The brain, the most complex of all the organs, is far from homogeneous in structure, and coordinated perfusion of its disparate regions appears to be under precise regulatory control, the nature of which has only recently begun to be appreciated. The surface of the brain is covered by pia mater incorporating a very rich and dense network of blood vessels. The cerebral arteries are thus distributed over the brain surface and they dive into the brain parenchyma to supply non-superficial structures. These vessels appear to contract and relax in response to metabolic activity taking place in the deeper recesses of the brain. The question is, how is this regulated? Numerous substances (adenosine, lactate, CO2, H+, nitric oxide etc.) appear to be involved in coupling metabolic demand to regional blood flow in specific tissues, but even when they are identified in the brain there remain the numerous unsolved questions surrounding how such mediators might be transported to the flow-controlling arterioles so rapidly and so precisely - it is not intuitively obvious how metabolic products swept downstream from a site of activity could directly influence the diameter of upstream cerebral vessels supplying very discrete regions of the brain.

As long ago as 1890 it had been hypothesized that, if function is coupled to metabolism and metabolism is coupled to blood flow, then regional heterogeneity in function must be accompanied by local changes in blood flow.1 Extraordinary progress has been made over the last decade in analysis and visualization of cerebral activity; and the assumption of a tight correlation between behavior, cerebral metabolic activity, and coupled changes in regional blood distribution underlies the logic of these current brain-imaging experiments. We have now become used to seeing representations of functional activity of the brain in which color-coding describes relative blood flow in various regions of the brain; these images have become progressively more complex and more highly resolved.

The purpose of this lecture is to give an oversight of the methods used to generate these images, to describe a little of what is known about the control of regional blood flow to the structures of the brain, and to indicate ways in which increased understanding of flow-metabolism coupling might lead to future clinical interventions.

Cerebral Blood Flow
Textbook descriptions of cerebral blood flow generally stress the abundance of flow, relative to other tissues, the constancy of blood delivery in the face of variations in systemic blood pressure, and the exquisite, precise coupling between blood flow and metabolic activity within specific regions of the brain. In gross terms, the brain is said to receive approximately 15% of total cardiac output; carrying with it metabolic substrates (principally glucose and oxygen) and sweeping away metabolic products (CO2). If blood flow is disrupted cognitive function and consciousness are lost within seconds. If disruption is maintained for more than 3-4 minutes irreparable, catastrophic, damage is done which cannot, at the present time, be reversed by any subsequent intervention. Although adequate blood perfusion of the brain is essential for function, blood perfusion alone is not sufficient; the essential component being delivered and consumed is oxygen (and, to a slightly less critical extent, glucose.) So it is that, despite adequate blood flow to the brain consciousness is lost rapidly when oxygen or glucose supply is restricted. Severe acute hypoxia occurs when aircraft at altitude suddenly depressurize; studies have shown that the time of useful consciousness following decompression without an oxygen mask is 10 seconds at 50,000 ft, this time being determined by the transit period for deoxygenated blood from the lung to the brain.2 It is more difficult to effect such acute decreases in glucose concentration in human subjects, but the intravenous administration of high doses of insulin certainly induces coma very rapidly.

Cerebral Blood Circulation is Complex
The arteries directing blood to the brain are connected in an interesting way. They terminate, on the floor of the cranial cavity, in the Circle of Willis, a vascular structure that encircles the brainstem, throwing off the major vessels supplying the cerebral hemispheres (the anterior, middle and posterior cerebral arteries). Because the Circle permits collateral flow (via the anterior and posterior communicating arteries), obstruction of any one of the four arteries supplying the brain (internal carotid and vertebral arteries, right and left), or even damage to the Circle itself may often be clinically insignificant because blood can readily be redirected. The cerebral vessels, themselves, on the other hand, have no anatomic reserve and any obstruction occurring within them will have catastrophic consequences. There is an amazingly intricate capillary bed serving the neuronal and supportive tissue of the cerebral hemispheres downstream from these major supply vessels. The flow of blood through this web of capillaries is regulated by the muscular tone of resistance elements (classic arterioles, and also contractile pericytes encircling passive capillaries) which have the capacity to alter the resistance in -and therefore flow through - specific regions of the network. There are relatively global modulators of cerebral vascular tone, such as CO2 and O2, and there are agents such as adenosine and nitric oxide that cause widespread increases in cerebral blood flow if injected into major supply vessels; but the more subtle question is how very specific local changes in vessel tone are regulated. How do local arterioles sense the blood supply requirements of local downstream regions of the brain? The first step in investigating this is to visualize the scale (temporal and spatial) of these blood flow changes.

