Anterior communicating artery division in the endoscopic endonasal translamina terminalis approach to the third ventricle: an anatomical feasibility study.

BACKGROUND: Endonasal endoscopic approaches (EEA) to the third ventricle are well described but generally use an infrachiasmatic route since the suprachiasmatic translamina terminalis corridor is blocked by the anterior communicating artery (AComA). The bifrontal basal interhemispheric translamina terminalis approach has been facilitated with transection of the AComA. The aim of the study is to describe the anatomical feasibility and limitations of the EEA translamina terminalis approach to the third ventricle augmented with AComA surgical ligation. METHODS: Endoscopic dissections were performed on five cadaveric heads injected with colored latex using rod lens endoscopes attached to a high-definition camera and a digital video recorder system. A stepwise anatomical dissection of the endoscopic endonasal transtuberculum, transplanum, translamina terminalis approach to the third ventricle was performed. Measurements were performed before and after AComA elevation and transection using a millimeter flexible caliper. RESULTS: Multiple comparison statistical analysis revealed a statistically significant difference in vertical exposure between the control condition and after AComA elevation, between the control condition and after AComA division and between the AComA elevation and division (p < 0.05). The mean difference in exposed surgical area was statistically significant between the control and after AComA division and between elevation and AComA division (p < 0.01), whereas it was not statistically significant between the control condition and AComA elevation (NS). CONCLUSION: The anatomical feasibility of clipping and dividing the AComA through an EEA has been demonstrated in all the cadaveric specimens. The approach facilitates exposure of the suprachiasmatic optic recess within the third ventricle that may be a blind spot during an infrachiasmatic approach.

Microsurgical anatomy of the central core of the brain.

OBJECTIVE The purpose of this study was to describe in detail the cortical and subcortical anatomy of the central core of the brain, defining its limits, with particular attention to the topography and relationships of the thalamus, basal ganglia, and related white matter pathways and vessels. METHODS The authors studied 19 cerebral hemispheres. The vascular systems of all of the specimens were injected with colored silicone, and the specimens were then frozen for at least 1 month to facilitate identification of individual fiber tracts. The dissections were performed in a stepwise manner, locating each gray matter nucleus and white matter pathway at different depths inside the central core. The course of fiber pathways was also noted in relation to the insular limiting sulci. RESULTS The insular surface is the most superficial aspect of the central core and is divided by a central sulcus into an anterior portion, usually containing 3 short gyri, and a posterior portion, with 2 long gyri. It is bounded by the anterior limiting sulcus, the superior limiting sulcus, and the inferior limiting sulcus. The extreme capsule is directly underneath the insular surface and is composed of short association fibers that extend toward all the opercula. The claustrum lies deep to the extreme capsule, and the external capsule is found medial to it. Three fiber pathways contribute to form both the extreme and external capsules, and they lie in a sequential anteroposterior disposition: the uncinate fascicle, the inferior fronto-occipital fascicle, and claustrocortical fibers. The putamen and the globus pallidus are between the external capsule, laterally, and the internal capsule, medially. The internal capsule is present medial to almost all insular limiting sulci and most of the insular surface, but not to their most anteroinferior portions. This anteroinferior portion of the central core has a more complex anatomy and is distinguished in this paper as the "anterior perforated substance region." The caudate nucleus and thalamus lie medial to the internal capsule, as the most medial structures of the central core. While the anterior half of the central core is related to the head of the caudate nucleus, the posterior half is related to the thalamus, and hence to each associated portion of the internal capsule between these structures and the insular surface. The central core stands on top of the brainstem. The brainstem and central core are connected by several white matter pathways and are not separated from each other by any natural division. The authors propose a subdivision of the central core into quadrants and describe each in detail. The functional importance of each structure is highlighted, and surgical approaches are suggested for each quadrant of the central core. CONCLUSIONS As a general rule, the internal capsule and its vascularization should be seen as a parasagittal barrier with great functional importance. This is of particular importance in choosing surgical approaches within this region.

Endovascular therapy for vasospasm after aneurysmatic subarachnoid hemorrhage.

