++
Vertebrate neurons are exquisitely specialized for the functions they perform. As explained in previous chapters, a single neuron may receive information from and relay information to thousands of other neurons; consequently, the nervous system is capable of remarkably complex functions. Moreover, the brisk flux of ions across neural membranes permits extremely rapid interneuronal signaling. However, this specialization comes at a cost. A tremendous amount of energy is required to maintain ionic gradients across the membranes of the approximately 100 billion neurons that comprise the human brain. Although the brain represents only 2% of the body’s total mass, it uses approximately 20% of the body’s oxygen supply, and blood flow to the brain accounts for about 15% of total cardiac output. Ischemia, or insufficient blood supply, results in oxygen and glucose deprivation and in the buildup of potentially toxic metabolites such as lactic acid and CO2. Interruption of blood flow to the brain can lead to complete loss of consciousness within 10 seconds, the approximate amount of time required to consume the oxygen contained in the brain.
++
Stroke occurs on disruption of blood flow to brain tissue caused by obstruction of blood flow or bleeding in the brain (hemorrhage). The exquisite vulnerability of neurons to energy deprivation caused by stroke results in vast medical, economic, and personal costs. In the United States alone, roughly 795,000 strokes occur each year. This equates to an average of one stroke every 40 seconds in the American population. Approximately 150,000 of these strokes are fatal, which equates to one death every 4 minutes, making stroke the fourth leading cause of death in the United States. Survivors of stroke often are beset by serious long-term disabilities, including paralysis and disruption of higher cognitive functions such as speech. Individuals with such disabilities may be unable to resume work and other daily activities, and often require extensive long-term care by healthcare professionals or friends and family.
++
The term stroke, now less commonly referred to as a cerebrovascular accident (CVA), broadly refers to neurologic symptoms and signs that result when blood flow to brain tissue is interrupted. The two primary types of stroke, as noted above, are occlusive and hemorrhagic. An occlusive stroke is caused by the blockage of a blood vessel and accounts for 87% of all strokes in the United States 20–1. Vascular occlusion, which generally restricts blood flow to a discrete area of the brain, results in neurologic deficits and in a loss of functions controlled by the affected region. Occlusive strokes typically are caused by embolic, atherosclerotic, or thrombotic occlusion of cerebral vessels 20–2.
++
++
++
A hemorrhagic stroke is caused by bleeding from a vessel and accounts for 10% of all strokes in the United States. Intracranial bleeding can occur in the intraparenchymal, epidural, subdural, or subarachnoid spaces. Intraparenchymal hemorrhage may be caused by acute elevations in blood pressure or by a variety of disorders that weaken blood vessels. Chronic hypertension is the most common predisposing factor, but coagulation disorders, brain tumors that promote the development of fragile blood vessels, amyloid deposition in blood vessels (ie, amyloid angiopathy; Chapter 18), and the use of cocaine or amphetamines—both of which cause rapid elevation of blood pressure—are among the risk factors for intraparenchymal hemorrhages. Intraparenchymal hemorrhaging can lead to the formation of blood clots (hematomas) in the cerebrum, cerebellum, or brainstem, which in turn may limit the blood supply to nearby brain regions and exacerbate the injurious effects of a stroke. Hemorrhagic stroke can also occur secondary to an initially ischemic stroke.
++
Epidural, subdural, and subarachnoid bleeding often results from head trauma or the rupture of an aneurysm. Subarachnoid hemorrhages account for the remaining 3% of strokes in the United States. In addition to the damage caused by the loss of blood supply to affected areas of the brain, hemorrhages can cause damage by increasing intracranial pressure that further compromises neuronal health. Moreover, through mechanisms that are not completely understood, subarachnoid hemorrhage can cause reactive vasospasm of cerebral surface vessels several days to weeks after the hemorrhage, which in turn can lead to a further reduction in blood supply and additional cerebral infarcts.
+++
Mechanisms of Neuronal Injury During Stroke
++
When neurons are deprived of the nourishment they require, they quickly become unable to maintain their resting membrane potentials and, as they depolarize, they fire action potentials. Their firing triggers the release of neurotransmitters, in particular, glutamate, which in turn promotes depolarization of neighboring neurons. Such activity sets the stage for a destructive cycle of neuronal activation, neurotransmitter release, and further activation. Prolonged periods of neuronal activation can lead to the disruption of ionic gradients, massive Ca2+ influx, cellular swelling, activation of cellular proteases and lipases, mitochondrial damage, generation of free radicals, and eventually widespread neuronal death. Ischemia of only a few minutes’ duration can result in permanent brain damage. The basic biochemical mechanisms responsible for these processes of neuronal injury and death, broadly termed either necrosis or apoptosis (programmed cell death), are described in detail in Chapter 18.