Regional Variations in Cerebral Blood Flow can be Visualized
There are methods for studying blood flow and methods for studying local metabolism; sometimes the same physical technique can be used to yield information on both modalities. To the non-specialist some of these techniques can be cryptic and it is worth describing some of the more common approaches in modest detail. First, the visualization of flow; there are both non-invasive and invasive techniques, but only techniques yielding rather generalized impressions of cerebral blood flow are currently applicable to the OR setting.

Fick principle: The size of a stream may be calculated by measuring the concentration difference, at various points or various times, of an indicator substance which is being added or removed from the stream itself. The direct Fick method uses endogenous oxygen concentrations as indicator, but more commonly dye dilution, thermal dilution or isotope dilution are used. Kety & Schmidt based their pioneering work on the Fick principle and used nitrous oxide as a reasonable approximation of an inert, freely diffusible tracer that would identify global blood flow to the brain.3 The method is quite cumbersome and requires cannulation of the jugular venous bulb; however it represents a standard of assay against which all other methods of global cerebral blood flow determination have been evaluated. If certain assumptions are made about the constancy of cerebral metabolic rate, then the arteriovenous oxygen content difference at the jugular venous bulb (AVDO2) can be used to monitor changes in cerebral blood flow. Provided that temperature and hemoglobin concentration also remain constant then a change in AVDO2 results in a reciprocal change in jugular venous oxygen saturation related to changes in cerebral blood flow.4

Transcranial Doppler insonation (TCD): A safe, reasonably reliable and inexpensive way of measuring cerebrovascular blood velocities. Pulses of ultrasound (2 MHz) from a hand-held transducer are pointed towards the major vessels in the base of the skull. The frequency shift in the reflected sound indicates the velocity of the reflecting surfaces (erythrocytes). Remember that flow velocity is inversely proportional to the fourth power of the vessel radius, and high velocities are therefore primarily a consequence of vessel constriction. Images can be reconstructed from the time-dependent intensity of the reflected sound and the cerebral arteries, internal carotids, basilar and vertebral arteries can all be insonated by altering transducer location angle and depth setting. Only large vessels can be observed (through relatively thin bone "windows"), most commonly in the orbit and in the temporal and suboccipital regions. Although flow velocity is not equivalent to blood flow, acute changes in velocity in the major cerebral supply vessels usually reflect corresponding relative changes in blood flow (and hence changes in the tone of downstream resistance elements) and the technique has been well validated in a large number of experimental and clinical settings. More recently, the introduction of laser Doppler flowmetry (LDF) techniques have had a dramatic impact in several areas of neurobiological research in living animals. LDF permits the continuous assay of blood flow in small vessels of brain parenchyma with a time resolution of several milliseconds; thus facilitating correlation between neuronal synaptic activity with local cerebral blood flow in real time. Similarly, laser-Doppler perfusion imaging (LDPI) now yields images of blood flow within the brain cortex with a time resolution of 2 seconds and a spatial resolution of 10 mm under optimal experimental conditions.

Magnetic Resonance Imaging (MRI): The key is that different protons resonate at slightly different frequencies (local electron density shields the magnetic field as seen by a given proton). As a result different groups of protons, as defined by their local electronic density, can be distinguished. A wave packet of radio-frequency (RF) radiation spanning some frequency interval is aimed at the patient; in its wake, the RF pulse leaves protons whose orientation has been disturbed. Because the pulse was composed of many frequencies, a wide variety of protons can be affected and, as they decay back to their ground states, characteristic "echoes" are generated, detected and analyzed. The workhorse method of MRI is spin-echo imaging: protons are exposed to RF twice, with a pause between the two exposures which allows exploitation of different longitudinal (T1) and transverse (T2) decay times. Contrast between tissues is obtained by changing the length of the pause in the spin-echo method. In this way, an image can be either "T1- or T2-weighted". Neuronal activity causes local changes in cerebral blood flow, blood volume, and blood oxygenation; high-speed spin-echo planar imaging techniques have been used to obtain completely noninvasive tomographic maps of human brain activity stimulated by visual and motor stimuli. Changes in blood oxygenation are detected by using a gradient echo imaging sequence (GE) sensitive to the paramagnetic state of deoxygenated hemoglobin. Blood flow changes are evaluated by a spin-echo inversion recovery sequence (IR). A series of images can be acquired continuously with the same imaging pulse sequence (either GE or IR) during task activation.

MRI Angiography: Flow angiography is the imaging of flowing blood in the arteries and veins of the body. In the past, angiography was only performed by introducing a x-ray opaque dye into the human body and making an X-ray image of the dye. This procedure produced a picture of the blood vessels in the body. It however did not produce an image which distinguished between static and flowing blood. It was therefore a less than adequate technique for imaging circulatory problems. Magnetic resonance angiography (MRA) on the other hand produces images of flowing blood. The intensity in these images is proportional to the velocity of the flow.

