Journal of Neuroscience Methods 131 (2003) 75–81

Continuous measurement of intracranial pressure in awake rats afterexperimental closed head injury

Servan Rookera,∗, Philippe G. Jorensb, Jos Van Reemptsc, Marcel Borgersc, Jan Verlooya

a Department of Neurosurgery, UZA, University Hospital of Antwerp, Wilrijkstraat 10, B-2650 Edegem, Belgiumb Department of Intensive Care Medicine, University Hospital of Antwerp, Edegem, Belgium

c Department of Life Sciences, Janssen Research Foundation, Beerse, Belgium

Received 27 April 2003; received in revised form 12 July 2003; accepted 21 July 2003

Abstract

The present study validates a method for continuous measurement of intracranial pressure (ICP) in freely moving rats after experimentalinduction of impact–acceleration injury. Rats subjected to either mild or moderate trauma were individually placed in a Bas-Ratturn® system,equipped with a sensor that synchronously turns the cage in response to the locomotor activity of the animal. In this way correct probepositioning is permanently assured and damage due to coiling is avoided. The evolution of ICP and mean arterial blood pressure (MABP) ininjured rats was compared with that of a non-traumatized sham group. Since the animals regained consciousness after surgery, interferenceof anaesthesia on these sensitive parameters should be minimised.

The results showed that immediately after induction of neurotrauma, ICP was significantly higher in traumatized rats (sham: 7.7±0.5 mmHg;mild trauma: 10.4±0.7 mmHg; moderate trauma: 14.9±2.4 mmHg;P < 0.05). Regression analysis showed a stable ICP up to 3 h post-insultfor all three conditions. From 4 h onwards till the end of the experiment at 10 h post-insult, a significant increase in ICP was seen forsham-operated and mildly traumatized rats (16.1±3.4 and 30.5±6.9 mmHg, respectively;P < 0.05), but not for moderately traumatized rats(47.3 ± 11.9 mmHg). The method allows observation of ICP for a critical period up to 3 h. As such the method can be regarded as clinicallyrelevant to study early pathological aspects of intracranial hypertension and to define a therapeutic window for pharmacological interventionafter traumatic brain injury (TBI).© 2003 Elsevier B.V. All rights reserved.

Keywords: Traumatic brain injury; Closed head injury; Rats; Long-term measurement; Intracranial pressure

1. Introduction

Traumatic brain injury (TBI) remains a major publichealth problem (Bullock et al., 1996, 2000; Goldstein,1990; Narayan and Michel, 2002). Several pharmacologicaltrials have been performed in recent years but have failedto demonstrate significant improvement in outcome, despitepromising preclinical data. The pathophysiology of TBI ishardly understood, and drug studies should target a knownmechanism occurring after TBI. Therefore experimentalmodels should mimic the situation in clinical practice.

One of the main shortcomings in recent preclinical exper-iments is the lack of adequate therapeutic window studies inanimal models. Even if the animal model mimics the clinical

∗ Corresponding author. Tel.:+32-3-821-3336; fax:+32-3-825-2428.E-mail address: servanrooker@belgacom.net (S. Rooker).

situation, it remains important to be able to translate tempo-ral changes occurring in the laboratory animals to humans. Ithas been questioned whether the assumption of the 2 h ther-apeutic window in the rat, for instance, translates to an 8 htherapeutic window in humans (Narayan and Michel, 2002).

After experimental closed head injury (CHI) the only re-liable, predictable and reproducible parameter is intracranialpressure (ICP). Since the introduction of ICP measurement(Guillaume and Janny, 1951), ICP has become an essentialparameter in the assessment of severe head injured patients.The therapeutic regimes are often instituted on the basis ofICP values. Contra-indications for long-term measurementare risks inherent to the employed method such as infection,bleeding, damage to brain tissue and technical problems withthe used probe.