++
Despite our growing knowledge of the mechanisms that underlie ischemic neuronal death, our ability to treat stroke remains limited. Among the treatment strategies currently available, the best are geared toward prevention through the maintenance of cardiovascular health, restoration of blood supply, and the slowing of metabolism with hypothermia. Although none of these therapies has capitalized on the sophisticated studies of the biochemical events underlying neuronal cell death, efforts to prevent stroke nevertheless have been successful. The incidence of stroke has been reduced markedly by primary preventive measures aimed at controlling hypertension, hypercholesterolemia, diabetes, and tobacco use. HMG-CoA reductase inhibitors, known as the statin class of cholesterol-lowering agents, appear to confer additional benefit in stroke prevention beyond cholesterol control, as treatment with statins in individuals with normal cholesterol levels significantly reduces the incidence of future stroke. Prophylactic use of drugs that inhibit platelet function, such as aspirin, clopidogrel, or dipyridamole, has proven to be effective in reducing the risk of occlusive stroke. While these drugs can result in a slight increase in the risk of hemorrhagic stroke, this small risk is outweighed markedly by the benefit conferred for ischemic stroke prevention.
++
One reason that stroke prevention is so important is that current stroke therapies are pitted against an unforgiving opponent: time. “Time is brain” is a fundamental concept to the treatment of acute ischemic stroke. As previously emphasized, serious neuronal damage can occur within minutes of an ischemic event. Barring round-the-clock observation of all individuals at risk for stroke, the effectiveness of treatment in humans is unlikely to approach that of laboratory animals because researchers in the laboratory have the luxury of administering therapy during or immediately after an ischemic insult. This underscores the need for protective therapy in high-risk populations.
+++
The Peri-Infarct Area: An Important Treatment Target
++
Many of the approaches to treatment discussed in the sections that follow involve actions that occur primarily in the peri-infarct area, the “penumbra,” and serve to salvage at-risk neurons that would otherwise be destined to die within hours or days after a stroke. The peri-infarct area constitutes compromised but potentially salvageable tissue between the severely ischemic core and adequately perfused brain tissue. Although potentially salvageable, the peri-infarct area is quite vulnerable because it is subject to high levels of excitotoxic neurotransmitters and free radicals, waves of cellular depolarization, and inflammatory processes (Chapter 18).
+++
Restoring Blood Supply
++
After an occlusive stroke, blood supply can be restored with thrombolytic agents, which dissolve the clots that impede the normal flow of blood. Thrombolytic agents have been found to improve the outcome of this type of stroke in clinical trials. However, strict protocols are utilized in determining patient eligible for this form of therapy in order to minimize the risk of hemorrhage. Consequently, the presence of hemorrhagic stroke must be ruled out by use of computerized tomography (CT), and other risk factors such as malignant hypertension, recent surgery, or prior cerebral hemorrhage must be excluded before thrombolytic agents are used. Even occlusive strokes are accompanied by a small but real risk of hemorrhagic transformation.
++
Thrombolytics such as tissue plasminogen activator (tPA; eg, alteplase, reteplase), urokinase, streptokinase, prourokinase, and desmoteplase are proteins that promote the conversion of the proenzyme plasminogen into plasmin, an enzyme that degrades fibrin, a key structural protein in most blood clots 20–3; 20–1. Currently, tPA is the only thrombolytic substance approved for intravenous use in acute ischemic stroke in the United States. Clinical trials demonstrate that intravenously delivered tPA reduces the disability of patients with acute ischemic stroke who were treated within the first 4.5 hours of the onset of symptoms.
++
++
20–1 Role of Nitric Oxide in Stroke
Nitric oxide (NO) functions as an intracellular and intercellular messenger in the brain (Chapter 8). It is synthesized by the Ca2+-activated enzyme, nitric oxide synthase (NOS). The importance of NOS activation in neuronal injury has been tested through the use of specific NOS inhibitors, such as L-nitroarginine. Initial studies produced inconsistent results: NOS inhibitors displayed neuroprotective effects in some experiments, especially those conducted in cell culture, and produced either no effect or a detrimental effect in others. These conflicting results most likely were attributable to the multiple roles and sources of NO in the brain.