INTRODUCTION: Balloon angioplasty and/or selective intra-arterial vasodilator therapies are treatment options in patients with vasospasm after subarachnoid hemorrhage (SAH). We analyzed the effect of balloon angioplasty and/or selective intra-arterial vasodilator therapy in our patients. METHODS: Twenty-six patients (vasodilation group, VDT) were treated with intra-arterial nimodipine. The balloon angioplasty with nimodiopine-group (BAP-N group) comprised 21 patients. The primary endpoint of this study was successful angiographic vessel dilation in vasospastic vessels after balloon angioplasty, together with nimodipine (BAP-N group), compared to intra-arterial vasodilator therapy (VDT group) with nimodipine alone. RESULTS: A significant effect of angioplasty plus nimodipine was found in the central arteries (composite endpoint) with an OR of 2.4 (95% CI: 1.4-4.2], p = 0.002), indicating a chance of improvement of the BAP-N group of more than twice compared to nimodipine infusions alone. Significant advantages for BAP-N-therapy were also encountered in the internal carotid artery (OR 5.4, p < 0.001) and basilar artery (OR 29.7, p = 0.003). A joint analysis of all arteries combined failed to show significant benefit of BAP-N therapy (OR 1.5, p = 0.079), which was also true for cerebral peripheral arteries (OR 0.77, p = 0.367). There was no difference in clinical outcome between both groups. CONCLUSIONS: In SAH patients with vasospasm, a combination therapy of balloon angioplasty and intra-arterial nimodipine resulted in a more than doubled vasodilative effect in the central cerebral arteries compared to the sole infusion of nimodipine. Regarding the ICA and BA arteries, this beneficial effect was even more pronounced. Although there was a tendency of better effects of the BAP-N group, regarding the overall effect in all territories combined, this failed to reach statistical evidence. In cerebral peripheral arteries, no differences were observed, and there was no difference in clinical outcome, too.

Microsurgical anatomy of perforated arteries in the hypothalamic area.

AIM: This study aimed to investigate the microsurgical anatomy of perforating arteries in the hypothalamic area, which are associated with diabetes insipidus. MATERIAL AND METHODS: A total of 20 adult cadaver heads soaked in formalin were infused with red latex through the carotid artery and vertebral artery, and supplementary perfusion was performed after 1 day. RESULTS: The perforating arteries in the hypothalamic area could be divided into three groups according to their origins, namely, the former, below and outside groups. The former group mainly comprised the perforating arteries near the current communicating arteries. The outside group comprised the perforating arteries from the upper clinoid segment of the internal carotid and posterior communicating arteries. The below group comprised the bottom hypophyseal arteries of the cavernous segment from the internal carotid artery. CONCLUSION: Vascular injuries that occur during surgery can be minimised by understanding the distribution of the aforementioned vessels.

Anticoagulation with dabigatran does not increase secondary intracerebral haemorrhage after thrombolysis in experimental cerebral ischaemia.

Dabigatran etexilate (DE) has recently been introduced for stroke prevention in atrial fibrillation, but management of acute ischaemic stroke during therapy with DE is a challenge. Thrombolysis is contraindicated because of a presumed increased risk of intracerebral haemorrhagic complications. We studied in different ischaemia models whether DE increases secondary haemorrhage after thrombolysis. C57BL/6 mice were anticoagulated with high-dose DE or warfarin. After 2 hour (h) or 3 h transient filament MCAO, rt-PA was injected. At 24 h after MCAO, secondary haemorrhage was quantified using a macroscopic haemorrhage score and haemoglobin spectrophotometry. Post-ischaemic blood-brain-barrier (BBB) damage was assessed using Evans blue. To increase the validity of findings, the duration of anticoagulation was prolonged in mice (5 x DE over 2 days), and the effect of DE after thrombolysis was also examined in thromboembolic MCAO in rats.Pretreatment with warfarin resulted in significantly more secondary haemorrhage (mean haemorrhage score 2.6 ± 0.2) compared to non-anticoagulated animals (1.7 ± 0.3) and DE (9 mg/kg, 1.6 ± 0.3) in 2 h ischaemia. Also after a 3 h period of ischaemia, haemorrhage was more severe in animals anticoagulated with warfarin compared to 9 mg/kg DE and non-anticoagulated control. Prolonged or enteral dabigatran pretreatment led to identical results. Also, thrombolysis after thromboembolic MCAO in rats did not induce more severe bleeding in DE-treated animals. Mice pretreated with warfarin had higher BBB permeability and increased activation of matrix-metalloproteinase 9. In conclusion, DE does not increase the risk of secondary haemorrhage after thrombolysis in various rodent models of ischaemia and reperfusion. The implications of this finding for stroke patients have to be determined in the clinical setting.

