Transcript
SimultaneousRecordingofEvokedPotentialsandT * 2 -WeightedMRImagesDuringSomatosensoryStimulationofRat GerritBrinker,ChristianBock,ElmarBusch,HenningKrep,Konstantin-AlexanderHossmann,andMathiasHoehn-Berlage Somatosensoryevokedpotentials(SEP)andT* 2 -weightednuclearmagnetic resonance (NMR) images were recorded simulta-neously during somatosensory stimulation of rat to investigatethe relationship between electrical activation of the brain tissueand the signal intensity change in functional NMR imaging.Electrical forepaw stimulation was performed in Wistar ratsanesthetized with -chloralose. SEPs were recorded with calo-mel electrodes at stimulation frequencies of 1.5, 3, 4.5, and 6 Hz.At the same time, T* 2 -weighted imaging was performed, and thesignal intensity increase during stimulation was correlated withthe mean amplitude of the SEP. Both the stimulation-evokedsignal intensity increase in T* 2 -weighted images and the ampli-tude of SEPs were dependent on the stimulation frequency, withthe largest signals at a stimulation frequency of 1.5 Hz anddecreasingactivationswithincreasingfrequencies.Thefeasibil-ity of simultaneous, artifact-free recordings of T* 2 -weighted NMRimages and of evoked potentials is proved. Furthermore, thestudy demonstrates—in the intact brain—the validity of func-tional magnetic resonance imaging for estimating the intensityof electrocortical activation. Magn Reson Med 41:469–473, 1999. 1999Wiley-Liss,Inc.Key words: fMRI; somatosensory cortex; evoked potentials;electrophysiology; rat Inrecentyearsnuclearmagneticresonance(NMR)imaginghas gained widespread use in the investigation of func-tional brain activation in both humansand animal experi-ments(1–5).Itiswidelyacceptedthattheincreasedenergydemandsof theactivated brain tissueresultin an increasein local cerebral blood flow (CBF), although the exactmechanismsofcouplingbetweenCBFandneuronalactiva-tion arestill under discussion (6–11). Theamountof CBFincrease during somatosensory stimulation exceeds theincrease in cerebral metabolic rate of oxygen (CMRO 2 )(12,13), with theconsequencethat thenet oxygen extrac-tion fraction decreases. This results in decreased localconcentration of deoxyhemoglobin, which is paramag-netic,andthereforeinachangeofthemagneticsusceptibil-ity of theblood. Asaconsequence, thesignal intensity in T* 2 -weighted MR images increases, a mechanism that hasbeen referred to as blood oxygenation level-dependent(BOLD) contrast (14,15). T* 2 -weighted MR imaging se-quences are widely used for the mapping of corticalfunctions in humans (16,17) and more recently also inanimals(18,19).Thefunctional responsetosomatosensorystimulation has been shown by perfusion-weighted imag-ingwith visualization of both thespatial distribution andthetemporal pattern of theactivation-induced changesinblood flow (20,21). However, itmustbekeptin mind thatimaging of functional activation relies on the intact cou-pling with blood flow and metabolism and, thereforereflects neuronal activity only indirectly. For a moreprecise evaluation, functional MR imaging should there-forebecompared with theelectrophysiological response. The present investigation demonstrates for the first timethefeasibilityofsimultaneousmeasurementsofsomatosen-sory evoked cortical potentials (SEPs) and T* 2 -weightedimagesintherat. MATERIALS AND METHODS Animal Preparation Five male Wistar rats (350–400 gBW) were anesthetizedwith 1.2% halothane in