Incomplete Infarct and Delayed Neuronal Death After Transient Middle Cerebral Artery Occlusion in Rats
Background and Purpose The clinical syndrome of transient ischemic attacks is accompanied in a significant percentage of patients by brain lesions or neuroimaging abnormalities whose structural counterparts have not been defined. The objective of this study was to analyze, in an experimental model of short-term (<25 minutes) focal ischemia and long-term (≤28 days) reperfusion, the extent and nature of the structural abnormalities affecting neurons and glia located within the territory of the transiently occluded artery.
Methods Adult Wistar rats (n=121) had the origin of one middle cerebral artery (MCA) occluded with a nylon monofilament for periods of 10 to 25 minutes. Experiments of transient MCA occlusion were terminated at variable periods ranging from 1 day to 4 weeks. Control experiments consisted of (1) MCA occlusion without reperfusion (n=7) lasting 7 to 14 days and (2) sham operations (n=2) followed by 1- to 4-day survival. After in situ fixation, brain specimens were serially sectioned and subjected to detailed morphometric evaluations utilizing light and electron microscopes. The statistical method used to evaluate the results was based on ANOVA followed by Bonferroni’s corrected t test and Student’s t test comparisons.
Results Brain lesions were not detectable in the sham-operated controls. All brains with permanent MCA occlusion (7 to 14 days) had large infarctions with abundant macrophage infiltration and early cavitation. Forty-five (37%) of the experiments involving transient MCA occlusion had no detectable brain lesions after 4 weeks. Selective neuronal necrosis was found in 76 of 121 rats (63%) with transient MCA occlusion. Neuronal necrosis always involved the striatum, and in 29% of the brains with ischemic injury, necrosis also included a short segment of the cortex. In the striatum, the length of the arterial occlusion was the main determinant of the number of necrotic neurons (20 minutes [22.6±19] is worse than 10 minutes [4.9±7]) (P<.0001). In the cortex, the length of reperfusion determined the number of necrotic neurons appearing in layer 3. Experiments with reperfusion of 4 to 7 days’ duration yielded more necrotic neurons per microscopic field (2.02±3) than those lasting fewer days (0.04±0.1) (P<.05). The histological features of these lesions underwent continuous change until the end of the fourth week, at which time necrotic neurons were still visible both in the striatum and in the cortex.
Conclusions Arterial occlusions of short duration (<25 minutes) produced, in 76 of 121 experiments (63%), brain lesions characterized by selective neuronal necrosis and various glial responses (or incomplete infarction). This lesion is entirely different from the pannecrosis/cavitation typical of an infarction that appears 3 to 4 days after a prolonged arterial occlusion. Delayed neuronal necrosis, secondary to a transient arterial occlusion or increasing numbers of necrotic neurons in experiments with variable periods of reperfusion, was a response observed only at a predictable segment of the frontoparietal cortex.
Transient ischemic attacks are brief episodes of focal loss of brain function, thought to be caused by focal ischemia, that can usually be localized to that portion of the brain supplied by one vascular system and for which no other cause can be found. TIAs commonly last 2 to 15 minutes and are rapid in onset. Each TIA leaves no persistent deficit.1 TIAs probably have many causes, but their common denominator is presumed to be temporary interference followed by restoration of the normal blood flow to a focal area of the brain.2 The possibility that these episodes of neurological deficit may be accompanied by some as-yet-undefined structural brain injury was suggested by the type of observations made on 62 patients with hemispheric TIAs lasting from 2 minutes to 23 hours. Each of these patients was evaluated by MR at an undetermined time after the ictus but always after the neurological deficit had resolved completely. Acute “ischemic brain lesions,” appropriately located to explain the transient neurological symptoms, were identified in 19 (31%) of these patients.3 Similar observations concerning residual brain lesions detectable by CT have been made by several other investigators who report ischemic brain lesions with a frequency ranging from 12% to 48% among large groups of patients with TIA.4,5
The histopathologic features of these “silent brain lesions” and their etiology would be extremely difficult to characterize in human specimens because, among other reasons, by definition these patients recover their neurological integrity. Detecting as-yet-undefined brain lesions months or years after the TIA when an autopsy examination would become available is an extremely challenging undertaking.
