Recovery of the rodent brain after cardiac arrest: A functional mri study

Recovery of the Rodent Brain After Cardiac Arrest: A Functional MRI Study Bernd Schmitz, Mathias Hoehn-Berlage, Christian M. Kerskens, Bernd W. Bottiger, Konstantin-Alexander Hossmann

Recovery of the cerebral cortex after 10 min of cardiac arrest was studied in rat using noninvasive MRI techniques. The apparent diffusion coefficient (ADC) of brain water was im- aged to document reversal of the metabolic impairment. Per- fusion-weighted imaging and blood oxygen level dependent (BOLD) imaging were performed to assess functional recovery. To this purpose, rats were anesthetized with a-chloralose, and somatosensory cortex was activated by electrical stimulation of the contralateral forepaw. In sham-operated controls, cortical ADC was 862 rf: 10 pm2/s, and stimulation of forepaw led to a focal increase of signal intensity in somatosensory cortex by 71 f 22% in perfusion-weighted images and by 6 f 1% in BOLD images. One hour after successful resuscitation following 10 min of cardiac arrest, ADC did not differ from control but functional activation was completely suppressed. After 3 hours of reperfusion, functional activity began to reappear but the recovery of the BOLD signal pro- gressed faster than that of the perfusion-weighted signal. The differences in the recovery of ADC, BOLD, and perfusion imaging are related to differences between metabolic and functional recovery on one hand and between blood flow and oxygen extraction on the other. The combination of these MRI methods thus provides detailed qualitative information about the progression of brain recovery after transient circulatory arrest. Key words: apparent diffusion coefficient; perfusion-weighted imaging; BOLD imaging; cardiac arrest.

INTRODUCTION

Circulatory arrest of the brain causes rapid breakdown of energy-producing metabolism, membrane ,depolariza- tion, and a shift of fluid from the extracellular into the intracellular compartment. These changes can be re- versed as long as restoration of blood flow results in the reactivation of the energy-producing metabolism (1).

Alterations of the intracellular/extracellular water homeostasis can be monitored by diffusion-weighted MRI. There is general agreement that the macroscopically observed apparent diffusion coefficient (ADC) of water is mainly a function of the extracellular fluid volume (2), although other mechanisms have also been discussed in

MRM 39783-788 (1998) From the Max-Planck-Institute for Neurological Research, Koln, Germany. Address correspondence to: Prof. Dr. K.-A. Hossmann, Max-Planck-lnsti- tute for Neurological Research, Department of Experimental Neurology, Gleueler Str. 50, 50931 Koln, Germany. Received July 25, 1997; revised September 22, 1997; accepted October 23, 1997. The study was supported by the Deutsche Forschungsgemeinschaft (SFB 194/B1). Presented in part at the 4th scientific meeting of the International Society for Magnetic Resonance in Medicine, New York, 1996.

Copyright 0 1998 by Williams & Wilkins All rights of reproduction in any form reserved.

0740-31 94/98 $3.00

the past (3). Since maintenance of water homeostasis is energy-dependent, diffusion imaging also provides indi- rect evidence of energy metabolism (4, 5). In our laboratory, this has been documented after cardiac or isolated cerebrocirculatory arrest in cats by correlating ADC images with pictorial measurements of adenosine 5’-triphosphate (6, 7). However, restoration of brain energy metabolism is not equivalent with the return of functional activity. We were, therefore, interested to know whether MR techniques are also able to monitor functional recovery of the brain after circulatory arrest.

The MR method most widely used for the evaluation of functional activity is the blood oxygenation level dependent (BOLD) imaging method, originally described by Ogawa et al. (8). This method is based on the observation that functional activity of the brain is coupled to an increase of blood flow that exceeds the oxygen demands of the activated tissue. The resulting decline of paramag- netic deoxyhemoglobin improves the homogeneity of the magnetic field and increases signal intensity of heavily T,*-weighted images. More recently, a perfusion- weighted imaging technique has been proposed that detects the regional increase of blood flow coupled to the functional metabolic activation (for reviews, see refs. 9 and 10). By using a-chloralose as the anesthetic drug, these imaging methods can also be applied to laboratory animals, as demonstrated in rats during somatosensory stimulation (11-14).