a 70%/30% mixture of N 2 O/O 2 .Body temperature was maintained at 37 C via a rectalthermometer and afeedback-controlled warmwater blan-ket. Two PE 50catheterswereplaced in theright femoralarteryandfemoral vein,todraw bloodsamplesforarterialblood gasanalysisand to infusedrugs, respectively. Aftertracheotomy thetracheawas cannulated and theanimalswereartificiallyventilatedtokeeppCO 2 inthephysiologi-cal range.Pancuroniumbromide(0.2mg/kg/hr)wasgivenformusclerelaxation.After placingtheanimalspronein aPlexiglasstereotac-ticheadholder,theskin and theperiosteumover theskullwere medially incised and retracted. A small nylon nutwasgluedtotheexposednasal boneandanotheronetotheskull at the level of the somatosensory cortex (1 mmanterior to the bregma). The remainingskull surface wascovered with dental cement to avoid NMR imaging arti-factscausedbysusceptibilitychanges.Forelectroencepha-lographic(EEG)recordinginsidethemagnet,twoL-shapedcalomel electrodeswerepreparedasdescribedbefore(22). The electrodes were introduced with their tip into eachnut,withtheactiveelectrodeatthelevel ofthesomatosen-sory cortex and the reference electrode above the nasalbone, and fixed at thestereotactic holder. With thesetupdescribed, electrophysiological recordinginsidethemag-netispossiblewithminimal artifacts.Sixty minutes before MR measurements, -chloralose(80mg/kg)wasgiven intravenously and supplemented by40mg/kgi.v.at90minintervals.Halothaneanesthesiawasdiscontinued, and N 2 O was substituted by N 2 . A pair of needleelectrodes was introduced into theskin of theleft Max-Planck-Institute for Neurological Research, Cologne, Germany.Grantsponsor:DeutscheForschungsgemeinschaft;Grantnumber:SFB194/B1.*Correspondence to: Dr. Mathias Hoehn-Berlage, In-vivo-NMR, Max-Planck-Institutfu¨rneurologischeForschung,GleuelerStr.50,D-50931Ko¨ln,Germany.E-mail:
[email protected] 20 March 1998; revised 30 July 1998; accepted 10August 1998. MagneticResonanceinMedicine41:469–473(1999)469 1999 Wiley-Liss, Inc. forepaw.Rectangularelectrical pulsesof0.3msecdurationand 0.5 mA intensity were applied at frequencies of 1.5,3.0,4.5,or6.0Hzforadurationof50sec.Thepulsesweredelivered byaconstantcurrentpower supply(FMI,Egels-bach, Germany). In addition to this protocol, in oneexperimentthelefthindpawwasstimulatedat1.5Hzwithelectrical pulsesof0.3msecand1.2mA intensity. NMR Methods NMR measurements were performed on a 4.7 T BrukerBIOSPEC MSL-X11 system (Bruker, Ettlingen, Germany)equippedwithactivelyshieldedgradientcoils(100mT/m,rise time 250 sec). Radiofrequency (RF) pulses weretransmitted usinga12cmdiameter Helmholtz coil, whilesignalswerereceived with a16mmdiameter inductivelycoupledsurfacecoil centeredoverthebregma.MRimageswere acquired using the fast low-angle shot (FLASH)sequence(23)withafieldofview of4cm.Multislicepilotscansof thebrain anatomy (flip angle 22.5 , TE 8msec, TR 400 msec) were performed to position the functionalimageslice. Single-sliceheavily T* 2 -weighted images (24)( 22.5 ,TE60msec,TR70msec,slicethickness2mm,64phaseencodingsteps,scan time4.5sec)wererecordedforfunctional imaging. Experimental Protocol Based on the sagittal pilot scans, the functional imagingslicewaspositioned coronallythrough thesomatosensorycortex1.0mmanteriortothebregma.EachsetofactivationexperimentsincludedtheacquisitionofeightT* 2 -weightedimages(T* 2 -WI) prior to stimulation (baselineimages) andeight images during electrical stimulation of paws. Thisprocedure