Several groups of investigators have developed animalmodels of TBI in an attempt to reproduce various aspects of

0165-0270/$ – see front matter © 2003 Elsevier B.V. All rights reserved.doi:10.1016/S0165-0270(03)00233-4

76 S. Rooker et al. / Journal of Neuroscience Methods 131 (2003) 75–81

the pathophysiological responses, neurological syndromes,and histopathological findings observed in human head in-jury (Foda and Marmarou, 1994; Marmarou et al., 1994;McIntosh et al., 1987, 1989). As reported earlier, we de-veloped an adequate and reliable model that allows mea-surement of cerebral ICP after graded impact–accelerationclosed head injury. These measurements were done in theacute phase as well as in the late phase (De Mulder et al.,1999; Engelborghs et al., 1998, 2000; Rooker et al., 2002).Since evolution of ICP after CHI can only be reliably inter-preted when the ICP is recorded in one animal at all timepoints, the present study aimed at developing a technique tostudy ICP and mean arterial blood pressure (MABP) changesover a prolonged time period in the awake rat. For this pur-pose we used the Bas-Ratturn® which allowed secure posi-tioning and continuous use of an ICP probe. In the presentstudy ICP and MABP were followed for 10 h in rats thatwere allowed to survive after a mild and moderate traumaticimpact in the absence of skull fractures, and compared tonon-traumatized sham rats.

2. Materials and methods

2.1. Animal preparation and trauma induction

Animal housing and treatment conditions complied withEU directive for Animal Welfare #86/609. Trauma induc-tion was performed as described earlier (De Mulder et al.,1999; Engelborghs et al., 1998, 2000). In short, non-fastedmale Sprague–Dawley rats, weighing 390–430 g, underwentstandard anaesthesia induction over 4 min with 4% isoflu-rane in a mixture of 30% oxygen and 70% nitrous oxide.Subsequently, rats were rapidly intubated and anaesthesiawas maintained at 2% isoflurane. The head of the rat wasfixed into a stereotaxic apparatus (Kopf Intsruments, Düs-seldorf, Germany) and a 2 cm midline incision of the scalpwas made. After removal of the periost the impact place wasdetermined on the exposed bregma, and the rat was trans-ferred to the trauma device. In the group subjected to mildCHI the weight was dropped from a height of 30 cm. To in-duce moderate CHI the weight was dropped from a heightof 50 cm. After the impact the rats were positioned in thestereotaxic apparatus and the trachea tube connected to thegas circuit with 2% isoflurane. When skull fractures or ap-nea was present, the animals were excluded from the ex-periment. Otherwise the animals were prepared for ICP andblood pressure measurement.

A total of 30 animals were used for the experiment: shamanimals (n = 10), mild CHI (n = 10) and moderate CHI(n = 10).

2.2. Preparation for the Bas-Ratturn®

The system consists of a swivel that activates a bowl whichturns in the opposite direction of the movement from the

swivel. The swivel is fixed to the head, thus reflecting thebody movements of the rat. With this principle, the cathetersare kept in place and coiling up is avoided.

The femoral artery was exposed and catheterised with aPE-50 catheter; the catheter was tunnelled subcutaneouslytowards the skull. To avoid displacement and consequentbleeding, the catheter was fixed on the muscles of the legwith Vicryl 3.0 (Ethicon, USA). The femoral vein was pre-pared in the same manner to allow the administration of painmedication or other compounds.

For implantation of the ICP probe (Codman & Shurtl-eff Inc., Randolph, MA) the rat was fixed in the stereotaxicapparatus and a hole was burred in the right parietal bone,at a point 4 mm from the midline and 2 mm caudal to thebregma suture. After puncturing the dura, the ICP probe wasinserted in the right parietal cortex using the stereotaxic ap-paratus. A screw was placed 1 mm dorsal of the burr hole toallow fixation of the ICP probe with a ligature. The swivelwas fixed to the ICP probe and catheter with tape. Subse-quently, the rat was disconnected from the gas circuit andplaced in the bowl. The trachea tube was removed when therat was breathing spontaneously. The arterial catheter wasconnected to a Y-shaped connector, to allow both infusionof normal saline 0.3 ml/h with heparine (1 ml 5000 IU solu-tion in 100 ml normal saline) as well as blood pressure mea-surement. MABP and ICP were recorded with a MacLab®/8(MK3 Version 3.5, ADInstruments, Australia). Mean valuesof MABP and ICP were calculated every 15 min and repre-sent one measurement point in the figures. During the entirerecording period the rats had free access to food and waterin the bowl.

2.3. ICP probe stability

To verify possible drift of the ICP probe, the probe waspositioned in a water column at room temperature to obtaina start pressure between 15 and 16.5 mmHg. Pressure wascontinuously measured with the same probe and MacLab®

recorder for 10 h. This was done three times during thecourse of the experiment. The drift was expressed as relativechange compared to the mean values of the first hour.