In the present investigation, we combined ADC imaging with BOLD and perfusion-weighted imaging to fol- low the regional and temporal recovery of rat brain after cardiac arrest. Our study revealed striking dissociations between the three imaging modalities and provides detailed insight into the temporal evolution of the recovery process.

MATERIALS AND METHODS

The study was approved by the ethical committee of the local authorities and was conducted according to the German legislation on animal care. The animals were allowed free access to food and water until the start of experimentation.

General Preparations

Eighteen male adult Sprague-Dawley rats (290-380 g) were used. Animals were divided into three groups to facilitate MRI at the different time points of the experimental protocol: nonischemic controls (group I), 10 min of cardiac arrest and 6 hours of reperfusion (group 21, and 10 min of cardiac arrest followed by 1 day of reperfusion (group 3) (n = 6 per group). MR data during cardiac arrest

783

784 Schmitz et al.

(n = 3) were taken from a previous study (15). The animals were anesthetized with 1.5% halothane in a 7:3 N,O/O, mixture. A femoral artery and vein were cannu- lated with polyethylene catheters (PE-50) for blood pressure monitoring (Kombidyn Monitoring Set, Braun, Melsungen, Germany), blood sampling, and drug admin- istration. A continuous saline infusion ( 2 ml/kg/h) was started to prevent dehydration and obstruction of the catheters. After tracheotomy, the animals were paralyzed with pancuronium bromide (0.2 mg/kg/h iv) and artifi- cially ventilated (Foehr Medical Instruments, Egelsbach, Germany) to keep arterial PCO, within the physiological range. Rectal temperature was kept at 37°C using a warm water jacket wrapped around the animal and connected to a feedback controlled heating system. Arterial blood pressure was recorded continuously on a polygraph and processed with a personal computer running the Dasy- Lab data acquisition software (DataLog, Monchenglad- bach, Germany). Arterial blood samples were collected and analyzed at regular intervals, using a blood gas ana- lyzer (Model 288, Ciba-Corning, Fernwald, Germany). In group 3, in which animals survived for 1 day after resuscitation, some modifications of the protocol were made to minimize animal discomfort during the recovery in- terval. For arterial access, the tail artery was used to prevent ischemic contractions of the leg, and the animals were intubated orotracheall y to allow extubation after cardiac resuscitation. One hour after return of spontaneous circulation, all wounds were infiltrated with lido- caine and closed by suture. The animals were weaned from the respirator, extubated, and placed in an 0,- enriched environment for 1 more hour before being returned to their cages. Twenty ml of Ringer’s solution were applied subcutaneously to prevent dehydration. One day later, these animals were prepared as described for groups 1 and 2.

Resuscitation Protocol

After baseline measurements of the physiological variables, a pacing catheter was advanced into the esophagus for electrical fibrillation. The halothane supply was discontinued 2 min before electrical fibrillation to minimize adverse side effects on cardiac resuscitability. Normo- thermic cardiac arrest was induced by electrical fibrillation (12 V, 50 Hz) and confirmed by the immediate drop of systemic arterial blood pressure. Then, artificial ventilation and saline infusion were stopped, and the heating system was switched off. After 10 min, cardiac arrest cardiopulmonary resuscitation was started as described previously (16). Closed chest cardiac massage was performed by applying sternal compressions at a rate of 200/min. Artificial ventilation was resumed with 100% oxygen at a frequency 50% above control, and bolus injections of epinephrine (0.02 mg/kg) and sodium bicarbonate (NaHCO,, 0.5 mEq/kg) were given. DC countershocks (5 J) were applied 2 min after beginning cardiopulmonary resuscitation and repeated after 30-60 s if necessary. Return of spontaneous circulation (ROSC) was confirmed by spontaneous cardiac action and the buildup of mean arterial blood pressure to higher than 50 mmHg. After ROSC, the heating pad was switched on

and saline infusion was resumed. Arterial blood was sampled after 5, 30, and 60 min of reperfusion and then hourly. Ventilator speed and inspired oxygen concentration were adjusted and NaHCO, was given according to the results of the blood gas analysis.