There are three general types of MRA angiography: time-of-flight, phase contrast and contrast-enhanced: Time-of-flight angiography can be performed in several ways. One method uses a spin-echo sequence where the two stimulating RF pulses have different frequencies. The first pulse excites protons spinning in one plane; the second pulse excites protons spinning in another plane. In the absence of flow, no signal is seen because no protons experience both the first and second pulses. In the presence of flow and the correct delay time, blood from the first plane flows into the second plane and produces an echo. Phase contrast angiography is a little more complicated. It uses a bipolar gradient pulse (one in which the gradient is turned on in one direction for a period of time then turned on in the opposite direction for an equivalent amount of time). A bipolar gradient pulse has no net effect on stationary protons, but protons having a velocity component in the direction of the gradient will be affected. The net effect is an image of the flowing protons. Contrast enhanced angiography is based on the difference in the T1 relaxation time of blood and the surrounding tissue when an intravascular paramagnetic contrast agent is present. This agent reduces the T1 relaxation times of the fluid in the blood vessels relative to surrounding tissues. When the data is collected the signal from the tissues surrounding the blood vessels is very small and the signal from the blood vessels dominates. The high quality of images from contrast enhanced MR angiography has made MRI the modality of choice for angiography.

There are also a number of techniques for observing blood flow that involve administration of (radioactive) tracers: For example 133Xenon can be mixed with air (5 mCi/L) and administered by 1 minute inhalation. For the next 10 minutes tracer concentrations in various regions of the brain are monitored with an array of scintillation detectors; then regional flow values are calculated and graphically displayed as topographic brain maps in up to 256 bihemispheric cortical regions. Similarly 14C antipyrene and 14C 2-deoxyglucose can be infused into the bloodstream of experimental animals which are subsequently decapitated; the brains are frozen, sectioned and exposed to radiographic film or plates. The subsequent autoradiogram yields quantitative information about the amount of antipyrene (marker for blood distribution) and deoxyglucose (marker for glucose utilization) in specified regions of the brain. Finally acrylic microspheres of defined size can be injected into the blood stream; these plastic particles are large enough to become entrapped in capillary beds and the assumption is that the number of spheres becoming trapped in any given tissue region is proportional to the blood flow to that region; quantitation is either by counting, radioassay or fluorimetry, depending on the nature of the microspheres used.

Does tissue oxygen distribution in fact parallel blood flow?
Visualizing blood flow does not automatically imply equivalent visualization of oxygen delivery. The physiology of oxygen distribution between the cerebral capillaries and the brain tissues is not well understood; in particular, the interface conditions between the different compartments of oxygen distribution in the brain and the capillaries themselves are not yet described in any detail.5 Confocal microscopy of rat brain cortex shows that 10-20% of cerebral capillaries may not contain a single erythrocyte at any given time.6 The subsequent assumption that non-hemoglobin oxygen transport may be clinically significant has led to investigation of the effect of hyperbaric oxygen therapy, in particular on patients with significant head injuries. Nevertheless, at the scale of resolution of most common imaging techniques, it is reasonable to equate demand for oxygen with demand for blood. It is not clear, yet, if demand for glucose also equates to demand for blood; i.e. if increased local glucose requirement - in the presence of abundant oxygen supply - triggers a commensurate increase in local blood flow.

Cerebral Metabolism
Although the brain performs no mechanical work its energy requirements are relatively very large. Energy is consumed by maintenance of membrane integrity, and in support of transmembrane ion gradients essential for electrical activity and cell survival. In addition, energy is required for the synthesis, sequestration and ordered release of a constellation of neurotransmitters. Energy costs of ion transport, and the maintenance of nonequilibrium ion gradients across cell membranes accounts for about 70% of the energy demands of the human brain.7 The preferred energy source for brain tissue is glucose; using glycolysis and oxidative phosphorylation 38 ATP molecules are yielded from each glucose molecule consumed. Greater than 95% of the glucose used by the brain undergoes this oxidative metabolism, and 43% of the energy released is captured in the form of ATP, the rest being lost as heat.8 Oxygen delivery to the brain is determined by the product of cerebral blood flow and arterial oxygen content