Arterial spin-labeling magnetic resonance imaging after revascularization of moyamoya disease.

Arterial spin labeling (ASL) magnetic resonance imaging (MRI) is a technique for depicting cerebral perfusion without contrast medium. The purpose of this study was to determine whether ASL can be used to detect hyperperfusion after revascularization for moyamoya disease as effectively as N-isopropyl-[123I]ß-iodoamphetamine ((123)I-IMP) single-photon emission computed tomography (SPECT). Fifteen consecutive patients with moyamoya disease were included in the study. All patients underwent surgical revascularization. Postoperatively, regional cerebral blood flow (rCBF) was measured by flow-sensitive alternating inversion recovery (FAIR) ASL and (123)I-IMP SPECT during the acute stage, and rCBF of the operative side was compared with the other side. The asymmetry ratio (AR) was then calculated from the rCBF as measured using each modality. The postoperative AR of ASL was moderately correlated with that of (123)I-IMP SPECT (y = 0.180x + 0.819; R = 0.80; P = .0003). In this series, 2 patients (13.3%) suffered symptomatic hyperperfusion after revascularization and accordingly exhibited increased AR of ASL. Our data indicate that early increases in rCBF in patients with hyperperfusion could be detected using FAIR ASL supplemental to (123)I-IMP SPECT after revascularization. Our data indicate that FAIR ASL is a convenient method for evaluating hyperperfusion that can be performed repeatedly without the use of contrast medium or radioisotopes.

Surgical treatment for musicogenic epilepsy.

We report a patient with medically intractable musicogenic epilepsy (ME) who was treated with surgery. Using the non-invasive methods of ictal and interictal electroencephalography (EEG), MRI, interictal single photon emission computed tomography and clinical manifestations, we first localized the musicogenic seizures (MS). The ictal onset zone was then further localized using intracranial EEG to the middle part of the left superior temporal gyrus. Surgical resection of the epileptogenic zone was then performed. The patient had two seizures within 2 weeks post-operatively, but has then had no seizures during the following year (Engel class II). The results suggest that patients who have medically intractable ME combined with unilateral ictal onset zones should be considered for the surgical treatment of epilepsy.

Hyperbaric oxygen reduces blood-brain barrier damage and edema after transient focal cerebral ischemia.

BACKGROUND AND PURPOSE: Hyperbaric oxygen (HBO) has been shown to protect the brain parenchyma against transient focal cerebral ischemia, but its effects on the ischemic microcirculation are largely unknown. We examined the potential of HBO to reduce postischemic blood-brain barrier (BBB) damage and edema. METHODS: Wistar rats and C57/BL6 mice underwent occlusion of the middle cerebral artery (MCAO) for 2 hours. Forty minutes after filament introduction, animals breathed either 100% O2 at 3.0 atmospheres absolute (ata; HBO group) or at 1.0 ata (control) for 1 hour in an HBO chamber. In rats, MRI was performed 15 minutes after MCAO and after 15 minutes and 3, 6, 24, and 72 hours of reperfusion. In mice, BBB permeability for sodium fluorescein was measured after 24-hour reperfusion. RESULTS: Increased BBB permeability on postcontrast T1-weighted (T1w) images had a biphasic pattern. HBO reduced volumes and intensity of enhancement. Mean abnormal enhancing volumes were 71+/-10 mm3 (control) versus 47+/-10 mm3 (HBO) at 15 minutes; 111+/-21 mm3 versus 69+/-17 mm3 3 hours; 147+/-44 mm3 versus 83+/-21 mm3 6 hours; 150+/-37 mm3 versus 89+/-14 mm3 24 hours; and 322+/-52 mm3 versus 215+/-21 mm3 72 hours (all P<0.05). Interhemispheric quotients of mean gray values on T1w were at 1.73+/-0.11 versus 1.57+/-0.07 15 minutes; 1.74+/-0.07 versus 1.60+/-0.06 at 3 hours; 1.77+/-0.07 versus 1.62+/-0.06 at 6 hours; 1.79+/-0.10 versus 1.60+/-0.05 at 24 hours; and 1.81+/-0.10 versus 1.62+/-0.07 at 72 hours (all P<0.05). HBO-treated mice had significantly lower postischemic BBB permeability than mice treated with either normobaric hyperoxia or room air. Vasogenic edema assessed on T2w images and histologic sections was significantly lower in HBO-treated rats. CONCLUSIONS: Intraischemic HBO therapy reduces early and delayed postischemic BBB damage and edema after focal ischemia in rats and mice.