was repeated in each animal for the variousstimulationfrequencies,witharestperiodofatleast5minbetweeneachstimulationsequence.Toavoidmisinterpre-tationsduetohabituation effects, theorder of stimulationfrequencies was varied. EEG recording artifacts due todesiccation of the electrodes were prevented by refillingthesalinesolution in thenylon nutsseveral timesduringthecourseoftheexperiment. Data Processing T* 2 -WI activation maps were obtained by subtracting theeight averaged baseline images from the eight averagedstimulation images of each stimulation sequence. Forlocalization of the functionally activated brain region,pixels with an intensity of morethan 1.5 standard devia-tions above the noise level of subtracted images wereoverlaidontothecorrespondinganatomical picture.Quan-titative analysis of functional (f)MRI was performed bymeasuringtheT* 2 -WIsignal intensityincreaseinaregionof interestplacedinthecenteroftheactivatedsomatosensorycortex.EEGs and evoked electrical responses were recordedusingalaboratorycomputer.Forepawswerestimulated atincreasingfrequencies(1.5,3,4.5,and6Hz,respectively),and thecortical evoked potentialswereaveraged over thetotal stimulation period of 50sec, resultingin averagesof 75, 150, 225, and 300 single responses, respectively. Thefrequency dependent changes of the peak-to-peak ampli-tude (difference between P1- and N1-signals) were ex-pressedinpercentofthemeanamplitudeat1.5Hz.Statistical analysis (Student’s t -test, unpaired, unequalvariances)wasperformedbycomparisonofthenormalizedpeak-to-peakSEPamplitudeandtheT* 2 -WIsignal intensityincrease, respectively, obtained at 1.5 Hz with the corre-spondingvaluesathigherstimulationfrequencies. RESULTS Physiological variables were kept within normal rangeduringthewholelength of experiments: arterial PO 2 was135 26 mmHg, arterial CO 2 was 37.5 4.2 mmHg,arterial pH was 7.37 0.05, and themean arterial bloodpressurewas105 9mmHg(all valuesaremeans SD).Duringelectrical stimulationphysiological variables,nota-blysystemicarterial pressure,didnotchange. The novel experimental setup allowed simultaneous,artifact-free measurements of SEP and T* 2 -WI (Fig. 1).ArtifactsinMRimages,whenpresent,wereduetoinhomo-geneitiesofsusceptibilityproducedbysmall airbubblesinthe dental cement or the nylon nuts, or they appearedwhentheelectrodeswereplacedontheskull withtoohighpressure.Theycouldbeeliminatedbyexchangeofcementor revision of electrodeplacement. Gross disturbances of theelectrical signal werevisibleduringgradientswitchingbutdisappearedalmostcompletelyonaveraging. The SEPs consisted of an early negative deflection(N1-wave) with a peak latency between 14 and 18 msec,followed by a later positive deflection (P1-wave) with alatency of about 40 ms (Fig. 1, right column). N2-wavesweredetectablein somecases. Theamplitudeof theSEPwasclearly dependenton stimulation frequency, with thelargestamplitudeatastimulation frequencyof 1.5Hz anddecreasingamplitudeswithincreasingfrequency.Electrical stimulation of theleft forepaw resulted in anincreasein T* 2 -WI signal intensity in theright somatosen-sory cortex (Fig. 1, left column). The signal intensityincrease,expressedinpercentoftherestinglevel,wasalsodependent on thestimulation frequency, with thehighestincrease at 1.5 Hz (5.91 0.83%, mean SEM) and de-creasing signal intensity increases with increasing fre-quency(3.42 0.58% at3Hz,1.56 0.56% at4.5Hz,and1.19 0.22% at 6Hz). Similarly, thepeak-to-peak