The Working Brain Uses More Energy than the Resting Brain:
In common with most other organs of the body, the brain has phases of greater and lesser metabolic activity. For the human body as a whole, the minimum amount of oxygen consumed (at complete rest) is approximately 4 mL/min/kg; any amount of physical activity (particularly shivering) causes an incremental increase in oxygen consumption. The brain, undertaking normal intellectual functioning, uses oxygen at a rate of approximately 35 mL/min/kg brain tissue; so, for a 70 kg man with 1.5 kg brain, basal whole body oxygen consumption is 280 mL/min with brain oxygen consumption of 50 mL/min. Adequate intellectual function thus represents approximately 20% of total basal metabolic demand of the body at rest.9 There is still very little knowledge about the precise type of nerve cell activity that generates increases of blood flow or oxygen and glucose consumption - one of the historical difficulties in the mechanistic interpretation of functional brain imaging experiments. Internal jugular venous sampling reveals that the blood leaving the brain typically has an hemoglobin oxygen saturation of 60-75%4, this does not change significantly during even strenuous intellectual activity; saturation decreases are, however, noted during major cerebral metabolic storms such as focal seizure activity. Interestingly, it has been noted that the onset of electrical activity, detected by EEG, following an interval of electrical silence, is accompanied by an increase in VMCA, blood flow velocity in the middle cerebral artery.10

At sea-level maximum possible oxygen uptake is approximately 3.5 L/min (for a 70 kg man), this is 70-fold greater than the basal cerebral oxygen demand and brain oxygen requirements are readily met almost regardless of the level of co-incident physical exertion; the only likely causes of cerebral hypoxic insult are traumatic (hypoperfusion states, anemia, asphyxiation etc.). More extreme environments may not provide the same margin of safety: on Everest maximum possible oxygen uptake at the summit is less than 1 L/min 9 and, although cerebral basal demand is still only 5% of this amount, the requirements of physical activity (climbing) and, more importantly, maintenance of core temperature can readily lead to progressive degradation of intellectual function and even cerebral integrity.11

Oxygen Supply to the Brain is Obligatory:
The brain has very limited capacity for anaerobic metabolism. It has been proposed that cerebral oxygen requirements can be divided conceptually into two broad categories: a proportion (60%) of the brain energy consumption is devoted to functional activity, i.e. the generation of electrical and chemical signals; the remaining 40% being dedicated to maintaining the integrity of brain structure.12 When oxygen supplies to the brain are limited for some reason, the brain can retrench, restricting energy consumption only to those processes vital to tissue integrity and sacrificing intellectual activity until energy flow is restored. If oxygen supplies are insufficient to maintain the integrity of neurons, damage accumulates rapidly.

Regional Variations in Cerebral Metabolism can be Visualized
In the normal, non-traumatized, brain, tissue metabolic demand varies with functional activity, and can now be visualized in vivo using optical signals from, for example, deoxy- and oxy-hemoglobin in the cerebral cortices. Biochemical modalities differ from structural ones in that they follow actual chemical substituents and trace their routes through the body. Since, usually, anatomical structures serve different functions and embody different biochemical processes, to some degree biochemical imaging can give anatomical information. However, the strength of these methods is to distinguish tissue according to metabolism, not structure.

Single Photon Emission Computerized Tomography (SPECT): A tracer is an analog of a biologically active compound in which one of the atoms has been replaced by a radioactive atom. When the tracer is introduced into the body, its site-specific uptake can be traced by means of the labeled atom. In SPECT applied to brain imaging these are commonly 99Tc and 123I (gamma emitters) and the patient receives a dose up to about 25 mCi of activity. In the brain, when one of these atoms decays, the emitted photons are detected directly with solid-state detectors. As the name implies, SPECT collects single photons, localization is maintained through the use of collimators - thick perforated steel sheets placed between the patient and detector. The photons can only enter the detector through the holes, which are narrow, and each signal can therefore be projected back to a site of origin. Multiple scans are performed from various angles and then reconstructed back to the original 3-dimensional tracer activity map. SPECT suffers from several drawbacks, with sensitivity being a main issue, since many photons are lost by absorption by the collimator. In addition, ability to image some chemical processes is hindered by the fact that the isotopes are relatively large atoms and cannot be used to label some compounds because of steric hindrance. However, SPECT is cheaper than PET and therefore more widely available.

Positron Emission Tomography (PET): PET is the modality which combines principles of CT and SPECT together. The key advantage of PET is added sensitivity which is obtained by naturally collimating the photons through a physical process, without the use of absorbing collimators. PET imaging requires access to a cyclotron, which produces positron-emitting elements or radioisotopes (11C,13N,15O,18F). These radioisotopes can be easily incorporated into other chemical compounds including normal body components, like oxygen (used to image blood flow) or a drug (used to visualize brain chemical systems), to make a radiopharmaceutical. For PET imaging, the patient is asked to lie on a table similar to that of a CT scanner and receives the radiopharmaceutical, which emits positrons. When the labeled atom disintegrates in the body, the emitted positron comes to rest and annihilates with an electron. Such an event produces two beams of gamma ray at 180 degrees to one another (a requirement of linear momentum conservation and the fact that the positron-electron system right before annihilation form essentially a zero-momentum frame). Within the scanner are rings of detectors containing special crystals that produce energy when struck by a gamma ray. The scanner's electronics record these detected gamma rays and map an image of the area where the radiopharmaceutical was located. The PET detector is set up in such a way as to accept events in which both gamma rays are detected in coincidence. Typically, two photons are identified as coming from a single event if they arrive at detectors within about 15 ns of one another. No collimators are required, since the two crystals which detected the photons must lie along the line, termed line-of-response or LOR, where the annihilation occurred. Absence of physical collimators increases the sensitivity of PET over SPECT by about 10-50 times. Attenuation correction is necessary because deep-seated tissues appear to contain less activity because the annihilation photons stand a higher chance of being absorbed before leaving the body. In PET this correction is almost trivial, since every photon pair along the same LOR is subject to the same attenuation. To perform the correction, a blank and a transmission scan are taken and used to correct the patient scan.

