There are different types of ischemic injury.  Ischemia can affect the entire brain, or just a particular vascular distribution.  It can be brief, as in a TIA, or it can endure.  The patterns of damage we see in gross and microscopic pathology reflect the focality and temporal course of the ischemia.

Like other organs, the brain is dependent on a constant supply of oxygen and nutrients to maintain its normal function.  However, the brain differs from other organ systems in three major ways: 1. Neurons are incapable of surviving any but the most transient interruptions in blood supply. 2. Small areas of damage are capable of producing significant functional deficits. 3. Due to the rigidity of the skull, increased material within the brain (blood or edema, for example) can cause devastating secondary damage. A reduction of, or disruption in, blood flow to the brain is the primary cause of a stroke. Reductions in cerebral perfusion even for even a short period of time can be disastrous, resulting in brain damage or even death. Normal cerebral function can continue for only 8-10 seconds following cerebral ischemia, and irreversible damage occurs after about 6-8 minutes. The site and size of the lesion and, consequently, the clinical and pathological picture it produces are dependent on a large number of modifying factors, including collateral circulation, the duration of the ischemia, the degree and the rapidity of the reduction of flow, and reperfusion injury.

Experimental data indicates that brain energy metabolism is maintained for several minutes even at pronounced degrees of hypoxemia, due to an autoregulatory increase in cerebral blood flow. In contrast, in experimental ischemia, the severity and extent of ischemic injury to neurons correlate extremely well with the percent drop in local cerebral blood flow. It therefore appears that hypoxia in the presence of preserved cerebral blood flow, while producing functional deficits (e.g. at high altitudes) is generally insufficient to produce neuronal necrosis. On the other hand, the brain is highly susceptible to damage consequent to reductions in cerebral blood flow.

Global cerebral ischemia typically produces one of two patterns of brain injury. With severe transitory cerebral hypoperfusion (e.g. cardiac arrest with resuscitation), it becomes apparent that certain cell populations and brain regions are more susceptible to cell death and necrosis. In general, neurons are more susceptible than glial cells. Among the various neuronal populations, the pyramidal neurons of Sommer’s sector (CA-1) of the hippocampus, the Purkinje cells of the cerebellum, and the pyramidal neurons of the cerebral cortex are selectively vulnerable.

Current evidence favors an excess of excitotoxic neurotransmitter receptors as the mechanism for selective vulnerability in the hippocampus. In the cerebral cortex, the neuronal loss and gliosis are more pronounced in some layers than others, and may produce a pattern of laminar (or pseudolaminar) necrosis in survivors of global cerebral ischemia. Several hours are needed for the neurons to develop the characteristic findings of ischemic neuronal death (cytoplasmic eosinophilia, loss of Nissl substance, nuclear pyknosis). While neurons die rapidly after critical ischemia, morphologic findings characteristic of ischemia at the resolution of the light microscope will not be discernable if the patient dies soon after the inciting event.

The second pattern of brain injury, usually seen in patients after severe systemic hypotension, is arterial border zone "watershed” infarction.  This affects the areas of the brain which are the most peripheral to the arterial circulation, presumably because they are the most easily deprived of adequate oxygenation.

The developing brain differs from the adult in its selective vulnerability. Because of the relatively high metabolic activity of the white matter during development (active myelination), white matter may be selectively or predominantly damaged with ischemic neonatal insults. This pattern of damage is often termed periventricular leukomalacia (literally, softening of the white matter near the ventricles), or perinatal telencephalic leukoencephalopathy (something bad happening to the cerebral white matter around the time of birth).

Cerebral infarction is the most common form of stroke, and occurs when the blood supply to a region of the brain is interrupted.  This is most often due to occlusion of the vascular lumen by clotted blood, either formed in situ (thrombosis) or formed elsewhere followed by migration into the cerebral circulation (thromboembolism).  While either event can occur in any vessel, extracranial and proximal intracranial arteries tend to undergo thrombosis (mostly related to atherosclerotic disease) while more distal arterial branches are more likely to be affected by emboli. 

These vascular occlusions produce territorial infarcts whose topology and symptomatology depends on the vascular anatomy of the brain. The selective cellular vulnerability and necrosis operative in global ischemia is usually not relevant in focal infarcts because the local severity of ischemia is much more severe (e.g. complete occlusion of a vessel for a prolonged time). The result is necrosis involving all of cellular elements of the affected area (usually excepting the capillary endothelial cells), rather than death of a single cell type within a region.

Acutely, cerebral infarcts are characterized by edema and eosinophilic neuronal degeneration, similar to that seen in vulnerable regions during global ischemia. Although glial cells also undergo necrosis within a region of infarction, they do not manifest striking morphologic abnormalities at the light microscopic level. The death of these cellular elements becomes apparent during organization of the infarct, when in contrast to the reactive astrocytic scar that replaces necrotic neurons in global ischemia, infarcted brain tissue begins liquefying as macrophages migrate into the lesion through the relatively intact capillary bed.  Within about a week, the infarct is characterized by scant residual eosinophilic neurons, abundant macrophages, and reactive astrocytosis in the adjacent non-infarcted brain.   As the infarct ages, the astrocytosis matures and the macrophages carry the necrotic debris out through the cerebral vasculature.  The end result is a cystic fluid filled space surrounded by glial scar.