ampli-tude of the SEP was largest at 1.5 Hz and graduallydeclinedwithincreasingfrequencies(Fig.1).InoneexperimentaT* 2 -WI signal intensityincreasewasdocumented during stimulation of the left hindpaw, butthe current required for activation was higher (0.3 msec,1.2 mA). Thecortical areaactivated by hindpaw stimula-tionwaslocatedinaslice1.5mmcaudallyfromthatoftheforepaw, i.e., the area corresponding to the anatomicalrepresentationofthehindpawintheratcerebralsomatosen-sorycortex(25)(Fig.2).In individual experiments, thecorrelation between SEPamplitude and T* 2 -WI signal intensity change at varyingstimulationfrequenciesresultedincorrelationcoefficientsranging between 0.61 and 0.85 (example of one animalgiven in Fig. 3). For overall characterization of the fre-quencydependenceofSEPandT* 2 -WI,theSEPamplitudesofindividual animalswerenormalizedtothemeanampli-tude at a stimulation rate of 1.5 Hz. The two variables 470 Brinkeretal. revealed a clearly similar behavior, namely, decreasingsignals with increasing stimulation frequencies (Fig. 4:5.91 0.83%at1.5Hz,3.42 0.58%at3Hz,1.56 0.56%at 4.5 Hz, and 1.19 0.22% at 6 Hz for BOLD signalintensity increase; 100 7% at 1.5 Hz, 52 2% at 3 Hz,34 5% at 4.5 Hz, and 22 2% at 6 Hz for normalizedSEP amplitude; all values mean SEM). The values re-corded at 1.5 Hz differed significantly from the higherstimulation frequencies, both for theelectrophysiologicalandfortheMRdata. DISCUSSION Theamplitudeofevokedcortical potential depends,inthefirstplace,ontheintensityoftheafferentcortical input.At FIG. 1. Simultaneous acquisition of activation maps (left) and ofsomatosensory evoked potentials (right) at increasing stimulationfrequencies (1.5, 3, 4.5, and 6 Hz) during left forepaw stimulation.Activation maps were derived from T* 2 -weighted images and overlaidon the corresponding anatomical image. The signal intensity in-crease is highest at 1.5 Hz (5.9 0.8% above background,mean SEM) and decreases at increasing frequencies. Similarly,theamplitudeoftheSEPislargestat1.5Hzandgraduallydecreasesat increasing frequencies. Note the position of the active electrode inthe same image plane as that of the activated cortical field.FIG. 2. Activation maps (left) and corresponding somatosensoryevoked potentials (right) during stimulation at 1.5 Hz of left forepaw(top) and hindpaw (bottom). The planes in which cortical activity arevisible are separated by 1.5 mm. Note the increased latency of theSEPandthemedio-caudalshiftoftheactivatedcorticalregionduringhindpaw stimulation as apparent by the relative shift of the epicenterof the activated regions for the representation of the forepaw andhindpaw in the somatosensory cortex (34).FIG. 3. Correlation of the increase of T* 2 -weighted imaging signalintensity with the peak-to-peak amplitude of the somatosensoryevoked potential (SEP) during forepaw stimulation at increasingfrequencies (data are from one individual animal; r 0.82). Evoked Potentials and T* 2 -Weighted Imaging 471 a given stimulation intensity, it also reflects the level of oxygen consumption, at least under conditions of de-creased metabolic rate. In dogs that underwent hypoxiaand reoxygenation,thepeak-to-peak amplitudeof theSEPcorrelatedwithCMRO 2 ,irrespectiveofbloodflowchanges(26), and in humans metabolic and electrophysiologicalchanges paralleled each other duringcontrolled hypoten-sion (27). On the other hand, activation of the neuronalnetwork,which underliesthegeneration