Functional Activation and NMR Measurements

Functional activation of somatosensory cortex was studied in all groups under a-chloralose anesthesia. To this purpose, halothane/N,O anesthesia (used during prepa- ratory surgery for group 1 and up to resuscitation for groups 2 and 3) was discontinued and replaced by intra- venous anesthesia with a-chloralose (60-80 mg/kg). Dur- ing a-chloralose anesthesia, the animals were ventilated with 30% oxygen and 70% nitrogen. Supplemental doses of a-chloralose (25 mg/kg) were given at 60- to 90-min intervals.

Somatosensory stimulation was performed by insert- ing needle electrodes under the skin of the forepaw and by delivering electrical pulses of 0.3 ms in duration and 0.5-1.0 mA at a frequency of 3 Hz for 1 min using an electrical stimulator (Foehr Medical Instruments, Egelsbach, Germany).

The ADC and functional response of the brain were monitored by MRI. The head of the animal was placed in a nonmagnetic headholder for accurate positioning in the magnet. MR measurements were performed at 4.7 T using a 47/30 Bruker BIOSPEC MSL-Xll system (Bruker Med- izintechnik, Karlsruhe, Germany) with actively shielded gradient coils (Bruker; maximum gradient strength, 100 mT/m; gradient rise time, <250 ps). The RF pulses were transmitted using a 12-cm-diameter Helmholtz coil. A 16- mm-diameter surface coil with inductive coupling was cen- tered over the animal’s skull for signal reception. The two coils were decoupled, the transmitter coil actively and the surface coil passively.

After positioning the animal in the magnet, multislice fast low angle shot (FLASH) pilot scans (TE = 8 ms, TR = 400 ms, FOV = 24 mm, 1 2 8 X 128 pixels) of the brain were obtained in sagittal orientation, to place the functional imaging slice coronally through the somatosensory cortex at 4.5 mm caudal to the rhinal fissure (12) for functional imaging and ADC imaging. Diffusion- weighted images were recorded with a pulsed gradient spin-echo sequence according to Stejskal and Tanner (17) (TR = 2000 ms, TE = 34 ms). The strong magnetic field gradients allowed the use of short TEs of only 34 ms, thus minimizing the T,-dependent signal loss (18). The diffusion-sensitizing gradient was always aligned parallel to the read direction of the imaging gradients, i.e., up-down within the image plane.

For quantitative determination of ADC, three diffusion-weighted images were acquired with different gradient strengths (b factor: 0 , 750, 1500 s/mmz). Correc- tions for background-noise were performed as described elsewhere (18). ADC was calculated according to the monoexponential intravoxel incoherent motion model of LeBihan et al. (19) using an INDY workstation (SGI, Mountain View, CA). Images were transferred to an Apple Macintosh (Apple, Cupertino, CA) equipped with

fMRI After Cardiac Arrest 785

the image processing software IMAGE (National Insti- tutes of Health, Bethesda, MD).

Perfusion-weighted images were recorded using a modification (13) of the arterial spin-tagging perfusion technique (20). The magnetic spins of blood flowing through the neck vessels were adiabatically inverted, followed by snapshot-FLASH imaging starting 3 s after the beginning of spin inversion (TE = 3.9 ms, TR = 7.4 ms, FOV = 40 mm, slice thickness = 2 mm, flip angle = 18). Matrix size was 128 X 64 pixels, and scan time was 3.5 s. For BOLD imaging, T,* weighting was performed with a FLASH sequence (TE = 60 ms, TR = 70 ms, FOV = 24 mm, slice thickness = 1.5 mm,a = 22.5'), a 64 X 64 matrix size and a scan time of 4.5 s (12).

Eight BOLD and eight perfusion-weighted images were recorded before (baseline) and during forepaw stimulation. Image calculations were performed with the IDL software (Research Systems Inc., Boulder, CO) running on an INDY workstation (SGI, Mountain View, CA). BOLD-activation maps were obtained from the difference between the averaged images obtained at rest and during functional stimulation. For localization of the activated regions, pixels with intensity changes of more than 1.5 SDs above noise were overlayed on the anatomical T,-weighted images. In perfusion-weighted images, activated brain regions could be detected without subse- quent image processing.