The present study aimed to define the nature of the brain alterations that accompany transient arterial occlusions of less than 30 minutes’ duration; for this, we utilized an experimental model of short-term (10 to 25 minutes) MCA occlusion followed by a period of reperfusion of up to 28 days’ duration. Specifically, we wished to define the brain lesion in terms of its topography and dissimilarity with the features of a brain infarction (or area of pannecrosis). In addition, we wished to determine whether delayed neuronal necrosis of the type observed in the hippocampus after transient bilateral carotid artery clamping6 can be induced elsewhere in the brain by transient occlusion of one MCA.
The results of our experiments suggest that transient MCA occlusion (lasting 10 to 25 minutes) produces (in 63%) brain lesions that have histological features different from those of infarctions. Moreover, transient MCA occlusion induces SNN that appears promptly (12 hours) in the striatum; this is followed, a few days later, by delayed neuronal necrosis in selected regions of the cerebral neocortex.
Materials and Methods
All experiments were conducted according to the guidelines issued by the institutional Animal Care Committee, in compliance with regulations formulated by the US Department of Agriculture.7
One hundred thirty male Wistar rats (body weight, 270 to 290 g) purchased from Charles River Laboratories (Wilmington, Mass) and fed Agway rat chow during the 4 to 6 days of quarantine were used in this study. Animals were divided into three experimental groups (Table 1⇓). The first group had one MCA transiently occluded for periods of 10 to 25 minutes followed by reperfusion lasting between 24 hours and 28 days (n=121). Observations made on these subjects were compared with those derived from two other groups: one having the artery permanently occluded for 7 to 14 days (n=7) and two animals subjected to MCA occlusion of less than 1 minute’s duration, followed by reperfusion of either 1 or 4 days’ duration.
The surgical procedure is an adaptation of the method originally described by Koizumi et al8 and Zea-Longa et al.9 Details of the method used to occlude the artery are described elsewhere.10 Briefly, under general anesthesia (halothane and nitrous oxide) the origin of the right MCA was occluded by inserting through the external carotid artery a short segment (18±0.77 mm) of a 4–0 nylon monofilament. Reperfusion was achieved in the subjects with transient MCA occlusion by pulling out the filament until its tip became visible in the cervical segment of internal carotid.
All experiments were terminated under analepsis (ketamine and xylazine) by cardiovascular perfusion with paraformaldehyde fixative at a pressure of 100 mm Hg and according to methods described in detail elsewhere.10 After removing the brain and allowing overnight fixation in 4% paraformaldehyde, we made five coronal sections (each 2 to 3 mm thick) using a rat brain matrix (Activational System, Inc). The slabs were labeled A (frontal) through E (occipital), and after tissue processing each slab was embedded in paraffin. Approximately 6.0-μm-thick histology sections were obtained from each slab and stained with H&E. Histology sections from paraffin block slab B (caudal surface) were selected for H&E staining and immunohistochemical demonstration of GFAP (Dako). The rostral surface of slab C, corresponding to the level of the anterior commissure, was fixed in 3% glutaraldehyde and processed for electron microscopy. Additional details of the methods used in these procedures have been published elsewhere.10
Reperfusion describes the period of time elapsed between the arterial reopening and the time of death. Necrotic neurons were identified by light microscopy as exhibiting one or more of the following alterations: pyknosis, karyorrhexis, and karyolysis as well as cytoplasmic eosinophilia or loss of affinity for hematoxylin.11 By electron microscopy, necrotic neurons showed discontinuities in plasma or nuclear membranes and flocculent densities in the mitochondrial inner matrix.12,13SNN is characterized by irreversible injury (pyknosis/eosinophilia, karyorrhexis, or karyolysis) limited to specific populations of brain neurons that have been empirically identified as being vulnerable to hypoxia/ischemia.14,15 SNN was accompanied by reactive structural changes in astrocytes and microglia; these glial responses will be the subject of a separate scientific communication. Pannecrosis describes histological changes consisting of the loss of affinity for hematoxylin that affects simultaneously all cell types (neuronal, glial, and vascular). Infiltrating inflammatory cells (ie, leukocytes) may be the only cell types showing normal stainability to hematoxylin in areas of pannecrosis. This term is synonymous with infarction, which in the brain is defined as an area of pannecrosis involving a defined vascular territory.14
