2002-03-29
Below extracts (italicized) from Scientific American, October 1978, p 50-59
Comments in normal fonttype. The original contains numerous informative colored illustrations that have not been included here.

p 50

Brain Function and Blood Flow

by Niels A. Lassen, David H. Ingvar and Erik Skinhøj
- - -
After an overview of the history of related brain studies,
the need for more acurate methods is stated

...A more direct approach to the localization problem is to study the intact cerebral cortex in various functional situations: during "brain work." One technique is to record with minute electrodes the altered firing rate of individual nerve cells in the cortex. Yet not even large arrays of microelectrodes are adequate for deciphering the vastly complex interplay among the 10 billion nerve cells in the brain. We have taken an entirely different and more holistic approach: the measurement of the enhanced blood supply to cortical areas that are activated by the performance of specific sensory, motor and mental tasks.

p 53

...Later it became possible to measure the blood flow in circumscribed regions of the intact human brain with the aid of radioactive isotopes. In 1961 the three of us developed this principle from studies in cats, and its application to diagnostic examinations in man soon followed. The method involves the use of Xenon 133, a radioactive isotope of the inert gas Xenon. The radioactive gas is dissolved in sterile saline solution, and a small volume (two to three millilitres, containing from three to five millicuries of radioactivity) is injected as a bolus into one of the main arteries to the brain. The arrival and subsequent washout of the radioactivity from many brain regions is followed for one minute with a gamma-ray camera consisting of a battery of 254 externally placed scintillation detectors, each of which is collimated to scan approximately one square cm of brain surface. Information from the detectors is processed by a small digital computer and is displayed in graphical form on a color-televison monitor, with each flow level being asigned a different color or hue. Owing to the attenuation of radiation from structures deeper in the brain, the gamma radiation detected comes from the superficial cerebral cortex. Thus the radiactive-Xenon technique provides a fairly specific picture of the activity of the cerebral cortex directly below the detector array.
We use the technique routinely on patients in whom cerebral arteriography has to be performed. Cerebral arteriography involves making an X-ray plate of the blood vessels in the brain after an X-ray opaque medium has been injected into the arteries. This injection is made through a catheter (a fine plastic tube) placed directly in the internal carotid artery through a puncture in the neck or passed up the femoral artery through a puncture in the groin. Since the same catheter is used for radioactive-Xenon injections, the measurement of cerebral blood flow incurs no independent risk to the patient. (The low level of the gamma radiation emitted by Xenon 133 is not considered harmful.)
In our laboratories in Copenhagen and Lund some 500 patients have had their brain examined with the radioactive-Xenon technique for diagnostic purposes such as the study of strokes, tumors or epilepsy. In retrospect we can say that about 80 of these patients had a normal brain at the time of measurement. This group consisted of patients who suffered from severe headache attacks, generalized epileptic seizures and other transient neurological symptoms that turned out not to be associated with permanent brain lesions or abnormalities. Electroencephalography ("brain wave" recordings) and other tests also served to confirm the normality of their brains. Our studies of the regional cerebral flow in these patients therefore enabled us to draw some conclusions about the localization of function in the normal cerebral cortex.
the flow is always substantialIy higher in the front part of the cortex than in the central and rear parts. Although as we have mentioned the mean flow in normal subjects is 50 millimeters per 100 grams of brain tissue per minute, the flow in the front part of the cortex is as much as 20 to 30 percent above this mean, and the flow in the rear regions, particularly in some parts of the temporal lobe. is correspondingly lower. The density of both the capillaries and the nerve cells in the various regions of the cortex is about the same, and so the remarkable difference in flow rate suggests that the overall activity level of the front part of the resting brain is about 50 percent higher than that of the rear parts.
This "hyperfrontal" resting flow pattern may contribute to an understanding of conscious awareness. It is weIl known that the frontmost parts of the frontal lobe, the prefrontal areas, are responsible for the planning of behavior in its widest sense, whereas the remaining regions of the cortex subserve motor and sensory functions. The hyperfrontal resting flow pattern therefore suggests that in the conscious waking state the brain is busy planning and selecting different behavioral patterns. In the same