Data Analysis

Regions of interest (ROIs; 4 X 4 pixels) were placed in the activated area of the somatosensory cortex to determine signal intensity increase in the T,*- and perfusion- weighted images. These ROIs were positioned in the center of the activated region to exclude analysis arte- facts caused by changes in the apparent size of the activated area. For the ADC analysis in the coronal activation plane, additional ROIs of equal size were placed in the striatum and the thalamus.

All values are expressed as means ? SEM. The data were tested for normal distribution with the Kolmogorov- Smirnow test and compared for significant differences using analysis of variance or Kruskal-Wallis analysis of variance on ranks followed by the Student-Newman- Keuls test (STATISTICA for Windows 4.5, StatSoft Inc.,

Tulsa, OK). A probability value of P < 0.05 was consid- ered significant.

RESULTS Physiological Observations

In control animals and in experimental animals before cardiac arrest, all physiological variables were within the normal range (Table 1). Ventricular fibrillation led to complete circulatory arrest within seconds, as evidenced by the immediate decrease in arterial blood pressure. After 10 min of cardiac arrest, all animals could be re- suscitated within 4 min (mean duration of cardiopulmonary resuscitation, 165 2 11 s). No animals required more than one epinephrine bolus; the amount of sodium bicarbonate (0.65 ? 0.04 mEq/kg) and the number of DC countershocks (2.3 ? 0.25) required for successful resuscitation were similar in all groups.

After ROSC, mean arterial blood pressure rose sharply, peaking at 176 -f 5 mmHg by 52 ? 4 s of reperfusion. Acidosis could be corrected within 60 min by adminis- tration of sodium bicarbonate in combination with hy- perventilation (Table 1). Pulmonary function was little affected as evidenced by the rapid normalization of P,CO, and P,O,, This allowed reduction of the oxygen concentration in the ventilation gas to prearrest values within 30-60 min.

Table 1 Physiological Variables Before and After 10 minutes of Cardiac Arrest of Rat

MRI

Before cardiac arrest, ADC in the somatosensory cortex was 862 ? 10 pm2/s. During cardiac arrest, ADC sharply declined to 520 ? 16 pm2/s (15). At 1 hour, after 10 min of cardiac arrest, the earliest reperfusion time measured in this study, ADC had returned to control (842 2 45 FmZ/s), and it remained at this level throughout the observation time (after 3 hours, 860 ? 48 pm2/s; after 6 hours, 780 ? 81 pm2/s; and after 1 day of reperfusion, 905 2 22 pm"/s, respectively. The pictorial evaluation of ADC maps demonstrated the fast normalization of ADC in the reperfusion phase but also confirmed changing regional heterogeneities during the first hours of reperfusion, a situation already reported in earlier investiga- tions in cat brain after cardiac arrest (6, 7, 21) (Fig. 1, Table 2).

Temperature ("C)

MABP PO, (torr) pC0, (torr) pH (units) Hct (percent) (mmHg)

Prearrest 106 t 2 124 t 5 41.4 t 0.6 7.41 i 0.01 41 2 37.0 t 0.3 37.0 i 0.1 Reperfusion

5 minutes 84 t 9" 235 t 14" 48.8 t 3.3" 7.18 i 0.03" n.d. 34.7 t 0.1" 30 minutes 118? !ia 225 Z 20" 48.4 t 2.3" 7.28 i 0.02" 50 t 0.5" 35.9 t 0.31" 1 hour 91 2 3 250 t 26" 43.9 i 2.9 7.40 t 0.02 47 i 0.8" 37.0 t 0.2 3 hours 98 i 5 168 t 9" 34.1 t 1.4a 7.44 + 0.02 40 t 0.8 36.8 i 0.2 6 hours 104 t 3 131 5 9 39.8 f 1.5 7.37 2 0.02 41 i 0.9 37.0 t 0.1 24 hours 96 t 1" 121 + 4 39.9 t 1.8 7.39 i 0.01 39 f 0.5 37.0 t 0.1

All values are means ? SEM. Prearrest values are control recordings from all three groups (n = 18); 5-minute to 1-hour reperfusion values are from groups 2 and 3 (n = 12); 3- and 6-hour reperfusion values are from group 2 (n = 6); and I-day reperfusion values are from group 3 (n = 6). a Significantly different from control (f < 0.05). MABP = mean arterial blood pressure; n.d. = not determined.

