymapsarethereinvisualcortex?CerebCortex

2000;10:633–43.

SwindaleNV,GrinvaldA,ShmuelA.Thespatialpatternofresponse

magnitudeandselectivityfororientationanddirectionincatvisualcortex.CerebCortex2003;13:225.

SwindaleNV,ShohamD,GrinvaldA,BonhoefferT,HübenerM.Vi-sualcortexmapsareoptimizedforuniformcoverage.NatNeurosci2000;3:822–6.

TakashimaI,KajiwaraR,IijimaT.Voltage-sensitivedyeversusintrinsic

signalopticalimaging:comparisonofopticallydeterminedfunctionalmapsfromratbarrelcortex.NeuroReport2001;12:2888–94.

TogaAW,CannestraAF,BlackKL.Thetemporal/spatialevolutionof

opticalsignalsinhumancortex.CerebCortex1995;5:561–5.

TolhurstDJ,DeanAF,ThompsonID.Preferreddirectionofmovement

asanelementintheorganizationofcatvisualcortex.ExpBrainRes1981;44:340–2.

TommerdahlM,FavorovO,WhitselBL,NakhleB,GoncharYA.Mini-columnaractivationpatternsincatandmonkeySIcortex.CerebCor-tex1993;3:399–411.

TommerdahlM,DelemosKA,VierckCJJ,FavorovOV,WhitselBL.

Anteriorparietalcorticalresponsetotactileandskin-heatingstimuliappliedtothesameskinsite.JNeurophysiol1996;75:2662–70.

TommerdahlM,DelemosKA,WhitselBL,FavorovOV,MetzCB.Re-sponseofanteriorparietalcortextocutaneousﬂutterversusvibration.JNeurophysiol1999a;82:16–33.

TommerdahlM,WhitselBL,FavorovOV,MetzCB,O’QuinnBL.Re-sponsesofcontralateralSIandSIIincattosame-sitecutaneousﬂutterversusvibration.JNeurophysiol1999b;82:1982–92.

TommerdahlM,DelemosKA,FavorovOV,MetzCB,VierckCJJ,Whitsel

BL.Responseofanteriorparietalcortextodifferentmodesofsame-siteskinstimulation.JNeurophysiol1998;80:3272–83.

TootellRB,SilvermanMS,DeValoisRL.Spatialfrequencycolumnsin

primaryvisualcortex.Science1981;214:813–5.

Ts’oDY,GilbertCD,WieselTN.Relationshipsbetweenhorizontalin-teractionsandfunctionalarchitectureincatstriatecortexasrevealedbycross-correlationanalysis.JNeurosci1986;6:1160–70.

Ts’oDY,RoeAW,GilbertCD.Ahierarchyofthefunctionalorganization

forcolor,formanddisparityinprimatevisualareaV2.VisionRes2001;41:1333–49.

Ts’oDY,FrostigRD,LiekeEE,GrinvaldA.Functionalorganization

ofprimatevisualcortexrevealedbyhighresolutionopticalimaging.Science1990;249:417–20.

TsunodaK,YamaneY,NishizakiM,TanifujiM.Complexobjectsare

representedinmacaqueinferotemporalcortexbythecombinationoffeaturecolumns.NatNeurosci2001;4:832–8.

UchidaN,TakahashiYK,TanifujiM,MoriK.Odormapsinthemam-malianolfactorybulb:domainorganizationandodorantstructuralfeatures.NatNeurosci2000;3:1035–43.

VanzettaI,GrinvaldA.Increasedcorticaloxidativemetabolismdueto

sensorystimulation:implicationsforfunctionalbrainimaging.Science1999;286:1555–8.

VersnelH,MossopJE,Mrsic-FlogelTD,AhmedB,MooreDR.Optical

imagingofintrinsicsignalsinferretauditorycortex:responsestonarrowbandsoundstimuli.JNeurophysiol2002;88:1545–58.

A.Zepedaetal./JournalofNeuroscienceMethods136(2004)1–21

VillringerA,ChanceB.Non-invasiveopticalspectroscopyandimaging

ofhumanbrainfunction.TrendsNeurosci1997;20:435–42.

VillringerA,DirnaglU.Couplingofbrainactivityandcerebralblood

ﬂow:basisoffunctionalneuroimaging.CerebrovascBrainMetabRev1995;7:240–76.

VnekN,RamsdenBM,HungCP,Goldman-RakicPS,RoeAW.Optical

imagingoffunctionaldomainsinthecortexoftheawakeandbehavingmonkey.ProcNatlAcadSciUSA1999;96:4057–60.

WangG,TanakaK,TanifujiM.Opticalimagingoffunctionalorganization

inthemonkeyinferotemporalcortex.Science1996;272:1665–8.WelikyM,BoskingWH,FitzpatrickD.Asystematicmapofdirection

preferenceinprimaryvisualcortex.Nature1996;379:725–8.

WhiteLE,CoppolaDM,FitzpatrickD.Thecontributionofsensory

experiencetothematurationoforientationselectivityinferretvisualcortex.Nature2001;411:1049–52.

WoolseyCN,MarshallWH,BardP.Representationofcutaneoustactile

sensibilityinthecerebralcortexofthemonkeyasindicatedbyevokedpotentials.BullJohnsHopkinsHosp1942;70:399–441.

WoolseyTA,VanderLoosH.ThestructuralorganizationoflayerIVinthe

somatosensoryregion(SI)ofmousecerebralcortex:Thedescription

ofacorticalﬁeldcomposedofdiscretecytoarchitectonicunits.BrainRes1970;17:205–42.

XiaoY,WangY,FellemanDJ.Aspatiallyorganizedrepresentationof

colourinmacaquecorticalareaV2.Nature2003;421:535–9.

YangX,RenkenR,HyderF,SiddeekM,GreerCA,ShepherdGM,

ShulmanRG.Dynamicmappingatthelaminarlevelofodor-elicitedresponsesinratolfactorybulbbyfunctionalMRI.ProcNatlAcadSciUSA1998;95:7715–20.

YousefT,BonhoefferT,KimDS,EyselUT,TothE,KisvardayZF.

Orientationtopographyoflayer4lateralnetworksrevealedbyopticalimagingincatvisualcortex(area18).EurJNeurosci1999;11:4291–308.

ZepedaA,VacaL,AriasC,SengpielF.Reorganizationofvisualcor-ticalmapsafterfocalischemiclesions.JCerebBloodFlowMetab2003;23:811–20.

ZuritaP,VillaAE,deRibaupierreY,deRibaupierreF,RouillerEM.

Changesofsingleunitactivityinthecat’sauditorythalamusandcortexassociatedtodifferentanestheticconditions.NeurosciRes1994;19:303–16.

因篇幅问题不能全部显示，请点此查看更多更全内容

查看全文