of evoked poten-tials, results in a coupled increase of glucose and—to alesser degree—oxygen consumption, which riseswith theintensity of the stimulation. It is therefore reasonable toassumethattheamplitudeofevokedpotentialsisamarkerofthemetabolicrateintheareainwhichthepotentialsaregenerated.Changes in metabolic activity are also reflected bychanges in T* 2 -WI signal intensity. A mathematical modeloftherelationshipbetweenCMRO 2 ,bloodflow andT* 2 -WIsignal intensityincreasewasrecentlypresentedbyBuxtonandFrank(15).Accordingtotheiranalysis,theincreaseof blood flow during stimulation is tightly coupled to themetabolic demand, although in a nonlinear relationship. Thus a small increase in metabolism requires a largeincreaseinbloodflow tocompensateforthehigherenergydemands. In their proposed model, the relation betweenblood flow and MR signal intensity increase is nearlylinear. Therefore, differences in the metabolic demandduringdifferent stimulation paradigms should havetheircounterpartsin parallel changesof thecorrespondingMRsignalintensity.ThisissupportedbyarecentNMRspectros-copy study that demonstrated parallel changes in T* 2 -WIsignal andoxidativeglucosemetabolismduringfunctionalactivationinrat(28). The dependence of T* 2 -WI signal intensity on the fre-quency of somatosensory stimulation was first describedby Gyngell et al in amodel similar to that in thepresentstudy (24). Although in our study theT* 2 -WI signal eleva-tions were somewhat smaller, the dependence of the MRsignal on thestimulation frequency was similar, with thehighest signal increaseat 1.5 Hz and agradual declineof signals at increasing frequencies (24). The most likelyexplanation for thefrequency-dependentsignal declineistheocclusionoftheelectrophysiologicalresponseathigherstimulation rates. In fact, in an earlier study of forepawstimulationincat,eachstimulusevokedamaximalelectro-cortical responseat low frequencies (2–3 Hz), whereas athigher frequenciesevery second or third evoked responsewassuppressed,resultinginasmalleraveragedSEPampli-tude(29). Theseauthors also showed that theincreaseinmicroflow declined with increasing stimulation frequen-cies, which is in line with the interpretation that thehemodynamic-metabolic responseiscoupled tothemeta-bolicdemandsoftheactivatedtissue.TheSEP amplitude,therefore, reflects the local cortical metabolic activation,which in turn can bedetected by recordingeither micro-floworT* 2 -WI signal intensity.Obviously, thechangesin T* 2 -WI duringfunctional acti-vation rely on theintactcouplingamongelectrical activa-tion,oxygenconsumption,andbloodflow.Inpathologicalsituations such as after global ischemia (30), this is notalwaysthecase.Afterprolongedglobal forebrainischemia,producedbythefour-vessel occlusionmodel inrat,recov-ery of electrical activity was much faster than that of thehemodynamic-metabolicresponse(31).Ontheotherhand,resuscitation after 10 min cardiac arrest in rats led to theparallel normalization of SEP and blood flow increaseduringelectrical forepaw activation (32,33). Theapplica-tionofT* 2 -WI forevaluationofpost-ischemicrecoverymaythereforeleadtodifferentresults,dependingonthepreser-vation of thecouplingmechanisms.Toavoid misinterpre-tations,thesimultaneousrecordingoftheelectrophysiologi-cal responseis,therefore,recommended. CONCLUSIONS Our study demonstrates the feasibility of simultaneouselectrophysiological and functional MRI measurementsandprovidesanew