Quantitation of SNN
The surface of the area involved by the ischemic lesion (as outlined in a GFAP preparation) was classified as being mild (+), moderate (++), or marked (+++). Mild involvement (+) corresponds to a lesion in which scattered necrotic cells (neurons) are confined to the dorsolateral striatum. In specimens with moderate involvement (++), the surface of the area containing necrotic neurons equals less than 50% of the surface area of the striatum. Marked involvement (+++) defines specimens in which necrotic neurons are scattered over an area that exceeds 50% of the surface area of the striatum. Fig 1⇓ illustrates the differences that exist between lesions induced by permanent occlusion (Fig 1A⇓ and 1B⇓) and those secondary to transient occlusion (Fig 1C⇓, 1D⇓, and 1E⇓).
Necrotic neurons were counted in each of the 76 specimens having “brain lesions” in the territory of the transiently occluded MCA. Fifteen nonoverlapping microscopic fields at a magnification of ×600 were collected from each histology specimen with the use of a Sony computer-controlled display videocamera interfaced with an Olympus microscope system (Global Laboratory Image Data Translation). Five of the 15 fields examined were from the granular insular and parietal cortex; the additional 10 fields were from the lateral striatum. The numbers of necrotic neurons per microscopic field were evaluated with respect to different durations of MCA occlusion and different durations of reperfusion. In a randomly selected number of animals, electron microscopic evaluation served to further define the features of cells undergoing necrosis.
Results of individual counts of necrotic neurons were expressed as mean±SD values for each experimental group. Differences among groups with MCA occlusion of different duration were determined by the paired Student’s t test. ANOVA, followed by Bonferroni’s corrected t test, was used to determine differences among subgroups with variable times of reperfusion.
There were no detectable changes in the histological preparations of the experiments based on sham operations (MCA occlusion lasting less than 1 minute) followed by reperfusion periods of 1 to 4 days’ duration. All experiments (n=7) based on permanent MCA occlusion resulted in infarction or pannecrosis accompanied by early cavitation on day 7 and by advanced cavitation on day 14 (Figs 1A⇑, 1B⇑, and 2A⇓).
The histological changes induced by MCA occlusion of short duration (10 to 25 minutes) followed by periods of reperfusion lasting up to 28 days were significantly different from those observed in experiments in which the artery remained occluded for seven days. Histological abnormalities were not detectable in 45 (37.2%) of the experiments involving MCA occlusion/reperfusion. The percentage of histologically intact brains was higher in experiments of MCA occlusion/reperfusion in which occlusion lasted 10 minutes (46.4%) compared with those lasting 15 (24.8%) or 20 (38%) minutes (Table 2⇓).
Sixty-three percent of the experiments involving MCA occlusion/reperfusion resulted in neuronal and glial alterations that were scattered over a large area of the cerebral hemispheres corresponding to the level of the anterior commissure. These alterations primarily involved the striatum and the parietal cerebral cortex on the hemisphere ipsilateral to MCA occlusion/reperfusion. Three types of injury were observed in the striatum: (1) The mildest type of neuronal lesion consisted of SNN confined to the dorsolateral portion of the striatum; this is illustrated in Figs 1C⇑ and 2B⇑. This type of SNN was more common in the group with 10-minute MCA occlusion (86%) than in the others (Table 2⇑). (2) An intermediate type of lesion (Figs 1D⇑ and 2C⇑) was more common in the 15-minute MCA occlusion group (37%) than in the others (Table 2⇑). (3) SNN distributed over an area that exceeds 50% of the surface of the striatum is identified in Table 2⇑ as +++ and is illustrated in Figs 1E⇑ and 2D⇑. This type of lesion was more frequent in the group with MCA occlusion of 25 minutes’ duration than in the others (Table 2⇑).