Branches of the anterior cerebral artery near the anterior communicating artery complex: an anatomic study and surgical perspective.

The anatomy of the branches of the anterior cerebral artery (ACA) near the anterior communicating artery (ACoA) complex were investigated to minimize neurovascular morbidity caused by surgical procedures performed in this region. Thirty-one cadaver brains were perfused with colored silicone, fixed, and studied under the operating microscope. The recurrent artery of Heubner (RAH), orbitofrontal artery (OFA), and frontopolar artery (FPA) were identified as the branches of the ACA arising near the ACoA complex. The OFA and FPA were identified in all hemispheres. Forty-nine (64%) of a total of 77 RAHs arose from the A2 segment. The OFA always arose from the A2 segment, was consistently the smallest branch, and coursed to the gyrus rectus, olfactory tract, and olfactory bulb. The mean distance between the ACoA and the OFA was 5.96 mm. The FPA arose from the A2 segment in 95% of the specimens, and coursed to the medial subfrontal region. The mean distance between the ACoA and the FPA was 14.6 mm. The RAH, OFA, and the FPA are three branches that arise from the ACA near the ACoA complex. These vessels have similar diameters, but can be distinguished by the final destination. Distinguishing these vessels is important since the consequences of injury or occlusion of the FPA and OFA are significantly less than of the RAH.

Surgery for deeply seated arteriovenous malformation: with special reference to thalamic and striatal arteriovenous malformation.

Surgery for deeply seated arteriovenous malformation (AVM) is controversial because stereotactic irradiation is applicable to the lesion. We have, however, experienced 30 deeply seated AVMs treated by direct surgery and/or endovascular treatment. The present study shows profiles of those patients and results of surgery. They include AVM in the thalamus in 12 cases, striatum in four cases, paraventricular area in five cases, medial temporal lobe in three cases, intraventricular area in three cases, and other regions in three cases. They were treated by surgery alone in 23 cases, embolization followed by surgery in four cases, and embolization alone in three cases. AVM in the mediodorsal thalamus and fornix (5 cases) was best treated by transcallosal approach. Venous aneurysm was commonly found in the AVM of this region and was a good navigator to the AVM. Pulvinar AVM was accessible through posterior interhemispheric approach (2 cases). None of these cases had additional neurological deficits. Cadaver dissection was useful for acquisition of surgical approach. Striatal AVM was approached through hematoma cavity with minimal manipulation to the surrounding structures, yet two of four cases showed progression of their weakness. The present study indicates that thalamic AVM can be approached surgically with careful selection of the approach. On the other hand, striatal AVM is not a good candidate for direct surgery and better treated by stereotactic irradiation.

Thread occlusion but not electrocoagulation of the middle cerebral artery causes hypothalamic damage with subsequent hyperthermia.

Moderate changes in body temperature can influence the outcome of cerebral ischemic insults and the effect of drugs. Body temperature was measured continuously for 24 hours in rats subjected to permanent occlusion of the middle cerebral artery (MCA) by either coagulation or thread insertion, and the results correlated with the histology of the hypothalamus. The body temperature did not change after MCA occlusion by coagulation and the hypothalamus was intact in all rats. In contrast, the body temperature rapidly increased from about 38 degrees C to more than 39.5 degrees C after MCA occlusion using intraluminal thread, and hyperthermia continued for at least 6 hours in all rats. Histological evaluation revealed neuronal damage in the preoptic area of the hypothalamus in all rats undergoing thread occlusion. Long duration hyperthermia must be prevented after permanent MCA occlusion when the intraluminal thread occlusion model is used in chronic experiments.

Effectiveness of revascularization surgery evaluated by proton magnetic resonance spectroscopy and single photon emission computed tomography.