786 Schmitz et al.

FIG. 1. Representative maps of the ADC of brain water before, during, and at various times after successful resuscitation from 10 min of cardiac arrest in rat. Measurements were performed in different animals (see Methods section). Note rapid recovery after restoration of spontaneous circulation.

Under control conditions, electrical stimulation of the forepaw led to a circumscribed increase in the signal intensity of the opposite somatosensory cortex by 71 t 22% in perfusion-weighted images and by 6 t 1% in BOLD images (Figs. 2 and 3). These changes were unre- lated to alterations of arterial blood pressure, which remained constant during somatosensory stimulation. Side-specificity of the response was tested by alternating the stimulation between both forepaws. The specificity of the activation was also confirmed by subtraction of images recorded without stimulation, which resulted in cancellation of all extracranial and intracranial signals. One hour after successful resuscitation, functional activation was completely suppressed in all animals. After 3 hours of reperfusion, the BOLD and perfusion-weighted signal intensity changes were reduced to approximately 10% of the control response. Between 6 hours and 1 day of reperfusion, a dissociation was observed between the BOLD signal, which recovered from 30 to 92% of the control response, and the signal in the perfusion- weighted images, which improved only from 8 to 49% of the control response (Figs. 2 and 3).

DISCUSSION

To the best of our knowledge, this is the first investigation in which ADC and fMRI techniques were used to

Table 2 ADC in Different Brain Regions Before and After 10 Minutes of Cardiac Arrest in Rat

Cerebral cortex Caudate putamen Thalamus (w-n’/s) (w-n2/s) (m2/s)

Control 862 i 10 827 + 29 788 2 26 Reperfusion

1 hour 842 + 45 766 2 23 776 2 27 3 hours 860 -f 48 760 t 22 754 -c 19 6 hours 780 i 81 726 t 47 702 2 37 1 day 905 5 22 828 i 24 776 t- 33

All values are means t SEM. Control recordings were carried out in group 1 (n = 6), 1- to 6-hour reperfusion was performed in group 2 (n = 6), and 1 -day reperfusion was performed in group 3 (n = 6).

monitor the recovery of the cerebral cortex after cardiac arrest. Our findings demonstrate that functional activation, as re- flected by the stimulus-evoked changes of local blood flow and blood oxygenation, is completely suppressed during the early reperfusion period, de- spite full recovery of the ADC of brain water. They also show that even after 24 hours of recirculation, this response does not recover to normal. To appreci- ate the pathophysiological sig- nificance of these results, the biophysical basis of ADC and fMRI must be discussed briefly.

BOLD MRI detects deoxyhemoglobin-induced susceptibil-

ity changes of blood magnetization and is essentially a measure of venous oxygen content. The spin-tagging technique used for perfusion-weighted imaging takes ad- vantage of differences in the mean transit time of labeled blood water for imaging differences in cerebral blood flow. Both methods can be used to localize functional activation of cerebral cortex because the transient increase of functional activity leads to an increase in the metabolic rate of glucose and, to a lesser degree, of oxygen, which is coupled to a parallel increase of blood flow. Since the setpoint of the flow coupling is at a higher level than the increase in oxygen utilization, oxygen availabil- ity is in excess of the oxygen consumption, leading to a decrease in deoxyhemoglobin and the observed change in magnetic susceptibility. In a previous investigation of the same model of cardiac arrest (22), we demonstrated that the recovery of electrophysiological activation, as assessed by recording of primary evoked somatosensory cortical potentials, precisely matches the observed flow increase as measured with laser Doppler flowmetry. This study also revealed that in the same model of 10 min of cardiac arrest, both the amplitude of evoked potentials and the associated flow response recovered to only 50% within the first week of recirculation. The present find- ing of partial recovery of signal intensity changes in perfusion-weighted image is, therefore, fully consistent with these data. The normalization of the BOLD signal within 24 hours is, therefore, not evidence for full functional recovery at this time point. However, it demonstrates that the flow overshoot that is characteristic for functional activation has reappeared and that the setpoint of this response has returned to the preischemic level. This is consistent with the previous observation that the CO, responsiveness of the cerebral circulation, which is part of the multifactorial coupling mechanism, returns to normal long before the electrophysiological and hemodynamic re- sponses have recovered (22). The reduced responsiveness of perfusion-weighted imaging to functional stimulation is, therefore, reliable evidence for the persisting functional impairment.