SNN involving layer 3 within a short segment of the frontoparietal cortex was observed in 29 of the 76 specimens (38%) with MCA occlusion/reperfusion that had ischemic lesions (Fig 1C⇑ and 1E⇑). Brain hemorrhages were not detected in any of the 76 brains that had lesions attributable to the transient occlusion of one MCA. Selective neuronal injury was confirmed in most cases by electron microscopy that showed entirely different responses in neurons adjacent to one another (Figs 3A⇓ and 3B⇓).
Effect of MCA Occlusion of Different Durations
In the striatum mean numbers of necrotic neurons per microscopic field were more abundant in experiments of 15 to 20 minutes’ duration compared with those lasting 10 minutes (Fig 4⇓ and Table 3⇓). In contrast, mean numbers of necrotic neurons in the cortex did not differ among experimental groups with MCA occlusion of different durations (Fig 4⇓ and Table 3⇓).
Effect of Duration of Reperfusion
Mean numbers of necrotic neurons in the striatum were highest 24 hours after the artery was reopened (Fig 5⇓, Table 4⇓). The subsequent decrease in the number of necrotic neurons is interpreted as a reflection of the inability of the observer to identify, in a given microscopic field, each of the lethally injured cells. This was attributed to the fact that over a period of several days some of these necrotic cells disintegrate locally.
In the cerebral cortex, in contrast, the mean numbers of necrotic neurons were higher in experiments terminated 72 hours to 7 days after MCA occlusion/reperfusion (2.02±3) compared with those terminated after 24 to 48 hours (0.04±0.1) (P<.05) (Fig 6⇓; Table 4⇑).
These experiments demonstrate significant differences in the features of the brain lesion induced by permanent occlusion of the MCA compared with the partly necrotic, gliotic lesion produced by transient arterial occlusion. The characteristic pattern of neuronal damage after MCA occlusion/reperfusion follows a well-defined sequence, with neuronal necrosis developing promptly in the striatum and in a delayed fashion at a predictable location of the cerebral cortex. The degree of SNN is greater in the caudoputamen than in the cerebral cortex, and in the caudoputamen the extent of neuronal damage is proportional to the duration of the arterial occlusion. The cortical lesion does not appear to be influenced by the duration of MCA occlusion but rather by the time elapsed after the artery is reopened. An analogy may exist between the lesions produced in our experiments and the human syndromes attributed to a brief, episodic arterial occlusion. Incomplete infarcts are characterized by subtle morphological changes; thus, their clinical detection, imaging, and quantitation may require application of labeled radiotracers that specifically bind to neuronal components, such as the central benzodiazepine receptors.