Proton magnetic resonance spectroscopy (MRS) and single photon emission computed tomography (SPECT) were used to evaluate chronic ischemic regions in 26 stroke patients before and 1, 3, and 6 months after revascularization surgery. The volume of interest for proton MRS was placed in an area including part of the frontal and temporal opercula, insular cortex, and basal ganglia. Twenty healthy volunteers served as controls for proton MRS. Patients were divided into three groups according to the preoperative proton MRS. Group A (n = 12) had significantly lower N-acetylaspartate/choline (NAA/Cho) and N-acetylaspartate/creatine (NAA/Cr) ratios on the operative side compared to those on the contralateral side, and also lower than those in normal subjects. In seven patients in Group A, postoperative serial proton MRS demonstrated no recovery of these ratios on the operative side. However, proton MRS of the other five patients indicated gradual improvement in these ratios on the operative side at 3 to 6 months after surgery, and SPECT indicated an increase in cerebral blood flow on the operative side in four of these five patients. In Group B (n = 9), proton MRS and SPECT showed no laterality before revascularization and no remarkable change during the postoperative course. In Group C (n = 5), NAA/Cho or NAA/Cr decreased on the contralateral side preoperatively. Two patients showed fluctuating values of NAA/Cho or NAA/Cr during the postoperative period. Serial proton MRS and SPECT studies may be useful for the evaluation of revascularization surgery on ischemic regions. The efficacy of revascularization surgery on the metabolism may appear gradually within 3-6 months.

A mouse model of focal cerebral ischemia for screening neuroprotective drug effects.

The aim of the present study was to investigate if the infarct area on the brain surface after middle cerebral artery (MCA) occlusion in the mouse is representative for the infarct volume and if this determination of brain injury can be used for screening neuroprotective drug effects. Cerebral infarction was induced by coagulating electrically the stem of the left MCA. After 48 hr, the brains were perfused with carbon black and the unstained infarct area was determined by means of an image analyzing system. The infarct volume was determined by calculating the infarct area on coronal slices and the distance between succeeding slices. The correlation between the area and the volume of infarction was significant (r = 0.81; p less than 0.001). N-methyl-D-aspartate (NMDA) antagonists, calcium antagonists, 5-hydroxytryptamine-1A (5-HT-1A) agonists, radical scavengers, and various drugs were investigated in the mouse model of MCA occlusion. Drugs were usually applicated before ischemia. The drugs that were found to be neuroprotective in the mouse model revealed similar effects in rat models of focal or global cerebral ischemia. These findings show that the presented mouse model with its simple technique of measuring the infarct size is suitable for screening purposes.

The thalamogeniculate perforators of the posterior cerebral artery: the microsurgical anatomy.

The thalamogeniculate (TG) arteries of 30 forebrain hemispheres were examined. These vessels varied from 2 to 12 in number (mean, 5.7), and from 70 to 580 microns in caliber (mean, 345.8 microns). The average caliber of all the TG vessels per posterior cerebral artery ranged from 700 to 3400 microns (mean, 1972 microns). The TG arteries most often originated as individual vessels; however, in 26.67% of the hemispheres examined they shared a common site of origin, and 33.33% of the hemispheres they arose from common stems. The common stems ranged from 320 to 800 microns in diameter (mean, 583 microns). The TG branches arose from the crural or ambient (P2) segment of the posterior cerebral artery in 80% of the hemispheres, from the P2 and the quadrigeminal (P3) segment in 20%, from both the distal segment of the posterior cerebral artery and the common temporal artery (13.33%), or from the distal segment and either the calcarine (3.33%) or parieto-occipital artery (3.33%). The TG arteries usually penetrated the medial geniculate body (100%), pulvinar thalami (80%), brachium of the superior colliculus (53.33%), or lateral geniculate body (13.33%). The collateral branches of the TG arteries were noted to reach the medial geniculate body (76.67%), pulvinar (70%), brachium of the superior colliculus (40%), crus cerebri (40%), and lateral geniculate body (6.67%). The anastomoses were present in 66.67%, usually between the TG vessels and the medial posterior choroidal artery (33.33%), or the mesencephalothalamic artery (26.67%). They ranged in number from 1 to 3 (mean, 1.2), and in caliber from 90 to 400 microns (mean, 197 microns).(ABSTRACT TRUNCATED AT 250 WORDS)

Therapeutic effects of hyperbaric oxygenation on acute focal cerebral ischemia in rats.

The effects of hyperbaric oxygenation on acute focal cerebral ischemia in rats were investigated. All rats suffered 4-hour middle cerebral artery occlusion. Nontreated controls had a 27.9% +/- 5.5% infarct volume, and a left/right hemispheric volume ratio of 109% +/- 3%. Animals treated between 2.5 and 3.5 hours following occlusion had an 18.1% +/- 9.7% infarct volume (p less than 0.01), and a 105% +/- 3% left/right hemispheric volume ratio (p less than 0.001). In conclusion, at least until 4 hours following an ischemic insult, hyperbaric oxygenation reduces ischemic neuronal injury and brain edema following middle cerebral artery occlusion in rats treated between 2.5 and 3.5 hours following occlusion.