Three different isoforms of NOS exist, each of which is the product of a distinct gene. Neuronal NOS (nNOS) is expressed exclusively in neurons, endothelial NOS (eNOS) originally was identified in endothelial cells, and inducible NOS (iNOS) originally was identified in certain immune system cells. Some neurons express eNOS and iNOS in addition to nNOS. NO produced by eNOS acts as an endothelial relaxing factor: it promotes the relaxation of the smooth muscle surrounding arterioles and leads to vasodilation and increased blood flow.
Gene knockout technology has helped to elucidate the roles of these NOS isoforms in neuronal injury. Compared with their wild-type counterparts, mice deficient in nNOS typically have smaller infarct volume (ie, amount of necrotic tissue) after an experimentally induced ischemic stroke. This finding suggests that nNOS activity may be detrimental to neuronal survival during ischemia. In contrast, mice deficient in eNOS tend to have greater than normal infarct volumes after experimentally induced stroke, which indicates that eNOS has neuroprotective activity. Most likely, eNOS exerts its beneficial effect by promoting the reperfusion of the ischemic area. This is demonstrated in the figure. Interestingly, iNOS knockout mice, like nNOS knockout mice, display diminished infarct volume after ischemic stroke; it is speculated that a decreased inflammatory response may reduce infarct size in these mice.
The multiple effects of NO on ischemic injury provide an excellent lesson in the complexity of the brain’s response to ischemia. Other events that take place during ischemia, such as Ca2+ entry into cells, also may have multiple and varied effects, and these actions must be carefully examined if effective therapies are to be devised. If, for example, NOS inhibition can be developed as a clinical treatment for ischemic neuronal injury, such inhibition most likely will have to be carefully targeted to nNOS or perhaps iNOS.
++
Prior to large initiatives to expedite the treatment of acute ischemic stroke, similar to efforts for acute myocardial infarction, by the time an individual was aware of the occurrence of a stroke, traveled to a hospital, and was diagnosed, hours had elapsed. In many centers around the world, the average onset to treatment time has dropped to under 60 minutes. However, despite these marked improvements in select centers, and the fact that administration of intravenous tPA within 4.5 hours of symptom onset is now standard of care in the United States, only 3% to 5% of stroke patients actually receive tPA because of the difficulty of administering it within this time frame.
++
Even when intravenous tPA is successfully administered in time, the affected vessel does not always open or open completely. At academic centers, intravenous therapy is sometimes followed by interventional procedures, where a catheter is guided from a peripheral artery (usually the femoral artery in the groin) to the affected cerebral artery. The clot is then treated with a combination of mechanical clot disruption or direct instillation of a thrombolytic agent into the clot itself. Because a smaller dose of the thrombolytic agent is used when given at the site of the clot itself, as compared with systemic (intravenous) administration, intra-arterial therapy is in theory safer and can be performed up to 8 hours after symptom onset. However, three large trials published in 2013 argued against widespread use of this approach due to negative results. Thus, current recommendations are for intra-arterial therapy to be used on a selective, case-by-case basis. In cases of basilar artery thrombosis, when the brainstem is at risk and the matter of opening the artery is literally life and death, intra-arterial therapy administered up to 48 hours after symptom onset may improve outcomes. Prourokinase, urokinase, and desmoteplase are additional thrombolytics that are used less commonly in intra-arterial therapy.
++
Heparin is a heterogeneous mixture of sulfated mucopolysaccharides. It is found in mast cells and in the extracellular matrix of most tissues. It has a molecular mass of 750 to 1000 kDa and is composed of long polymers of glycosaminoglycan chains that are attached to a core protein 20–4. Because of its structure, heparin is not effective after oral administration and must be given parenterally. It inhibits clot formation by enhancing the activity of antithrombin III, a protein that forms equimolar complexes with the various proteases activated during the clot formation process (see 20–3). By binding directly to antithrombin III, heparin causes a conformational change in the protein that enhances its binding to the clotting factor proteases. Heparin has been shown to cause more harm than benefit as an acute treatment of stroke and carries a risk for increased hemorrhage.