An important subset of ischemic infarction is that due to arteriosclerosis of the penetrating vessels supplying the basal ganglia, thalamic nuclei, internal capsules, brainstem, and cerebellum. Occlusion of these vessels, which is usually seen in patients with long-standing systemic hypertension, produces a small infarct which liquefies to become a small cavity (<1.5 cm) after the macrophages have removed the dead tissue. Resembling (in someone’s mind) small lakes, these are referred to as lacunar infarcts.

Virtually all lacunar infarcts occur in the arterial territories of the lenticulostriate and thalamoperforant arteries, paramedian branches of the basilar artery, and branches of the anterior choroidal artery.  These arteries have in common both a single unbranching end artery anatomy and a tendency to arise directly from much larger arteries.  Their small size (<500 mm) and proximal origins in the arterial tree are thought to expose them to stresses more severe than found elsewhere in the circulation. Although lacunar infarcts involving the basal ganglia and anterior limb of the internal capsule are asymptomatic, those involving the posterior limb of the internal capsule present as pure motor hemiparesis.

Cerebral edema is defined as an increase in brain volume that is due to increased tissue water content. The normal water content of gray matter is 80% (by weight) while that of the white matter is 68%. In cerebral edema, typical water content values would be 81-82% in the gray matter and 76-79% in the white matter. Therefore, accumulation of edema fluid is greater in white matter compared with gray matter, presumably because water can more easily move through parallel fiber bundles of the white matter. Three major subtypes of cerebral edema are distinguished: vasogenic edema, cytotoxic edema, and interstitial edema.

Vasogenic edema is the result of blood brain barrier dysfunction with increased microvascular permeability, and is typically most pronounced in white matter. Accumulation of water is largely in the extracellular compartment. Vasogenic edema may be seen in association with a wide variety of cerebral processes (e.g. infarcts, abscesses, tumors) and may be responsible for much of the associated morbidity and mortality.

 
Cytotoxic edema occurs consequent to cellular energy failure, the most common cause of which is cerebral ischemia. In cytotoxic edema, energy failure disables the Na-K pump, allowing large amounts of Na to enter the cell. The result is intracellular water accumulation and intracellular swelling. As opposed to vasogenic edema, cytoxic edema is most pronounced in gray matter.
 
Interstitial edema refers to increased water content of periventricular white matter due to obstructive hydrocephalus. In these cases, it appears that fluid under high pressure is forced through the ependyma and into the periventricular white matter.

The adult brain is contained within a rigid compartment and has poor tolerance for any process that increases the intracranial contents, be it a neoplasm, hematoma, abscess, or cerebral edema. Moreover, because the brain parenchyma is poorly compressible, its major physical reaction to a space-occupying mass is to become displaced. When the displacement leads to a shift of brain parenchyma from one anatomic compartment to another, it is referred to as herniation. Herniation phenomena are generally life threatening. The major types of herniation are transtentorial or “uncal” herniation, subfalcine or cingulate herniation, and cerebellar tonsillar herniation.

A space occupying lesion in the temporal lobe or other lateral portions of the cerebral hemisphere is likely to lead to herniation of the medial temporal lobe over the free edge of the tentorium (transtentorial herniation), with multiple potential consequences: 1) alterations in consciousness to the point of coma secondary to midbrain compression by the herniating medial temporal lobe structures; 2) stretching or compression of the oculomotor nerve (ipsilateral, contralateral, or both), with partial ophthalmoplegia and pupillary dilatation; 3) compression of the ispilateral cerebral peduncle with contralateral hemiparesis; 4) compression of the contralateral cerebral peduncle against the contralateral tentorial edge (resulting in Kernohan’s notch) with ipsilateral hemiparesis (sometimes referred to as false localizing sign); 5) compression of the cerebral aqueduct with hydrocephalus, further increasing intracranial pressure; 6) compression of the posterior cerebral artery (ispilateral, contralateral, or both) with resulting ischemic infarction in the posterior cerebral artery territory. Severe shift of the brainstem against a relatively fixed vasculature sometimes results in vascular tears and secondary brainstem hemorrhage, often called Duret hemorrhages, a fatal complication of transtentorial herniation.

Downward movements of the contents of the posterior fossa, as may occur with a posterior fossa mass or diffuse cerebral swelling, results in herniation of one or both cerebellar tonsils through the foramen magnum (tonsillar herniation). Compression of the medulla and its vital respiratory centers may occur and is usually rapidly fatal.

 
Expansion of a mass in the frontal or parietal lobe will eventually result in herniation of the ipsilateral cingulate gyrus under the free edge of the falx (cingulate or subfalcine herniation). This may compromise circulation through the pericallosal arteries and result in infarction of the parasagittal cortex, manifest clinically as weakness and/or sensory loss in one or both legs.

Further Reading

Huang BY, Castillo M. Hypoxic-ischemic brain injury: imaging findings from
birth to adulthood. Radiographics. 2008 Mar-Apr;28(2):417-39; quiz 617.

Liebeskind DS, Kidwell CS. Back to the future: reconsidering the hemodynamics of cerebral ischemia. Neurology. 2009 Mar 31;72(13):1118-9.

Cordonnier C, Leys D. Stroke: the bare essentials. Pract Neurol. 2008 Aug;8(4):263-72.

Marmarou A. A review of progress in understanding the pathophysiology and treatment of brain edema. Neurosurg Focus. 2007 May 15;22(5):E1.

Qureshi AI, Suarez JI. More evidence supporting a "brain code" protocol for
reversal of transtentorial herniation. Neurology. 2008 Mar 25;70(13):990-1.