Branches of the anterior communicating artery. Microsurgical anatomy.

The anterior communicating artery (ACoA) and its branches were examined in 22 human brains after injecting Indian ink or methylmethacrylate. The ACoA branches were divided into the small and the large. Small branches were from 1 to 5 in number (mean 2), and from 70 to 270 microns in diameter (mean 151 microns). Seventy-six percent of the branches originated directly from the ACoA. They tended to arise closer to the left than to the right anterior cerebral artery. Fourteen percent of them arose from the junctional site of the ACoA with the anterior cerebral arteries, and 10% from the site of origin of the subcallosal artery. Large branches were identified as the median artery of the corpus callosum, and the subcallosal artery, respectively. The former vessel was present in 9% of the patients, and the latter in 91%. The subcallosal artery was from 320 to 640 microns in size (mean 486 microns). It tended to arise from the middle of the ACoA. In spite of the very frequent anastomoses involving the ACoA branches, care must be taken to avoid injury to these important vessels during operations of the ACoA aneurysms.

Morbidity and mortality after early aneurysm surgery--a prospective study with nimodipine prevention.

Based on the outcome in 116 consecutive patients who were subjected to early aneurysm operation combined with additional nimodipine treatment, and who were controlled by transcranial Doppler (TCD) sonography, a morbidity and mortality analysis was performed. Of the 84 patients who preoperatively were in Hunt & Hess grades III, 79 patients (94%) were considered to show a favourable (good-fair) late recovery, while one patient (1%) had a poor outcome, and four patients (5%) died. Of the 32 poor condition patients (H & H IV-V), 17 (53%) showed a favourable recovery, while seven (22%) had a poor outcome, and eight patients (25%) died. Altogether, 20 patients (17%) had an unfavourable (poor-dead) outcome. Only two of these patients showed delayed ischaemic deterioration, one as a consequence of a secondary occlusion of perforating branches from the basilar artery and one with decompensated vasospasm after the evacuation of an epidural haematoma and a longlasting severe systemic hypotension; both these patients died. In another six of the patients with an unfavourable outcome, this was mainly related to a complicated surgery. The unfavourable outcome was related to primary brain damage produced by the subarachnoid haemorrhage (SAH) in ten patients and in two patients to internal medical complications. In addition to the two patients who died following delayed deterioration, secondary neurological dysfunction occurred in 11 patients. In 10 of these patients transient neurological dysfunction was attributed to vasospasm or to a combination of vasospasm with intraoperative or postoperative complications. One further case of delayed deterioration was attributed to secondary occlusion of the internal carotid artery after a complicated operation.(ABSTRACT TRUNCATED AT 250 WORDS)

Unruptured fusiform aneurysms of the posterior circulation with thalamic infarction.

Three patients with unruptured fusiform aneurysms of the posterior circulation presented with nonhemorrhagic thalamic infarctions. All of the aneurysms were seen on enhanced computed tomographic (CT) scans preangiographically. Although unruptured fusiform aneurysms are probably a rare cause of nonhemorrhagic thalamic infarction, their importance lies in the therapeutic implications of this diagnosis. In patients with nonhemorrhagic thalamic infarction, we suggest careful scrutiny of the blood vessels on enhanced CT scans.

Hyperbaric oxygen as an adjunct to acute revascularization of the brain.

Two recent cases suggest that hyperbaric oxygen may be an important adjunct to the surgical treatment of occlusion of major cerebral arteries within the first few hours after onset of neurological deficit. In both patients, one with an embolus to the right middle cerebral artery and one with a surgical occlusion of the left internal carotid artery, circulation to the ischemic area was restored more than eight hours after occlusion. In the patient with the middle cerebral artery embolus, hemiplegia cleared after a six-minute exposure to hyperbaric oxygen. The patient with occlusion of the internal carotid artery was revascularized by anastomosis of a superficial temporal artery less than 1 mm in diameter to a branch of the middle cerebral artery. Her hemiplegia and aphasia cleared rapidly and concomitantly with intermittent exposure to hyperbaric oxygen during the first nine postoperative days. Postoperative angiograms demonstrated patency in both cases. The implications of these observations are discussed.