This disturbance is clearly at variance with the rapid normalization of the ADC of brain water. There is com-

fMRI After Cardiac Arrest 787

Our data imply the preserva- tion of coupling, in contrast to earlier results after longer peri- ods of cerebral ischemia, in which the coupling was found to be disturbed (24). These di- verging results raise the ques- tion of the physiological importance of the functional coupling process. It is widely held that coupling serves to fuel the increased energy demands of the neurons and to remove metabolic by-products from the activated tissue (for a recent re- view, see Villringer and Dirnagl (25)). However, this relation- ship may not be a direct one. It has been proposed that glucose

FIG. 2. Functional activation of the right somatosensory cortex by electrical stimulation of the left is metabolized mainly anaerobi- forepaw before and at various times after successful resuscitation from 10 min of cardiac arrest cally in astrocytes to provide in rat. Functional MRI was performed using BOLD and perfusion-weighted imaging. Note post- the energy required for the ischemic suppression of the MR response followed by slowly progressing recovery. transport of glutamate across

pelling evidence that under conditions of acute ischemia, alterations of ADC closely reflect parallel changes in energy metabolism (6, 7). This is explained by the energy dependence of ion exchange pumps, which control both ion and water homeostasis between intracellular and extracellular fluid compartments. The normalization of the ADC at the earliest reperfusion time of 1 hour is, therefore, evidence that, at this time, energy metabolism has already fully recovered.

In view of these considerations, the results of the present investigation are interpreted as follows. At 1 hour after 10 min of cardiac arrest, energy metabolism and ion homeostasis have recovered but cortical functional activity is severely suppressed. Within 24 hours after cardiac arrest, the coupling mechanism of functional activation to increase blood flow also returns to normal, but functional activation itself is only partially restored. This persisting disturbance of functional performance of cerebral cortex confirms earlier observations that the restoration of integrative neuronal function is a process that recovers much slower than that of the electrophysiological or metabolic activity of the brain (23).

Cortical ADC 1000, HFq l*fiflfi 600 500

C CA I h 3h 6h Reoelfusion

1 BOLD-imaging

the plasma membrane (26). The produced lactate, in turn, serves the neurons as substrate for aerobic metabolism. Obviously, the main energy ex- penditure during functional activation is not related to the generation of action potentials, which require only 0.3-370 of cortical energy consumption (27). This is far below the reported increase in glucose consumption during somatosensory stimulation, which under chloralose anesthesia, amounts to approximately 50% (24). There- fore, other biochemical activities, including glutamate transport, must be responsible for the increased metabolic demands during brain activation. If these activities would impose a substantial workload on the brain, the suppression of the flow coupling after ischemia could, in fact, reflect disturbances of the functional performance. This interpretation is underlined by the fact that the recovery of the flow response correlates not only with the amplitude of evoked potentials but also with neurological outcome (22).

In conclusion, our study demonstrates the usefulness of fMRI studies for the analysis of brain resuscitation after cardiac arrest in experimental animals. The present study documents the much slower recovery of the func-

Perfusion-weighted imaging

P 20

C I h 3h 6h Id Reperfusion Reperfusion

FIG. 3. Measurements of the ADC and of the signal intensity changes of fMRl before and at various times after successful resuscitation from 10 min of cardiac arrest in rat. Recordings were taken from somatosensoy cortex; functional MRI (BOLD and perfusion-weighted imaging) was performed during electrical stimulation of contralateral forepaw imaging. Values are means f SEM, n = 3 for cardiac arrest and n = 6 for all other times (for group assignments see Methods section). Note rapid normalization of ADC as compared to the much slower recovey of functional activation. *, Significantly different from control; +, significantly different from ADC recovey. P < 0.05.

tional performance of the brain than that of the ADC, which, in turn, depends on the recovery of cerebral energy metabolism. The dissociation between BOLD and perfiision-weighted imaging is explained by the different contrast mechanisms leading to disproportionately more rapid recovery of the BOLD effect. Therefore, BOLD imaging alone cannot be used to grade the severity of the ischemic impact or to monitor recovery from cerebral ischemia, but it must be combined

788 Schmitz et al.

with perfusion-weighted imaging to avoid misinterpreta- tions of the quality of functional recovery.