2.4. Neuropathology

Three rats, out of the mild trauma group, were used for his-tological observation of the area surrounding the ICP probe.After the experiment, brains were immersion-fixed in 4%formalin solution for 24 h. Coronal vibratome sections of50�m were prepared through the probe traject, and stainedwith azure-eosin.

2.5. Statistical analysis

Statistical computations were performed using a commer-cially available software package for exact statistical infer-ence (StatXact 4.0.1 for Windows).

S. Rooker et al. / Journal of Neuroscience Methods 131 (2003) 75–81 77

First, ICP measurements were compared between thethree groups at 0, 1, 2 and 3 h after onset of the recording.When significant difference was found (P < 0.05) with aKruskal–Wallis test, a two-sided Wilcoxon–Mann–Whitneyrank-sum test was used for analysis between pairs of groupsseparately. Two-sided probability values of less than 0.05were regarded as statistically significant.

Secondly, the continuous course of the recorded datawas analysed. Calculations were performed for ICP, MABPand cerebral perfusion pressure (CPP), whereby CPP=MABP–ICP. The values on different time points within onechallenge-group were compared with a two-sided Wilcoxonsigned-rank test. Two-sided probability values of less than0.05 were regarded as statistically significant.

Finally, stability of the probe was evaluated over a timeperiod of 10 h. Baseline ICP was calculated over the firsthour. The percentile change in relation to this baseline valuewas compared with the recordings after 2, 3 and 10 h, us-ing a two-sided Wilcoxon–Mann–Whitney rank-sum test.Two-sided probability values of less than 0.05 were regardedas statistically significant.

3. Results

All animals included in the experiment, survived a 10 hrecording period. They looked healthy as could be derivedfrom adequate food intake and normal locomotor activity.At no time their activity was complicated or hindered by theRatturn system. The implanted probes remained correctly inplace and no mechanical damage could be detected. In themoderate trauma group 2 animals had to be excluded from

Time (hours)0 1 2 3

ICP

(m

mH

g)

0

5

10

15

20

25

Sham Mild CHIModerate CHI

N.S.

Fig. 1. Comparison of ICP at 0, 1, 2 and 3 h after challenge. Values are expressed as mmHg with standard error of the mean. Significant differences arefound at all time points between sham and trauma groups (P < 0.05). The ICP values were also significantly different between the two trauma groups(P < 0.05), except at the onset (N.S.).

the experiment since MABP data were not complete due toclots in the catheter.

3.1. Comparison of intracranial pressure measurementsbetween the groups

Absolute ICP values, up to 3 h, for each group are shownin Fig. 1. At the onset of the experiment the ICP values were7.7 ± 0.5, 10.4 ± 0.7 and 14.9 ± 2.4 mmHg for the sham,mild and moderate CHI group, respectively (mean±S.E.M.).These values remained rather constant up to 3 h: 6.0 ± 0.7,9.9 ± 1.4 and 16.9 ± 2.5 mmHg for the sham, mild andmoderate CHI group, respectively.

ICP values were significantly higher at all time points inthe trauma groups as compared with the sham-operated rats.With exception at onset of the experiment, moderate traumaresulted in significantly higher ICP than mild trauma.

3.2. Time-related evolution of ICP, MABP and CPPwithin each experimental group

3.2.1. ICP measurementsSham-operated as well as traumatized animals showed a

sharp increase in the ICP at 4 h post-insult as can be seenin Fig. 2A. At 10 h ICP amounted to 16.1± 3.4, 30.5± 6.9and 47.3±11.9 mmHg in the sham, mild and moderate CHIgroup, respectively. Analysis of trendlines for two distin-guishable periods (0–3 and 4–10 h) were calculated usinglinear regression analysis. The trend lines revealed almosthorizontal lines up to 3 h after onset of the measurement(sham:y = −0.10x + 6.88; R2 = 0.57; mild CHI: y =0.02x + 10.10; R2 = 0.02; moderate CHI:y = 0.24x+

78S.