++
++
As discussed in Chapters 4 and 11, aspirin inhibits cyclooxygenase, which in platelets catalyzes the conversion of arachidonic acid to thromboxane A2, among other products 20–4. Thromboxane A2 is a critical intermediate in the recruitment of platelets necessary for the clotting cascade. Aspirin administered to patients during hospital admission for stroke produces a small but significant net benefit in that it reduces mortality by 14% compared with placebo. Clopidogrel and related thienopyridine class agents (eg, prasugrel and ticlopidine) are other antiplatelet agents in current use. They act as adenosine diphosphate (ADP) antagonists, whereby they inhibit the binding of ADP to P2Y12 receptors (Chapter 8) on platelet membranes. They also irreversibly modify platelet P2Y12 receptors, and therefore their effects last for the lifespan of the platelets (approximately 7–10 days). The P2Y12 receptor is responsible for activation of the glycoprotein GPIIb/IIIa complex (also known as integrin αIIbβ3), the major receptor for fibrinogen.
++
Dipyridamole, used for similar purposes clinically, inhibits clot formation and causes vasodilation, although it is not known which of its many actions (eg, phosphodiesterase, adenosine reuptake, or adenosine deaminase inhibition) is responsible for the drug’s clinical effects. A combination drug of aspirin–dipyridamole has been shown to offer added benefit in stroke prevention, relative to either drug given alone. Recent evidence suggests that dual antiplatelet therapy with aspirin and clopidogrel may also have added benefit in select populations. However, combination therapy historically has been avoided due to the increased risk of hemorrhage reported in the MATCH clinical trial.
++
Patients at risk for cardioembolic strokes, such as those with atrial fibrillation or a mechanical heart valve, conditions that predispose patients to form intracardiac clots, often are treated with warfarin or one of the more recently developed oral anticoagulant alternatives (eg, dabigatran, rivaroxaban, and apixaban) 20–4. Warfarin is a synthetic derivative of a related compound in sweet clover, which was found in the early 20th century to promote bleeding. It acts as a functional vitamin K antagonist; it does not directly antagonize the function of vitamin K; rather, it depletes vitamin K by inhibiting its recycling. Vitamin K is a required cofactor for the enzymes that activate several clotting factors, including II, VII, IX, and X (see 20–3). Warfarin and related compounds are the most potent oral anticoagulants known; indeed, they are so potent that severe hemorrhage is a significant side effect of their use. (This effect, at high doses, is exploited in warfarin’s use as a rat poison.) Patients who take warfarin must have regular blood tests to ensure that their bleeding times are within safe boundaries and that dose adjustments are made accordingly. Unlike warfarin, the newer, oral anticoagulants do not require monitoring, allowing for standardized dosing; importantly, they appear equally efficacious. These drugs each target a specific factor in clotting cascades: dabigatran is a direct inhibitor of thrombin, while rivaroxaban and apixaban are inhibitors of factor Xa (see 20–3).
++
Aspirin, warfarin, and other oral anticoagulants are used not only to treat stroke but also in the treatment of transient ischemic attacks (TIAs), which are brief periods of brain ischemia that resolve without a lasting neurologic deficit (ie, without appreciable neuronal death). These attacks are believed to be caused by transient occlusions of the cerebral vasculature. The symptoms of TIAs are similar to those of stroke, except that they resolve within minutes, to less commonly hours, of onset.
+++
Minimizing Ca2+ Influx Into Cells
++
Because Ca2+ appears to be critically involved in promoting the biochemical processes that lead to neuronal destruction, the reduction of Ca2+ influx might be considered a promising strategy in the treatment of stroke. However, the effectiveness of drugs that reduce the influx of Ca2+ into neurons (see 20–1) has yet to be demonstrated in clinical trials. Inhibitors of voltage-dependent Ca2+ channels (Chapter 2) such as nimodipine, an L-type Ca2+ channel blocker that penetrates the brain, and flunarizine, a T-type Ca2+ channel blocker, have been investigated as potential therapies for stroke but thus far have not been shown to improve the functional outcome of patients after ischemic stroke. NMDA receptor antagonists exhibit a robust protective effect on neurons in culture and in vivo in animal models but have not proven to be effective in humans. Even if they were effective, many of these antagonists have phencyclidine-like adverse effects, such as psychosis and dissociation (Chapter 17), which severely limit the dose that can be used. Magnesium blocks the NMDA receptor, and a current phase 3 trial of intravenous magnesium sulfate given within 2 hours of symptom onset is under way. Because of the relative safety of magnesium and the narrow time window defined by the study, the trial design allows intravenous magnesium to be given by paramedics in the field, prior to arrival and evaluation in the emergency room. Antagonists of voltage-dependent Na+ channels, such as phenytoin, which can be very effective in the treatment of seizure disorders (Chapter 19), have failed to improve clinical outcomes of stroke. Likewise, drugs that promote GABAergic function in brain have been considered for stroke, but efficacy of this mechanism too has not yet been demonstrated in humans.