Coupling of Blood Flow to Metabolism
The assays described above: direct quantitation of regional blood flow and contemporaneous, corresponding regional metabolic activity, show that regional flow-metabolism coupling is very tight. The next logical step is towards a more detailed understanding of the precise mechanisms relating blood flow to oxygen consumption.

Neuro-anatomic studies:
To achieve the observed match between blood flow and metabolic activity in all the disparate regions of the cortex the pial arteries, arterioles and capillary pericytes must adjust the microcirculation in a highly complex and precise fashion, both temporally and spatially. If the mechanism of control involves purely local release and detection of chemical mediators from relatively ischemic tissue these would have to be conveyed to flow-controlling resistance vessels very rapidly and precisely. Apparently, regulatory messengers, generated locally and released downstream from a site of metabolic activity, would need to influence the tone of specific upstream vessels; it is not intuitively obvious how this might be accomplished. In an attempt to answer this paradox it has frequently been postulated that neuronal control might be an important element of regulation, with projection of selected neurons from control centers to specific cerebral vessels. These regulatory vessels, under central control, could coordinate regional blood flow to facilitate local metabolic activity. This concept was controversial for a time because there was no obvious link between morphology and function, however it subsequently became apparent that the cerebral blood vessels are indeed directly innervated by several different systems of nerve fiber (including components of the autonomic and sensory networks). Goldman-Rakic and co-workers, for example, were able to demonstrate the existence, in rhesus monkeys, of dopaminergic axons, strategically positioned to adjust both the microcirculation and neuronal activity in discrete cortical regions.13 We also now know that stimulation of specific brain regions such as the basal forebrain elicits changes in cerebral blood flow in distant regions of the cerebral cortex 14 and nerve fibers from the subcortical regions have been shown to project to the basal lamina or perivascular astrocytes of arterioles and capillaries of the cerebral cortex. There is, in fact, now good evidence that cortical arterioles respond directly to local neurally released vasoactive substances and neurotransmitters and also that they posses strategic receptors that regulate cortical perfusion. The local release of regulatory mediators is, in turn, coordinated from areas of the brain quite remote from the region of blood flow control. Experiments in which particular nerves or nuclei are stimulated or inhibited show, for example, that there is an entire network of cerebrovascular sensory innervation derived, principally, from the trigeminal ganglion.15 Activation of the system by stimulation of the nasociliary nerve elicits an increase in cerebral blood flow 16 via a process involving release of substance P and neurokinin-A leading, ultimately, to production of nitric oxide. The trigeminal sensory nerve fibers are situated at the interface between the brain and the circulation facilitating continuous sampling of the microenvironment of the cerebral vessel walls by the trigeminal nerve fibers and continuous modulation of vascular tone, and hence regional blood flow.

Fine resolution mapping is still problematic in human subjects, and much of this type of work has been, of necessity, confined to more submissive animal subjects; it is likely, however, that the flow-control systems of interest have developed within, and been preserved across, a wide range of species. Recent studies in rats have begun to dissect regional differences in cerebral blood flow at a very fine spatial resolution.17 In rats local electrical stimulation of the cerebellar surface evokes action potentials in a very narrow corridor of "parallel" fibers leading to monosynaptic excitatory potentials in the apical parts of Purkinje cells. There is an alternative way of stimulating the same Purkinje cells by exciting "climbing" fibers originating from the brain stem. These fibers wind around the Purkinje cell soma making numerous contacts to the cell body, and their stimulation leads to strong activation of the soma and proximal part of the Purkinje cell’s dendritic tree, yielding complex spikes of activity.18 Using laser Doppler flowmetry it has been possible to show that when parallel fibers are stimulated there is an increase in cerebral blood flow in the upper cortical layers, with decreased flows in deeper layers suggestive of intracortical steal; when climbing fibers are stimulated the blood flow increases in the entire cortex; strong evidence that cerebral blood flow increases are closely coupled spatially to increased neuronal activity; the depth differences studied being of the order of 100 mm. It was also noted that nitric oxide and potassium ions appeared to be the signal molecules connecting the parallel fibers with blood vessels whereas nitric oxide and adenosine coupled the climbing fiber discharges to target vessel relaxation.