Rooker

etal./JournalofN

euroscienceM

ethods131

(2003)75–81

0 2 4 6 8 100

10

20

30

40

50

60

70ICP

(A)

mm

Hg

0 2 4 6 8 1080

90

100

110

120

130

140

150MABP

(B)

Time (hours)

0 2 4 6 8 1080

90

100

110

120

130

140

150

CPP

(C)

Sham Mild CHIModerate CHI

Fig. 2. Continuous registration of intracranial pressure (ICP) (A); mean arterial blood pressure (MABP) (B); and cerebral perfusion pressure (CPP)(C). Four hours after the start of the measurementthere is a sharp increase of ICP in all groups, including the sham group. This increase resulted in extremely high ICP values after 10 h. The MABP also increases, in the trauma groups, but to a lesserextent. CPP decreases during the late phase of the experiment due to the increase in ICP.

S. Rooker et al. / Journal of Neuroscience Methods 131 (2003) 75–81 79

14.26;R2 = 0.70). A significant difference was present be-tween the early phase of the sham and moderate traumagroup (P < 0.05).

On the other hand, none of the trendlines for the latephase, between 4 and 10 h, were horizontal (sham:y =0.35x + 8.78; R2 = 0.89, mild CHI: y = 0.85x + 9.44;R2 = 1.00, moderate CHI:y = 1.24x + 17.71; R2 = 1.00).There were no significant differences in trendlines betweenthe three groups for the late phase. The trendline of thelate phase differed significantly from the early phase in thesham-operated and the mild trauma group (P < 0.05).

3.3. MABP measurements

Shortly after onset of the experiment, there was a tempo-rary decrease in blood pressure in all groups. An increaseof MABP was seen in the trauma groups from 3 to 4 h on-wards (Fig. 2B). MABP values started at 132± 3, 124± 5and 124± 4 mmHg and changed to 123± 6, 132± 10and 139± 4 mmHg in the sham, mild and moderate CHIgroup, respectively. The trend lines in the three experimen-tal groups up to 3 h after onset of the measurement werey =−1.05x + 127.28; R2 = 0.70 in the sham group and in themild CHI group:y = −0.53x + 124.09; R2 = 0.58 and formoderate CHI:y = 0.00x+121.84;R2 = 0.00. Trend linesfor the late phase were: sham:y = 0.25x + 117.44; R2 =0.75, mild CHI: y = 0.69x + 118.95; R2 = 0.84, moderateCHI: y = 0.54x + 124.63; R2 = 0.93. For all experimen-

Fig. 3. Microscopic appearance of the brain tissue in the vicinity of the probe traject (arrow) 10 h after induction of mild trauma (A). The box in (A)indicates the detailed image as shown in (B). Details of the ipsilateral (B) and contralateral hemisphere (C) show a comparable staining aspect. Withexception of glial proliferation and leucocyte invasion at the edge of the probe traject (asterisk), there is no obvious difference in numerical density ofneuronal cells (50�m vibratome section; azure-eosin staining).

tal groups the increase of MABP per hour was significantlydifferent in the late phase compared to the early phase (P <

0.05). In addition, significant difference was found in thelate phase between sham and mild groups and in the earlyphase between mild and moderate groups (P < 0.05).

3.4. CPP measurements

The cerebral perfusion pressure was calculated by sub-tracting the ICP from the MABP. Time-related changes areshown in Fig. 2C. The trend lines for the first time pe-riod were: sham:y = −0.95x + 120.39; R2 = 0.71, mildCHI: y = −0.54x + 114; R2 = 0.50, moderate CHI:y =−0.24x + 107.59; R2 = 0.41. Trend lines for the late phasewere: sham:y = −0.10x + 108.65; R2 = 0.22, mild CHI:y = −0.16x + 109.52; R2 = 0.22, moderate CHI:y =−0.71x + 106.92; R2 = 0.93. For all experimental groupsthe changes of CPP per hour were significantly different inthe late phase compared to the early phase (P < 0.05). Inthe sham and mild trauma groups there was an increase inslope of CPP, but in the moderate trauma group a significantdecrease. In addition, significant difference was found in theearly phase between sham and moderate groups (P < 0.05).

3.5. ICP probe stability

Baselines ICP for the three performed measurements were15.4, 16.8 and 15.9 mmHg, respectively. The continuous data

80 S. Rooker et al. / Journal of Neuroscience Methods 131 (2003) 75–81

of the three individual measurements were stable and thepercentile change after 10 h is 5.5% (range 3.1–8.3%), whichcorresponds to a rise of less than 1 mmHg.