Beyond the present investigation on brain recovery from cardiac arrest, our results demonstrate an aspect of more general interest: The complementary information of blood oxygenation and perfusion, provided by the T2*- and perfusion-weighted imaging modalities, allows unequivocal conclusions about the degree of activation, which either of the imaging modalities alone could not provide. This may be of particular importance for the evaluation of functional deficits, e.g., after stroke. There- fore, the combination of both T,*- and perfusion- weighted imaging is strongly recommended for the ap- plication of fMRI to pathological situations.

ACKNOWLEDGMENTS

The authors thank Dr. Ferdi van Dorsten for help with the ADC analysis and Mr. B. Radermacher for excellent technical assistance. We also gratefully acknowledge the artwork of Mrs. I. Miihlhofer and Mr. B. Huth and the secretarial assistance of Mrs. D. Schewetzky and Mrs. M. Hahmann.

REFERENCES 1. K. -A. Hossmann, Cortical steady potential, impedance and excitabil-

ity changes during and after total ischemia of cat brain. Exp. Neurol. 32, 163--175 (1971). ,

2. H. Bemeniste, L. W. Hedlund, G. A. Johnson, Mechanism of d e t e c tion of acute cerebral ischemia in rats by diffusion-weighted magnetic resonance microscopy. Stroke 23, 746-754 (1992).

3. J. A. Helpern, R. J. Ordidge, R. A. Knight, The effect of cell membrane water permeability on the diffusion coefficient of water, in “Proc., SMRM, 12th Annual Meeting, Proc., SMRM, 11th Annual Meeting, Berlin, Germany, 1992,” p. 1201.

4. A. L. Busza, K. L. Allen, M. D. King, N. v. Bruggen, S. R. Williams, D. G. Gadian, Diffusion-weighted imaging studies of cerebral ischemia in gerbils -potential relevance to energy failure. Stroke 23, 1602-1612 (1992).

5 . C. Decanniere, S. Eleff, D. Davis, P. C. M. v. Zijl, Correlation of rapid changes in the average water diffusion constant and the concentra- tions of lactate and ATP breakdown products during global ischemia in cat brain. Magn. Reson. Med. 34, 343-352 (1995).

6. K. -A. Hossmann, M. Fischer, K. Bockhorst, M. Hoehn-Berlage, NMR imaging of the apparent diffusion coefficient (ADC) for the evaluation of metabolic suppression and recovery after prolonged cerebral ischemia. J. Cereb. Blood Flow Metab. 14, 723-731 (1994).

7. M. Fischer. K. Bockborst, M. Hoebn-Berlage, B. Schmitz, K. -A. Hoss- mann, Imaging of the apparent diffusion coefficient for the evaluation of cerebral metabolic recovcry after cardiac arrest. Magn. Reson. Imaging 13, 781-790 (19951.

8. S. Ogana, T. M. Lee, A. S. Nayak, P. Glynn, Oxygenation-sensitive contrast in magnetic resonance image of rodent brain at high magnetic fields. Magn. Reson. Med. 14, 68-78 (1990).

9. R. Turner, Functional mapping of the human brain with magnetic resonanize imaging. Semin. Neurosci. 7, 179-194 (1995).

10. A. G. Sorensen, S. H. Wray, R. M. Weisskoff, J, L. Boxerman, T. L. Davis, F. Caramia, K. K. Kwong, C. E. Stern, J. R. Baker, H. Breiter, I. E. Gazit, J. W. Belliveau, T. J. Brady, B. R. Rosen, Functional MR of brain activity and perfusion in patients with chronic cortical stroke. A m . J. Neuroradiol. 16, 1753-1762 (1995).