Clinical Significance
The combination of functional imaging, detailed ultrastructural analysis and neurophysiologic investigation is being applied with enormous energy in a constellation of research centers and to a wide variety of animal species. Two things are certain: firstly, that the reason we have less than complete understanding of the precise mechanisms regulating regional blood flow in the brain is because the systems involved are, like everything else in the brain, extraordinarily complex, defying superficial analysis; secondly, that the intense research activity will yield a very much more detailed understanding of flow control which will, in time, influence clinical practice.

Organ vitality depends on perfusion:
Adequate perfusion to vital organ beds must be maintained at all times; general anesthesia modifies the normal homeostatic mechanisms of the body - this being a major reason for careful monitoring of patients undergoing surgery. ECG S-T segment analysis permits some degree of warning of cardiac ischemia, urine flow monitoring reassures that renal perfusion is adequate, pulse oximetry, capnography and arterial blood gas monitoring give information about the behavior of the pulmonary beds and all are considered routine intraoperative aids facilitating some degree of oversight of the effectiveness of perfusion and avoidance of organ ischemia in the heart, kidney and lungs. Currently, we have no such reassuring oversight of perfusion to the brain, instead we rely on relatively empirical approaches to brain protection involving maintenance of adequate global perfusion coupled with indirect indicators of cerebral function.19

Volatile anesthetics are vasodilators but coupling is preserved:
Most volatile anesthetics are direct cerebral vasodilators, some also reduce cerebral metabolic demand significantly. If flow-metabolism coupling is intact then regional cerebral blood flow will show only modest increases under general anesthesia (the direct vasodilatation being offset by reduction in metabolic demand); in the absence of coupling anesthesia will induce relative cerebral hyperemia: unhelpful during neurosurgical procedures and dangerous during orthopedic or general surgery on patients with coincident traumatic brain injury. A recent study using blood flow velocity in the middle cerebral artery (VMCA) as an indicator of relative cerebral blood flow 10 showed that the cerebral hyperemia induced by halothane, isoflurane or sevoflurane remains constant for the duration of anesthetic administration, then returns to normal flow during emergence and recovery. The same study also showed that during the relative brain inactivity (burst suppression pattern in EEG) observed at 1.5 MAC isoflurane and sevoflurane cerebral blood flow was relatively reduced, but that the onset of cerebral electrical activity was associated with an increase in VMCA, which persisted for the duration of the burst of electrical activity 10; this type of observation is interpreted to mean that flow-metabolism coupling survives volatile anesthetic administration.20 It is significant that, even in the presence of apparently adequate general anesthesia, sensory outflow elicited by surgical stimulus can influence cerebral blood flow. TCD studies show VMCA increases significantly for several minutes after skin incision in patients under 1-2 MAC isoflurane anesthesia 21 and there is evidence, in children at least, that this effect is abolished or attenuated when there is effective concomitant epidural local anesthesia.

Invasive studies in experimental animals (rats) using radiotracer markers for glucose utilization (14C 2-deoxyglucose) and for regional blood flow (14C antipyrene) show that there is tight coupling between regional blood flow and local glucose metabolism in all of 40 representative brain structures studied and that this coupling persists under 1-2 MAC of isoflurane, desflurane or sevoflurane and in most brain structures although the relationship is reset to higher blood flows. 22, 23

Head trauma:
Post-traumatic hypoxia and hypotension are known to have devastating effects after severe head injury and cerebral oxygen delivery is so low as to be in the ischemia range in a significant proportion of patients experiencing such injury.24 Perversely, local, and even global increases in glycolysis and general oxygen demand occur in 30-40% of patients after severe head injury.25 Secondary cerebral ischemia resulting from this imbalance between oxygen supply and demand is one of the major factors influencing prognosis and outcome in patients with significant head injury.5 Early impairment of cerebral blood flow in patients with significant head injury correlates with poor brain tissue oxygen delivery, and cerebral ischemia occurring after significant head injury is associated with poor outcome.24, 6 Low brain tissue oxygen tension occurs in approximately 30% of patients with severe head trauma in the first 12 hours post-injury.5 The principal therapeutic objective in severe head injury is maintenance of cerebral perfusion and avoidance of tissue hypoxia. How do we monitor cerebral perfusion in the clinical setting?