++
+++
Reducing Free Radical Damage and Cell Death Pathways
++
There is evidence that the increased generation of free radicals in ischemic brain tissue may contribute to neuronal injury and death 20–5; 20–1 (Chapter 18). Free radical scavengers are agents that are oxidized by oxygen-reactive species without deleterious effects to the cell and thus might be expected to have positive effects in stroke. Tirilazad, a nonglucocorticoid steroid that inhibits lipid peroxidation, has been shown to reduce infarct area in animals treated within 10 minutes of complete focal ischemia. However, tirilazad had no effect on functional outcome when administered to humans approximately 4 hours after stroke. Ebselen, another free radical scavenger, is now in phase 3 clinical trials. Disofenin, a free radical trapping agent, showed modest initial promise in one trial in terms of functional outcome at 90 days, but those results were not replicated in a subsequent trial. Thus, currently there are no accepted stroke therapies based on blocking free radical damage.
++
++
Knowledge of the biochemical basis of necrotic and apoptotic mechanisms of neuronal death (Chapter 18) has suggested many additional potential approaches to the treatment of stroke. Indeed, genetic manipulation of numerous cell death or survival proteins in rodents has been shown to alter the brain’s vulnerability to a stroke. However, none of these strategies has to date been validated in clinical trials in humans. Among such strategies that were previously investigated in animal models is the use of caspase inhibitors; caspases (short for cysteine aspartate proteases) are enzymes that promote apoptosis and necrosis (Chapter 18).
++
Another potential strategy is the use of agonists of peroxisome proliferator–activated receptor-γ (PPARγ). PPARγ is a member of the nuclear receptor superfamily, which also includes the receptors for steroid hormones, vitamin D, and retinoic acid (Chapter 4). PPARγ, and its PPARα and PPARδ isoforms, heterodimerizes with the retinoid X receptor to form an active transcription factor complex that regulates many genes involved in intermediary metabolism. While the endogenous ligands for PPARγ remain uncertain (prostaglandins may be involved), synthetic agonists exert beneficial clinical effects: the thiazolidinediones, pioglitazone and rosiglitazone, are effective antidiabetic agents and act by increasing the sensitivity of peripheral tissues to insulin. Early studies demonstrated that diabetics treated with these agents exhibited a lower incidence of stroke; however, subsequent studies have suggested a higher risk of cardiovascular events (myocardial infarction and stroke) in diabetic patients exposed to rosiglitazone.
++
Still another experimental approach involves inhibitors of nitric oxide synthesis, based on the animal literature that the generation of nitric oxide during ischemia, mediated by excessive glutamatergic transmission and intracellular Ca2+ levels, may contribute to neuronal injury perhaps via free radical formation. However, other studies of nitric oxide in stroke suggest that it might be protective, highlighting some of the complexities of translating findings from animal models to the clinical situation 20–1. Indeed, over 1000 neuroprotective agents have been tested in preclinical studies, with many showing promising results. In contrast, of the 200 ongoing or completed clinical trials, no agent has yet been successful in being translated to clinical practice.
+++
Strategies With Pleiotropic Agents
++
Several neuroprotective strategies act at multiple levels in the cascade of postischemic damage. Statins, mentioned earlier as agents that play a key role in preventing strokes from occurring, appear to also confer benefit on stroke recovery in several animal and human studies. Effects of statins include improving endothelial function, reducing inflammation, and increasing cerebral blood flow. In animal models, statins reduce infarct size and improve functional recovery. These benefits of statins are bearing out in clinical studies as well. It will be interesting in future research to understand the mechanisms underlying these diverse, beneficial effects of statins.
++
Another broad-spectrum strategy is hypothermia. Controlled hypothermia confers benefit in cases of global ischemia after cardiac arrest, and so interest has been sparked as to its effect in cases of focal ischemia or stroke. In the setting of neonatal hypoxic-ischemic injury, hypothermia has been shown to offer clinical benefit. Several studies have demonstrated neuroprotective effects of hypothermia in animal models of adult stroke, and hypothermia has proven benefits in other mechanisms of human brain injury, including cardiac arrest. The clinical utility in human stroke is a matter still under investigation, with mixed results in initial clinical trials.