Incomplete infarction developed in 76 of 121 Wistar rats with transient occlusion of the MCA (63%). We observed clear-cut differences between the pannecrosis (or infarction) typical of prolonged arterial occlusions, in which coagulation necrosis involves the entire territory of the occluded artery (Fig 1A⇑ and 1B⇑) and the selective necrosis of individual neurons with preservation of other cells characteristic of incomplete infarction (Fig 1C⇑, 1D⇑, and 1E⇑). The morphological differences between the two types of lesions were particularly apparent several days after the arterial occlusion, when the areas of pannecrosis underwent cavitation while the architecture of the incompletely infarcted tissues remained intact, except for the reactive gliosis that often could be subtle. Differences in the gross features of the brain lesions secondary to focal ischemia have been noted in human brains injured by MCA occlusion of undetermined duration16,17 and in nonhuman primates with experimental MCA occlusion of less than 4 hours’ duration followed by reperfusion of several weeks. The predominant brain lesion in these experiments of MCA occlusion/reperfusion was called selective necrosis.18
The chronological difference in the appearance of necrotic neurons at two sites of the brain, the striatum and a short segment of the cerebral cortex, suggests different mechanisms of neuronal injury at these two sites. In the striatum, the number of necrotic neurons per unit area was highest at 24 hours after MCA occlusion/reperfusion; only a few ghostlike structures remained visible after 21 days (Fig 2⇑). In contrast, necrotic neurons were not detectable in the cortex during the first 48 hours of reperfusion; these cells became clearly apparent only after 72 hours, and their numbers increased at the end of the first week. The apparent decreasing numbers of necrotic neurons after 7 days (Fig 3⇑) is attributed to the inability of the observer to identify partly disintegrating neurons. Nakano et al19 and Du et al20 have described SNN restricted to the dorsolateral striatum and to layer 3 of the neocortex in rats with short-term (approximately 15 minutes) MCA occlusion followed by long-term (4 weeks) reperfusion. In both experiments several days elapsed between the reopening of the artery and the appearance of necrotic neurons in the cortex.19,20 Du et al20 noted that brain infarction can develop in a surprisingly delayed fashion after transient focal ischemia. Our findings confirm the observations concerning the delayed appearance of necrotic neurons in the cortex after brief periods of arterial occlusion. However, infarcts or pannecrosis were not present in any of our specimens. It appears to us that Du et al20 used the word “infarct” to describe the lesion known to us as SNN accompanied by gliosis. Also, Du et al20 suggested that delayed neuronal death in the cortex may be mediated by apoptosis. We have not seen evidence of apoptosis in any of our electron microscopic preparations, but we provide statistical evidence in support of the observation that SNN occurs at a predictable site of the cerebral cortex in a delayed manner. This belated appearance of necrotic neurons in the cerebral cortex, which has also been documented after permanent MCA occlusion,21 is perhaps related to the persistent cortical hypoperfusion reported by Nagasawa and Kogure22 in experiments of MCA occlusion/reperfusion, as well as other as-yet-unknown factors.
In our experiments, the neuronal injury after transient MCA occlusion/reperfusion was more severe in the caudoputamen than in the cortex (Figs 1⇑ and 4⇑ and Tables 2⇑ and 3⇑). This difference may reflect the level of residual CBF at each of these sites. Occluding one MCA in Wistar rats induces changes in rCBF that are heterogeneous in their topographic distribution. During the occlusive period rCBF values are lower in the striatum than in the cortex.22 This difference also was demonstrated in a group of 10 Wistar rats with MCA occlusion of 1 hour’s duration followed by reperfusion of 15 minutes, in which Memezawa et al23 noted that under these conditions the lateral striatum is more severely ischemic than the frontoparietal cortex.