11. F. Hyder, K. L. Behar, M. A. Martin, A. M. Blamire, R. G. Shulman, Dynamic magnetic resonance imaging of the rat brain during forepaw stimulation. J. Cereb. Blood Flow Metab. 14, 649-655 (1994).

12. M. L. Gyngell, C. Bock, B. Scbmitz, M. Hoehn-Berlage, K. -A. Hoss- mann, Variation of functional MRI signal in response to frequency of somatosensory stimulation in alpha-chloralose anesthetized rats. M a p . Reson. Med. 36, 13-15 (1996).

13. C. M. Kerskens, M. Hoehn-Berlage, B. Schmitz, E. Busch, C. Bock, M. L. Gyngell, K. -A. Hossmann, Ultrafast perfusion-weighted MRI of functional brain activation in rats during forepaw stimulation: com- parison with T2*-weighted MRI. NMR Biomed. 8, 20-23 (1996).

14. C. Bock, B. Schmitz, C. M. Kerskens, M. L. Gyngell, K. -A. Hossmann, M. Hoehn-Berlage, Functional MRI of somatosensory activation in rat: effect of hypercapnic upregulation on perfusion- and BOLD- imaging. Magn. Reson. Med. 39, 457-461 (1998).

15. M. Hoehn-Berlage, T. Back, D. G. Norris, Early changes in apparent diffusion coefficient of rat brain following total circulatory arrest.

16. B. Bottiger, J. Krumnikl, P. Gass, B. Schmitz, J. Motsch, E. Martin, The cerebral “no-reflow” phenomenon after cardiac arrest in rats: influence of low-flow states before stabilisation. Resuscitation 34, 79-87 (1997).

17. E. Stejskal, J. Tanner, Spin diffusion measurements: spin echoes in the presence of a time-dependent field gradient. J. Cbem. Phys. 42, 229-292 (1965).

18. M. Eis, M. Hoehn-Berlege, Correction of gradient cross-talk and op- timization of measurement parameters in diffusion MR imaging. Magn. Reson. Med. 107, 222-234 (1995).

19. D. LeBihan, E. Breton, D. Lallemand, M. -L. Aubin, J. Vignaud, M. Laval-Jeantet, Separation of diffusion and perfusion in intravoxel incoherent motion MR imaging. Radiology 168, 497-505 (1988).

20. J. A. Detre, J. S. Leigh, D. S. Williams, A. P. Koretsky, Perfusion imaging. Magn. Reson. Med. 23, 37-45 (1992).

21. B. Schmitz, M. Fischer, K. Bockhorst, M. Hoehn-Berlage, K. -A. Hoss- mann, Resuscitation from cardiac arrest in cats: influence of epinephrine dosage on brain recovery. Resuscitation 30, 251-262 (1995).

22. B. Schmitz, B. W. Bottiger, K. -A. Hossmann, Functional activation of cerebral blood flow after cardiac arrest in rat. J. Cereb. Blood Flow Metab. 17, 1202-1209 (1997).

23. K. -A. Hossmann, R. Schmidt-Kastner. B. Grosse Ophoff, Recovery of integrative central nervous function after one hour global cerehro- circulatory arrest in normothermic cat. J. Neurol. Sci. 77, 305-320

24. M. Ueki, F. Linn, K. -A. Hossmann, Functional activation of cerebral blood flow and metabolism before and after global ischemia of rat brain. J. Cereb. Blood Flow Metab. 8 , 486-494 (1988).

25. A. Villringer, U. Dirnagl, Coupling of brain activity and cerebral blood flow: basis of functional ueuroirnaging. Cereb. Brain Metab. Rev. 7, 240-276 (1995).

26. L. Pellerin, P. Magistretti, Glutamate uptake into astrocytes stimu- lates aerobic glycolysis: a mechanism coupling neuronal activity to glucose utilization. Proc. Natl. Acad. Sci. 91, 10625-10629 (1994).

27. 0. D. Creutzfeld, Neurophysiological correlates of different functional states of the brain, in “Brain Work” (D. H. Ingvar, N. A. Lassen, Eds.), p. 21-47, Munksgaard, Copenhagen, 1975.

MAGMA 2, 39-42 (1994).

(1987).

Documents

Recovery of the rodent brain after cardiac arrest: A functional mri study