Low venous oxygen concentration in blood samples taken from the jugular venous bulb generally indicates insufficient oxygen delivery to the brain, and this global ischemia is almost universally unhelpful; however, the converse is not necessarily true, i.e. that normal jugular oxygen concentrations indicate that all is well. After head trauma much of the sensitive autoregulation of the brain may be impaired. If the damage is extensive, global assays of oxygen utilization and lactate production, such as jugular venous sampling, may reveal deficits in tissue oxygenation (or altered oxygen extraction ratios). A normal balance between flow and metabolism will yield a normal jugular oxygen saturation of 60-75% 4 but may provide no information at all about the presence of focal ischemia. Indeed, some recent studies in patients with focal brain pathology using small oxygen electrodes to monitor regional variations in local oxygen tension show that significant focal ischemia can coexist with normal jugular venous blood oxygen extraction ratios.26 The sensors used, typically about 0.5 mm in diameter, contains a miniature Clark electrode for measuring partial pressure of oxygen in brain tissue (after insertion into brain parenchyma via a modified Camino bolt). However it is worth noting that currently the sensors tend to be inaccurate at low oxygen tensions and there is no practical way to test the accuracy of the probe in vivo.24

Similarly, autoregulation of cerebral blood flow in the face of changes in systemic blood pressure or changes in blood CO2 tension are often affected by traumatic brain injury; in fact at many centers autoregulation studies are frequently performed on trauma victims with suspected closed head injury who are about to undergo general anesthesia for associated orthopedic or visceral injuries in order to determine criteria for safe intraoperative management. But, even here, regional heterogeneity of, for example, CO2 vasoresponsivity is much more common than global estimates have suggested 27 leading to the probability that the studies, as currently conducted, underestimate the extent of deficit in many surgical patients.

Anesthesia for neurosurgery:
A principal goal is to preserve the vitality of brain tissue and to facilitate exposure of the surgical target site with minimal tissue retraction. In general the strategies involve maintenance of adequate blood perfusion (and oxygen supply) to the brain and coincident administration of hyperosmolar solutions such as mannitol, reduction of blood CO2 tension and preservation of venous drainage from the cranial vault. In some centers these interventions are monitored objectively (with intraoperative EEG, TCD, jugular venous blood monitoring etc.) in most there are undertaken empirically. It is difficult to imagine comprehensive intraoperative assessment of regional blood flow and metabolism rapidly becoming a practical possibility during open neurosurgical procedures. Local blood flow is monitored sporadically using very small Doppler ultrasound probes directed at specific small vessels in the surgical field, and it is possible that this type of approach could also be used to monitor local oxygen tension, using miniature Clark electrodes. Much more likely, however, is the incorporation of imaging techniques into stereotactic neurosurgical procedures. In common with other vascular surgery (aortic aneurysm repair, coronary artery stenting or bypass procedures) neurovascular procedures are increasing being conducted either percutaneously under imaging guidance, or stereotactically with minimally invasive techniques. For these procedures the prediction must be that, as greater insight is gained into the ways in which cerebral blood supply is regionally regulated, so this knowledge will increasingly, and rapidly, be incorporated into clinical practice. Although it is not yet common, a certain number of surgical procedures are conducted with the assistance of intra-operative CT, and intra-operative MRI and PET imaging are becoming practical realities. Clearly, if we understand how stimulation of a specific control center of the brain affects local blood supply or metabolic activity in distant region, we will increasingly be able to manipulate conditions in the surgical site with precision, avoiding unnecessary global derangement. Similarly, functional imaging of cerebral activity during endovascular or stereotactic surgery is likely to revolutionize the treatment of neurovascular defects, intra-parenchymal tumors and ablation therapy for motor defects, seizure disorder and behavioral pathology. In common with other, once purely experimental, intraoperative monitors (ECG, EEG, TEE etc.) whose routine use was facilitated by miniaturization of components, advanced signal processing and ergonomically acceptable display of information, many of the imaging and monitoring systems discussed in this paper are likely to become commonplace in advanced operating suites within the next decade; we should embrace this development with enthusiasm.