The incremental numbers of necrotic neurons that we observed in the caudoputamen with increasing time of arterial occlusion (Fig 4⇑) suggest that, after an arterial occlusion, the numbers of necrotic neurons at a given site are dictated by the severity of the local ischemia. This concept is supported by the results of several experiments of transient intracranial artery occlusion in which a direct correlation existed between the degree of histologically detectable injury (ie, number of necrotic neurons per unit area) and the severity of local ischemia, ie, percent of drop in rCBF.24–29 Moreover, at multiple sites within the territory supplied by the MCA, rCBF values are consistently lower 6 hours after MCA occlusion compared with 1 to 3 hours after the arterial occlusion.22 Curiously, however, this relationship does not apply to the cerebral cortex, where the number of necrotic neurons was about the same after 10, 15, or 20 minutes of MCA occlusion. This response may be related to the fact that despite reopening of the artery, a state of hypoperfusion persists in the cerebral cortex in experiments of MCA occlusion/reperfusion.22 In addition to the effects of regional CBF, selective neuronal injury may be the result of other factors, including intrinsic neuronal features. SNN has been observed in the dorsolateral striatum of Wistar rats exposed to transient forebrain ischemia followed by up to 8 days’ reperfusion, and in these experiments the necrotic neurons in the caudoputamen were mostly GABAergic, while the adjacent cholinergic neurons retained their structural integrity.29
By analogy with the human condition, we suggest that incomplete infarction may be the anatomic substrate of some TIAs. Our experimental design of short-term arterial occlusion resembles the type of focal ischemic episodes thought to be the cause of many TIAs. Furthermore, in previous work on the same experimental model, we have documented a close correlation (r=.951; P=.001) among duration of MCA occlusion, number of necrotic neurons, and severity of the neurological deficit30; in addition, several authors have demonstrated by CT and MRI residual brain lesions among large numbers of patients with TIA.3,5 Thus, we conjecture that some of the silent brain lesions observed in these patients may have features similar to those of the incomplete infarctions reported herein.
These incomplete infarcts are characterized by preservation of the tissue architecture and absence of cavitation. The spectrum of these lesions includes some that are subtle and difficult to see with the naked eye or to be depicted with conventional neuroimaging methods. Their demonstration may require the application of radiotracers that selectively bind to central benzodiazapine receptors. Sette et al31 demonstrated in macaques with MCA occlusion/reperfusion of several weeks’ duration two types of lesions. In the basal ganglia there were CT-visible cavitary infarcts. In addition, adjacent areas that on CT were isodense with the normal brain showed a 20% decrease in the uptake of flumazenil. These areas probably correspond to sites of incomplete infarct where early reperfusion may have prevented the progression to pannecrosis. Similarly, Nakagawara et al32 noted, in ischemic brain lesions among humans who had proven arterial reopening, topographic differences in the uptake of the radioligand iomazenil (almost identical to flumazenil) at sites where neuroimaging revealed no abnormality. This suggests that the extent of SNN in areas that appear normal on conventional CT or MRI may be quantifiable by methods that measure the uptake of these radioligands.
The term “incomplete infarction” describes brain lesions produced by moderate ischemia, such as may be induced by a transient arterial occlusion; this condition primarily injures neurons without inducing pannecrosis and subsequent cavitation. The term accurately refers to the “incomplete” necrosis of the tissue components and explicitly implicates focal ischemia in its pathogenesis. The term “SNN” applied to lesions that can be induced by hypoglycemia, cyanide, and carbon monoxide, among others, defines the type of anatomic lesion but does not indicate its pathogenesis.
Selected Abbreviations and Acronyms
|CBF||=||cerebral blood flow|
|GFAP||=||glial fibrillary acidic protein|
|H&E||=||hematoxylin and eosin|
|MCA||=||middle cerebral artery|
|rCBF||=||regional cerebral blood flow|
|SNN||=||selective neuronal necrosis|
|TIA||=||transient ischemic attack|
This study was supported by US Public Health Service grant NS31631 (Dr Garcia). The authors thank Karen Kuhrt (Detroit, Mich) for excellent secretarial support, as well as Jun Xu and Wenji Jiang for their technical expertise in the laboratory.
Presented at the 22nd International Joint Conference on Stroke and Cerebral Circulation of the American Heart Association, Anaheim, Calif, February 7, 1997, and published in abstract form (Stroke. 1997;28:247).
- Received July 3, 1997.
- Revision received August 7, 1997.
- Accepted August 20, 1997.
- Copyright © 1997 by American Heart Association
Special report from the National Institute of Neurological Disorders and Stroke. Stroke. 1990;21:637–676.
Millikan CH. The pathogenesis of transient focal cerebral ischemia. Circulation. 1965;32:438–450.
Fazekas F, Fazekas G, Schmidt R, Kapeller P, Offenbacher H. Magnetic resonance imaging correlates of transient cerebral ischemic attacks. Stroke. 1996;27:607–611.