3.6. Neuropathology

Evaluation of gross morphologic changes on 50�m vi-bratome sections (Fig. 3) showed relatively few bleedingsin the close vicinity of the probe. There were no signs ofedema in the tissue surrounding the probe, as could be de-rived from the absence of prominent eosinophilic staining,the preservation of a normal numerical density of the nu-clear profiles and the uniform section thickness. Residualblood in microvessels was identified as poor intracerebralfixation rather than thrombosis, since it was present also inthe contralateral hemisphere.

4. Discussion

In an attempt to develop a method for long-term ICP mea-surement in awake animals, temporal changes were evalu-ated in the rat after mild and moderate closed head injury.Immediately after induction of trauma the ICP was signifi-cantly elevated as well in the mild as in the moderate traumagroup. Up to 4 h after implantation of the probe, ICP re-mained stable. Thereafter in most animals a sudden increasewas observed, including in sham-operated animals. ICP val-ues amounted up to 47 mmHg after moderate CHI. Driftof the probe could be an explanation for this rise as it oc-curred in all groups. However, long-term ICP measurementin a water column resulted only in a 5.5% deviation after10 h whereas the ICP, after experimental head injury, raised209% in the sham, 293% in the mild and 280% in moderatetrauma group after 10 h. As such, drift of the probe alonecould not account for this significant rise.

Edema formation in tissue surrounding the implanted ICPprobe could have a mass effect and subsequently give rise tohigher ICP values. Therefore the presence of edema, asso-ciated with the probe traject, was histopathologically eval-uated in mildly traumatized rats. This was done on 50�mvibratome sections prepared at the level of the probe. Fluidaccumulation should result in a pale staining aspect of suchsections (Van Reempts and Borgers, 1990, 1994, 2000). Inedematous areas also the section thickness after drying onthe slide should be drastically reduced and the numeric den-sity of basophilic cell nuclei lowered. Presence of a normalhistologic staining picture, uniform section thickness andpreservation of normal cell density were in favour of the ab-sence of fluid accumulation at the moment that highest ICPvalues were recorded. Since the rats were freely moving inthe bowl, increased motor activity might provide an expla-nation of the ICP rise. However, visual observations duringmeasurement showed comparable activity of the animals inthe acute phase as well as in the late phase of the experi-ment. A plausible explanation could also be that trepanation

and/or implantation of the ICP probe might have elicited adelayed reaction, e.g. increase of cerebral blood volume orcontraction of the surrounding meninges, which both mightresult in ICP rise. Finally undefined pathological events af-ter neurotrauma could account for the high rise in ICP in thelate phase after neurotrauma. Elevated levels of ICP havebeen reported in the present model 24 h after neurotrauma(De Mulder et al., 1999; Engelborghs et al., 1998, 2000). Inthese studies ICP was only measured at that particular timepoint and the recorded values were considerably lower thanthe ones described in the present study. This might indi-cate that ICP, after having reached peak levels, progressivelylowers at later time points.

Among the various models of experimental head injury,the most commonly used are closed head injury and fluid per-cussion models (De Mulder et al., 1999; Engelborghs et al.,1998, 2000; Foda and Marmarou, 1994; Marmarou et al.,1994; McIntosh et al., 1987, 1989; Rooker et al., 2002). Afluid percussion model preferentially produces a focal braincontusion whereas an impact–acceleration injury results in amore diffuse challenge of the brain. Artefacts resulting fromthe experimental methodology, such as surge of blood pres-sure and craniotomy, not only complicate the findings, butthey also show biomechanical differences with most com-mon clinical trauma cases. Therefore an impact–accelerationmodel was chosen to induce a clinically relevant CHI. Ourmodel is characterised by several clinically relevant fea-tures, including increased ICP, diffuse axonal injury, con-tusions, impairment of cerebral blood flow autoregulation,and reduction of brain oxygenation (De Mulder et al., 1999;Engelborghs et al., 1998, 2000; Rooker et al., 2002). Mostlythe ICP is the parameter that is reliable, predictable and re-producible.

Volatile anaesthetics, like isoflurane, cause a dose depen-dent increase in cerebral blood flow and therefore they mayinfluence ICP as well (Lee et al., 1994). Due to this anaesthe-sia effect, interpretation of experimental ICP data after CHImay become difficult. Measurement in conscious animalsto our opinion might help to avoid this problem. This wassuccessfully achieved with the Bas-Ratturn® system. Nev-ertheless, several drawbacks had to be overcome. Cathetersfor BP measurement needed to be adapted to avoid occlu-sion and they had to be tunnelled subcutaneously to the skulland fixed to a screw in the rats skull. This small screw en-abled firm fixation and stabilisation of the ICP probe in thecortical bone of the rat and prevented its displacement inthe parenchyma. The special design of the Bas-Ratturn® al-lowed correct fixation of catheters and probes without riskfor coiling up or hindering the rat movements. A specialsensor in the swivel detects the direction of movement andsynchronously signals an opposite turning of the cage.