approachforstudyingtherelationshipbetween electrophysiological activity and MR-visiblechanges in blood flow and blood oxygenation. In thehealthybrain,theT* 2 -WI signal intensityincreaseslinearly FIG. 4. Increase of signal intensity in T* 2 -weighted images (top) andnormalized SEP amplitudes (bottom) at increasing stimulation fre-quencies. SEPamplitudes were normalized to the 1.5 Hz value. Notethe similar frequency dependence of MR and SEP (mean SEM,statistically significant difference from 1.5 Hz: *, P 0.05, **, P 0.01). 472 Brinkeretal. withtheSEP amplitudes,butthisrelationshipmaybelostunder pathological conditions. The simultaneous record-ing of SEPs together with functional MRI facilitates theinterpretationofsuchconditionsandprovidesevidenceonthehemodynamic-metabolic couplingof functional activ-ity. REFERENCES 1. AsheJ, Ugurbil K. Functional imagingof themotor system. Curr OpinNeurobiol 1994;4:832–839.2. Hyder F, Behar KL, Martin MA, Blamire AM, Shulman RG. Dynamicmagnetic resonance imagingof the rat brain duringforepaw stimula-tion.J CerebBloodFlowMetab1994;14:649–655.3. Le Bihan D, Jezzard P, Haxby J, Sadato N, Rueckert L, Mattay V.Functional magnetic resonanceimagingof thebrain. Ann Intern Med1995;122:296–303.4. KleinschmidtA,LeeBB,RequardtM,FrahmJ.Functional mappingof color processing by magnetic resonance imaging of responses toselectiveP- and M-pathway stimulation. Exp Brain Res1996;110:279–288.5. Scanley BE, Kennan RP, Cannan S, Skudlarski P, Innis RB, Gore JC.Functional magneticresonanceimagingofmediannervestimulationinratsat2.0T.MagnResonMed1997;37:969–972.6. Sokoloff L. Relationships among local functional activity, energymetabolism, and blood flow in the central nervous system. Fed Proc1981;40:2311–2316.7. Lou HC, Edvinsson L, MacKenzieET. Theconcept of couplingbloodflow to brain function: revision required? Ann Neurol 1987;22:289–297.8. Villringer A, Dirnagl U. Couplingof brain activity and cerebral bloodflow: basis of functional neuroimaging. Cerebrovasc Brain Metab Rev1995;7:240–276.9. Jueptner M,Weiller C.Review:Doesmeasurementof regional cerebralblood flow reflect synaptic activity?—Implications for PET and fMRI.Neuroimage1995;2:148–156.10. Lindauer U, Megow D, Schultze J, Weber JR, Dirnagl U. Nitric oxidesynthaseinhibitiondoesnotaffectsomatosensoryevokedpotentialsintherat.Neurosci Lett1996;216:207–210.11. Wolf T, Lindauer U, Villringer A, Dirnagl U. Excessive oxygen orglucosesupplydoesnotalterthebloodflowresponsetosomatosensorystimulation or spreadingdepression in rats. Brain Res 1997;761:290–299.12. Fox PT, RaichleME. Focal physiological uncouplingof cerebral bloodflow and oxidative metabolism during somatosensory stimulation inhumansubjects.ProcNatl AcadSci USA 1986;83:1140–1144.13. Fox PT, Raichle ME, Mintun MA, Dence C. Nonoxidative glucoseconsumptionduringfocalphysiologicneuralactivity.Science1988;241:462–464.14. Ogawa S, Tank DW, Menon R, Ellermann JM, Kim SG, Merkle H,Ugurbil K.Intrinsicsignal changesaccompanyingsensorystimulation:functional brain mappingwith magneticresonanceimaging. ProcNatlAcadSci USA 1992;89:5951–5955.15. BuxtonRB,FrankLR.A model forthecouplingbetweencerebral bloodflow and oxygen metabolismduringneural stimulation. J CerebBloodFlowMetab1997;17:64–72.16. Belliveau JW, Kennedy DN, Jr, McKinstry RC, Buchbinder BR, Weis-skoff RM, Cohen MS, Vevea JM, Brady TJ, Rosen BR. Functionalmappingof thehuman visual cortex by magnetic