+++
Promoting Neural Recovery
++
Another approach to stroke therapy is to promote the self-repair of damaged neurons or the growth of healthy neurons to help compensate for the loss of neurons destroyed during an ischemic attack. Two general strategies can be used to this end. A nutritive strategy involves ensuring that neurons have the molecules they need for repair and growth, and a signaling strategy involves providing the chemical signals that instruct neurons to grow. The first strategy has been attempted with administration of the phospholipid precursor citicoline. Citicoline is a key intermediate in the biosynthesis of phosphatidylcholine, an important component of the neural cell membrane. While past studies had suggested some positive effects, a recent large trial demonstrated no clinical benefit relative to placebo.
++
Neurotrophic factors such as basic fibroblast growth factor (bFGF; Chapter 8) also are being considered for restorative therapy. In animal stroke models, bFGF reduces infarct volume when given shortly after the onset of focal ischemia. Although bFGF did not reduce infarct size when given 24 hours after experimentally induced stroke, it did improve outcome as measured by behavioral tests. However, this neurotrophic factor and others listed in 20–1 have either shown no benefit or are yet to be tested in human clinical trials.
++
Despite all of the advances in our understanding of neural injury related to stroke, the best-established way to ensure long-term recovery of function is through rehabilitation. Research involving laboratory animals has provided insight into how rehabilitation might work at the neurobiologic level 20–2. The results of such research raise the possibility that neurobiologic mechanisms underlying the inherent plasticity of the brain may be exploited in the future to ensure maximal return of function after stroke.
++
20–2 Neurobiologic Basis of Rehabilitation
Stroke patients with large initial deficits often can exhibit striking improvement. The length of the recovery process (typically 1–2 years) suggests that events other than the resolution of edema and inflammation are responsible for improvement in function. In some cases, restoration of blood flow through the development of collateral circulation may contribute to the regaining of function. However, several lines of evidence indicate that neurons undergo anatomic and functional changes that significantly assist in functional improvement after a stroke. In rats, for example, increased expression of the growth cone–associated protein GAP-43 and of the synaptic vesicle protein synaptophysin have been detected near experimentally induced infarct areas. Growth cones are specialized endings of growing axons before they form mature synapses. Increased expression of GAP-43 also has been noted in the periphery of infarcted human brain tissue examined at autopsy. Interestingly, dendritic sprouting has been observed contralateral to cortical lesions produced by electrocauterization in rats. This finding suggests that recovery of function may occur as the corresponding, contralateral area of the brain assumes the function of its injured counterpart.
The occurrence of compensatory neural remodeling after stroke has been demonstrated experimentally. Electrophysiologic experiments have revealed a reassignment of function after ischemic infarct in squirrel monkeys: when small infarcts are produced in the area of the motor cortex that corresponds to the hand, new areas of the cortex, previously responsible for movements of the arm and shoulder, slowly gain the ability to control hand motions. This topographic reorganization of neuronal function required that the monkeys perform tasks that necessitated the use of their debilitated hand. Thus, frequent activation may stimulate the growth of remaining neuronal processes responsible for control of the hand into arm-and-shoulder territory. Alternatively, such activation may increase the potency of a small number of “hand neurons” that preexisted in the arm and shoulder space.
Similar reassignment of function likely takes place in the human brain. Positron emission tomography (PET) and magnetic resonance imaging (MRI) studies indicate that adjacent or contralateral brain regions may indeed work to compensate for damaged tissue. Verbal tasks typically cause activation of speech areas in the left hemispheres of normal subjects; however, in some recovered aphasic patients, increased activation of homologous areas in the right hemisphere has been observed. Is this an indication that the right hemisphere can undergo changes that allow compensation for the damaged left hemisphere? Unfortunately, current experiments have not conclusively answered this question. It is possible, for example, that speech centers in the right hemispheres of certain aphasics participated in verbal tasks to an unusual degree before the onset of stroke. However, in light of anatomic and physiologic data from animal models, dendritic and axonal growth and other forms of neural plasticity are mechanisms worth investigating in stroke patients. Transcranial magnetic stimulation and transcranial direct current stimulation are noninvasive brain stimulation techniques currently being investigated for their utility as adjuvants to intensive rehabilitation poststroke. Knowledge of the molecular and cellular events that influence such rearrangement may eventually lead to other techniques for aiding the recovery of stroke victims. Moreover, because shifts of function depend on the use of an affected area after stroke, it is likely that aggressive physical or speech therapy will continue to be a critical tool in promoting recovery after stroke.