Eliasziw M, Streifler JY, Spence JD, Fox, AJ, Hachinski VC, Barnett HJM. Prognosis for patients following a transient ischemic attack with and without a cerebral infarction on brain CT. Neurology. 1995;45:428–431.
Toole JF. The Willis lecture: transient ischemic attacks, scientific method, and new realities. Stroke. 1991;22:99–104.
Animal welfare: final rules (9 CFR parts 1, 2, and 3). Federal Register. August 31, 1989;54:36112–36163.
Koizumi J, Yoshida Y, Nakazawa T. Experimental studies of ischemic brain edema: a new experimental model in which recirculation can be reintroduced. Jpn J Stroke. 1986;8:1–8.
Zea-Longa E, Weinstein PR, Carlson S, Cummins R. Reversible middle cerebral artery occlusion without craniectomy in rats. Stroke. 1989;20:84–91.
Jennings RB, Ganote CE, Reimer KA. Ischemic tissue injury. Am J Pathol. 1975;80:179–187.
Graham DI. Hypoxia and vascular disorders. In: Adams JH, Duchen LW, eds. Greenfield’s Neuropathology. New York, NY: Oxford University Press; 1992;153–268.
Pulsinelli WA. Selective neuronal vulnerability and infarction in cerebrovascular disease. In: Welch KMA, Caplan LR, Reis DJ, Siesjö BK, Weir B, eds. Primer on Cerebrovascular Diseases. San Diego, Calif: Academic Press; 1997;104–107.
Spatz H. Pathologische Anatomie der Kreislaufstörungen. Z Neurol. 1939;167:301–324.
Scholz W. Die nicht zur Erweichung führenden unvollständigen Gewebsnecrosen (Elektive Parenchymnekrose). In: Lübarsch O, Rössle R, Henke F, eds. Handbuch der speziellen pathologischen Anatomie und Histologie, XIII: Band Nervensystem, 1: Teil, Bandteil B. Berlin, Germany: Springer-Verlag; 1957.
Du C, Hu R, Csernansky CA, Hsu CY, Choi DW. Very delayed infarction after mild focal cerebral ischemia: a role for apoptosis? J Cereb Blood Flow Metab. 1996; 16:195–201.
Garcia JH, Liu K-F, Ho K-L. Neuronal necrosis after middle cerebral artery occlusion in Wistar rats progresses at different time intervals in the striatum and the cortex. Stroke. 1995;26:636–643.
Nagasawa H, Kogure K. Correlation between cerebral blood flow and histologic changes in a new rat model of middle cerebral artery occlusion. Stroke. 1989;20:1037–1043.
Memezawa H, Smith M-L, Siesjö BK. Penumbral tissues salvaged by reperfusion following middle cerebral artery occlusion in rats. Stroke. 1992;23:552–559.
Marcoux FW, Morawetz RB, Crowell RM, DeGirolami U, Halsey JH. Differential regional vulnerability in transient focal cerebral ischemia. Stroke. 1982;13:339–346.
Mies G, Auer LM, Ebhart G, Traupe H, Heiss WD. Flow and neuronal density in tissue surrounding chronic infarction. Stroke. 1983;14:22–27.
Garcia JH, Wagner S, Liu K-F, Hu X-J. Neurological deficit and the extent of neuronal necrosis attributable to middle cerebral artery occlusion: statistical validation. Stroke. 1995;26:627–635.
Sette G, Baron J-C, Young AR, Miyazawa H, Tillet I, Barré L, Travère J-M, Derlon J-M, MacKenzie ET. In vivo mapping of brain benzodiazepine receptor changes by positron emission tomography after focal ischemia in the anesthetized baboon. Stroke. 1993;24:2046–2058.
Nakagawara J, Sperling B, Lassen NA. Incomplete brain infarction of reperfused cortex may be quantitated with iomazenil. Stroke. 1997;28:124–132.