The main advantage of the present method is that it al-lows continuous observation of a critical and highly vari-able parameter in individual rats during an acute phase upto 3 h after head injury. To our knowledge such studieshave not been performed previously in awake rats. The yet

S. Rooker et al. / Journal of Neuroscience Methods 131 (2003) 75–81 81

unexplained rise in ICP after that period certainly requiresfurther investigation but does not necessarily have to be re-garded as a limitation. Assuming that it is a delayed sideeffect caused by the implantation of the ICP probe, it can beconsidered as a penetration injury that also mimics certainhuman situations. As such the phenomenon might be use-ful to study therapeutic measures that avoid development ofdelayed intracranial hypertension.

We conclude that the presented method could provide auseful tool to investigate ICP evolution after neurotraumawithout interference of long-term anaesthesia effects andtheir unpredictable effects on cerebral blood flow, autoregu-lation and intracranial pressure. Furthermore, such a methodcould be used to identify the optimal time point for pharma-cological interventions to decrease ICP in a clinically rele-vant manner.

References

Bullock R, Chesnut RM, Clifton G, et al. Guidelines for the managementof severe head injury. Brain Trauma Foundation. Eur J Emer Med1996;3:109–27.

Bullock R, Chesnut RM, Clifton G, et al. Guidelines for the managementof severe traumatic brain injury. J Neurotrauma 2000;17:451–627.

De Mulder G, Van Rossem K, Van Reempts J, et al. Validation of aclosed head injury model for use in long-term studies. Acta Neurochir1999;76(Suppl):409–13.

Engelborghs K, Verlooy J, Van Reempts J, et al. Temporal changes inintracranial pressure in a modified experimental model of closed headinjury. J Neurosurg 1998;89:796–806.

Engelborghs K, Haseldonckx M, Van Reempts J, et al. Impaired autoreg-ulation of cerebral blood flow in an experimental model of traumaticbrain injury. J Neurotrauma 2000;17:667–77.

Foda MA, Marmarou A. A new model of diffuse brain injury in rats. PartII. Morphological characterization. J Neurosurg 1994;80:301–13.

Goldstein M. Traumatic brain injury: a silent epidemic. Ann Neurol1990;27:327.

Guillaume J, Janny P. Manometrie intra-crannienne continue. La PresseMedicale 1951;59:953–5.

Lee JG, Hudetz AG, Smith JJ, et al. The effects of halothane and isofluraneon cerebrocortical microcirculation and autoregulation as assessed byLaser–Doppler flowmetry. Anesth Analg 1994;79:58–65.

Marmarou A, Foda MA, van den Brink WA, et al. A new model ofdiffuse brain injury in rats. Part I. Pathophysiology and biomechanics.J Neurosurg 1994;80:291–300.

McIntosh TK, Noble LJ, Nadrews B, Faden AL. Traumatic brain injuryin the rat: characterisation of a midline fluid percussion model. CentNerv Syst Trauma 1987;4:119–34.

McIntosh TK, Vink R, Noble LJ, et al. Traumatic brain injury in therat: characterization of a lateral fluid percussion model. Neuroscience1989;28:233–44.

Narayan R, Michel M. Clinical trials in head injury. J Neurotrauma2002;19(5):503–57.

Rooker S, De Visscher G, Van Deuren B, et al. Comparison of intracranialpressure measured in the cerebral cortex and the cerebellum of therat. J Neurosci Methods 2002;119:83–8.

Van Reempts J, Borgers M. Structural damage in experimental cerebralischemia. In: Schurr A, Rigor BM, editors. Cerebral ischemia andresuscitation. Boca Raton: CRC Press; 1990. p. 235–57.

Van Reempts J, Borgers M. Histopathological characterization of photo-chemical damage in nervous tissue. Histol Histopathol 1994;9(1):185–95.

Van Reempts J, Borgers M. Animal models of stroke: compromise be-tween consistency and clinical relevance? Neurosci Res Commun2000;26:161–72.