resonanceimaging.Science1991;254:716–719.17. Frahm J, Merboldt KD, Hanicke W. Functional MRI of human brainactivation at high spatial resolution. Magn Reson Med 1993;29:139–144.18. BockC,SchmitzB,KerskensCM,Gyngell ML,HossmannK-A,Hoehn-BerlageM.Functional MRI of somatosensoryactivationinrat:effectof hypercapnic upregulation on perfusion- and BOLD-imaging. MagnResonMed1998;39:457–461.19. van Bruggen N, Busch E, Palmer JT, Williams S-P, Beaulieu C, deCrespigny A. High resolution functional MRI mapping of the ratcerebral cortex.J CerebBloodFlowMetab1997;17:S441.20. KerskensCM, Hoehn-BerlageM, Schmitz B, Busch E, Bock C, GyngellML, Hossmann K-A. Ultrafast perfusion-weighted MRI of functionalbrain activation in rats duringforepaw stimulation: comparison with T2-weightedMRI.NMRBiomed1996;9:20–23.21. KimSG.Quantificationofrelativebloodflow changebyflow-sensitivealternating inversion recovery (FAIR) technique: application to func-tional mapping.MagnResonMed1995;34:293–301.22. Busch E, Hoehn-Berlage M, Eis M, Gyngell ML, Hossmann K-A.Simultaneous recordingof EEG, DC potential and diffusion-weightedNMRimagingduringpotassiuminducedcortical spreadingdepressioninrats.NMRBiomed1995;8:59–64.23. Frahm J, Haase A, Matthaei D. Rapid NMR imaging of dynamicprocesses usingtheFLASH technique. Magn Reson Med 1986;3:321–327.24. Gyngell ML,SchmitzB,Hoehn-BerlageM,HossmannK-A.Variationof functional MRI signal in response to frequency of somatosensorystimulation in alpha-chloralose anesthetized rats. Magn Reson Med1996;36:13–15.25. PaxinosG.Theratnervoussystem.SanDiego:AcademicPress;1995.26. McPherson RW,Zeger S,Traystman RJ.Relationship of somatosensoryevoked potentials and cerebral oxygen consumption during hypoxichypoxiaindogs.Stroke1986;17:30–36.27. YamadaS,BrauerF,KnierimD,DayesL,DjernasM,IaconoR,MorgeseV,HaywardW.Cansomatosensoryevokedpotential monitoringpredictenergy dynamicsduringcontrolled hypotension?Neurol Res1992;14:325–329.28. Hyder F, Rothman DL, Mason GF, Rangarajan A, Behar KL, ShulmanRG. Oxidative glucose metabolism in rat brain duringsingle forepawstimulation: a spatially localized 1 H( 13 C) nuclear magnetic resonancestudy.J CerebBloodFlowMetab1997;17:1040–1047.29. Leniger-Follert E, Hossmann K-A. Simultaneous measurements of microflow and evoked potentialsin thesomatomotor cortex of thecatbrain during specific sensory activation. Pflugers Arch Eur J Physiol1979;380:85–89.30. Hossmann K-A, Schmitz B, Bock C, Kerskens C, Hoehn-Berlage M.Functional magnetic resonanceimagingafter cardiac arrest in rat. In:Krieglstein J, editor. Pharmacology of cerebral ischemia. MedpharmScientificPublishers,Stuttgart1996.p357–365.31. Ueki M,LinnF,HossmannK-A.Functional activationofcerebral bloodflow and metabolism before and after global ischemia of rat brain. JCerebBloodFlowMetab1988;8:486–494.32. Schmitz B, Bo¨ttiger BW, Hossmann K-A. Functional activation of cerebral bloodflowaftercardiacarrestinrat.J CerebBloodFlowMetab1997;17:1202–1209.33. Schmitz B, Bock C, Hoehn-Berlage M, Kerskens CM, Bo¨ttiger BW,Hossmann K-A. Recovery of the rodent brain after cardiac arrest: afunctional magnetic resonanceimagingstudy. Magn Reson Med 1998;39:783–788.34. Bock C, Krep H, Brinker G, Hoehn-Berlage M. Brainmapping of alpha-chloraloseanesthetized rats with T* 2 -weighted imaging: distinc-tion between the representation of the forepaw and hindpaw in thesomatosensorycortex.NMRBiomed1998;11:115–119. Evoked Potentials and T* 2 -Weighted Imaging 473