1. Roy CS, Sherrington CS: On the regulation of the blood-supply of the brain. J Physiol 11: 85-, 1890
2. Lovelace WR, Gagge AP: Aero medical aspects of cabin pressurization for military and commercial aircraft. J Aeronaut. Sci. 13: 143-150, 1946
3. Kety SS, Schmidt CF: The determination of cerebral blood flow in man by the use of nitrous oxide in low concentrations. Am J Physiol 14: 353, 1945
4. Lam AM, Mayberg TS: Jugular bulb venous oximetry monitoring. Anesth Clin North America 15: 533-549, 1997
5. Menzel M, Doppenberg EM, Zauner A, Soukup J, Reinert MM, Bullock R: Increased inspired oxygen concentration as a factor in improved brain tissue oxygenation and tissue lactate levels after severe human head injury. J Neurosurg 91(1): 1-10, 1999
6. Kuschinsky W, Paulson OB: Capillary circulation in the brain. Cerebrovasc Brain Metab Rev 4: 261-8, 1992
7. Erecinska M, Silver IA: ATP and brain function. J Cereb Blood Flow Metab 9: 2-19, 1989
8. Bickler PE: Energetics of cerebral metabolism and ion transport. Anesth Clin North Am 10: 563-573, 1992
9. West JB: High Life. Oxford, Oxford University Press, 1998
10. Kuroda Y, Murakami M, Tsuruta J, Murakawa T, Sakabe T: Blood flow velocity of middle cerebral artery during prolonged anesthesia with halothane, isoflurane, and sevoflurane in humans. Anesthesiology 87(3): 527-32, 1997
11. Hornbein TF, Townes BD, Schoene RB, Sutton JR, Houston CS: The cost to the central nervous system of climbing to extremely high altitude. N Engl J Med 321: 1714-19, 1989
12. Michenfelder JD, Theye RA: Cerebral protection bt thiopental during hypoxia. Anesthesiology 39: 510-517, 1973
13. Krimer LS, Muly EC, 3rd, Williams GV, Goldman-Rakic PS: Dopaminergic regulation of cerebral cortical microcirculation. Nat Neurosci 1(4): 286-9, 1998
14. Cohen Z, Bonvento G, Lacombe P, Hamel E: Serotonin in the regulation of brain microcirculation. Prog Neurobiol 50(4): 335-62, 1996
15. Suzuki N, Hardebo JE, Owman C: Origins and pathways of cerebrovascular nerves storing substance P and calcitonin gene-related peptide in rat. Neuroscience 31(2): 427-38, 1989
16. Suzuki N, Hardebo JE, Kahrstrom J, Owman C: Effect on cortical blood flow of electrical stimulation of trigeminal cerebrovascular nerve fibres in the rat. Acta Physiol Scand 138(3): 307-16, 1990
17. Fabricius M, Akgoren N, Dirnagl U, Lauritzen M: Laminar analysis of cerebral blood flow in cortex of rats by laser- Doppler flowmetry: a pilot study.
J Cereb Blood Flow Metab 17(12): 1326-36, 1997
18. Akgoren N, Dalgaard P, Lauritzen M: Cerebral blood flow increases evoked by electrical stimulation of rat cerebellar cortex: relation to excitatory synaptic activity and nitric oxide synthesis. Brain Res 710(1-2): 204-14, 1996
19. Mayberg TM, Bramhall JS: Brain monitoring in neurosurgery. Problems in Anesthesia 9: 87-98, 1997
20. Lam AM, Matta BF, Mayberg TS, Strebel S: Change in cerebral blood flow velocity with onset of EEG silence during inhalation anesthesia in humans: evidence of flow-metabolism coupling? J Cereb Blood Flow Metab 15(4): 714-7, 1995
21. von Knobelsdorff G, Kusagaya H, Werner C, Kochs E, Schulte m Esch J: The effects of surgical stimulation on intracranial hemodynamics.
J Neurosurg Anesthesiol 8(1): 9-14, 1996
22. Lenz C, Rebel A, van Ackern K, Kuschinsky W, Waschke KF: Local cerebral blood flow, local cerebral glucose utilization, and flow- metabolism coupling during sevoflurane versus isoflurane anesthesia in rats. Anesthesiology 89(6): 1480-8, 1998
23. Lenz C, Frietsch T, Futterer C, Rebel A, van Ackern K, Kuschinsky W, Waschke KF: Local coupling of cerebral blood flow to cerebral glucose metabolism during inhalational anesthesia in rats. Anesthesiol 91: 1720-3, 1999
24. Valadka AB, Gopinath SP, Contant CF, Uzura M, Robertson CS: Relationship of brain tissue PO2 to outcome after severe head injury. Neurologic Critical Care 26: 1576-81, 1998
25. Bergsneider M, Hovda DA, Shalmon E: Cerebral hyperglycolysis following severe traumatic brain injury in humans: a positron emission tomographic study.
J Neurosurg 75: 685-693, 1997
26. Gupta AK, Hutchinson PJ, Al-Rawi P, Gupta S, Swart M, Kirkpatrick PJ, Menon DK, Datta AK: Measuring brain tissue oxygenation compared with jugular venous oxygen saturation for monitoring cerebral oxygenation after traumatic brain injury. Anesth Analg 88: 549-53, 1999
27. Marion DW, Bouma GJ: The use of stable xenon-enhanced computed tomographic studies of cerebral blood flow to define changes in cerebral carbon dioxide responsivity caused by severe head injury. Neurosurgery 29: 869-73, 1991