p 54

state the motor and sensory regions of the cortex are not very active; they are perhaps even inhibited. This interpretation seems to agree with subjective experience. While one is at rest one is not continuously aware of one's sensory input; only occasionaly does one perceive distinct visual, auditory or tactile signals that stand out from the background "noise" of the resting state. Most of resting awareness is focused on inner thoughts, particularly on reflections on one's own situation and its relation to past events and to possible future ones. The resting conscious brain can therefore be said to be primarily engaged in the simulation of behavior. What is the effect of simple sensory stimuli on the pattern of regional blood flow in the cortex? For these experiments our computer was programmed so that only departures from the resting pattern of blood flow were displayed in colors on the screen. When the subject opened his eyes and looked at an object, the pattern of the cortical blood flow changed dramatically: an increase of about 20 per cent was seen in the visual association cortex, located in the temporal and the occipital lobes. (The primary visual cortex deep in the occipital lobe at the rear of the brain was not seen because this area is supplied by the vertebral artery and hence did not receive the radioactive Xenon injected into the carotid artery.) In addition a well-localized part of the premotor cortex, the frontal eye field, became active. Auditory stimulation in the form of a loud, meaningless noise increased the blood flow near the upper rear part of the temporal lobe on each side of the brain, where the primary auditory cortex and the auditory association cortex are located. Flow in these areas was further increased by hearing simple spoken words such as "bang," "zoom" and "crack." The activated region includes Wernicke's area in the left hemisphere. which is involved in the understanding of spoken language. When spoken words were heard with the eyes closed. the frontal eye field in the premotor cortex was slightly activated. More complex verbal stimuli caused an increase in the regional blood flow in the lower rear part of the frontal lobe, where Broca's speech center is located on the left side. The effects of tactile perception were studied by Per Roland in Copenhagen, working in collaboration with one of us (Lassen). The subjects were asked to indicate verbally which was the larger of two objects (small metal bars) placed one after the other in the palm of the hand, with the fingers kept motionless. This tactile stimulus activated the hand area of the primary somatosensory cortex in the central part of the opposite

(The illustration is edited compared with the original: enlarged text)

CELLULAR ARCHITECTURE of the cerebral cortex is organized into columnar modules made up of vertically arranged circuits of nerve cells. A typical column, some 250 micrometers (10-3 millimeter) in diameter, is shown in this highly simplified diagram based on one by János Szentágothai of the Semmelweis University Medical School in Budapest. Hundreds of incoming nerve fibers carrying sensory information converge on the spiny stellate cells in Layer IV of the cortex. The vertical circuits of interneurons, arranged in a highly specific spatial configuration, transform this raw data into the subtleties of conscious experience and behavior.

p 55

cerebral hemisphere, as weIl as the adjacent association cortex in the parietal lobe. (As is weIl known, the sensory and motor functions of the limbs are controlled by the hemisphere on the opposite side of the body.)
The performance of these simple sensory tests requires certain memory functions: the sensory inputs preceding the input arriving at any given moment must be retained, and the cumulative input must be compared with previous experience. A hypothesis of what the perceived entity means is then formed, and identification is completed by an active search for characteristic features. This process involves associations, in which memory plays a key role. In our experiments we noted that each type of sensory stimulus activated both the primary sensory cortex and the adjacent association area for that specific sensory modality. These areas were the only ones that were consistently activated by sensory input, suggesting that modality-specific forms of memory are localized in the association cortex specific for that modality. It should be emphasized that we have not studied memory functions other than those directly associated with sensory perception. Short-term memory (which operates, for example, in the immediate recall of a telephone number or a person's name) is generally considered to be localized in the region called the hippocampus in the deeper parts of the temporal lobe. Like the primary visual cortex, the hippocampus gets its blood from the vertebral artery and hence is not represented in our carotid-artery injection pictures. Voluntary movements of the hand have been studied by Jes Olesen in Copenhagen in collaboration with one of us (Lassen). In 1970 it was observed that when the fist on the side of the body opposite the cerebral hemisphere being examined was rhythmically clenched, there was an increase in blood flow in the hand area of the primary motor cortex in the central part of the brain. Hand movements also caused an increase of flow in an adjacent area of the primary somatosensory cortex, which receives feedback signals from the skin, tendons and muscles of the hand while the hand is moving. As one would expect, however, when the hand being clenched was

p 56

on the same side of the body as the hemisphere being scanned, there was no change in the flow pattern. Recently Marcus E. Raichle and his colleagues at the Mallinckrodt Institute of Radiology in St. Louis studied the controI of hand movements by injecting a radioactive isotope of oxygen into the brain arteries in order to directly follow regional oxygen consumption by the brain tissue. They found that hand movements increased the oxygen uptake in the same regions of the cortex where we had observed an increase in regional blood flow. This finding provides direct support for the basic assumption on which our interpretation of the blood-flow data rests: that local changes in blood flow reflect local variations in the intensity of nerve-cell metabolism. Voluntary movements of the mouth in speech cause a well-defined activation of the cortical area that controls the movements of the mouth, the tongue and the larynx; voluntary movements of the foot activate the part of the motor cortex in the opposite hemisphere above the part activated by movements of the hand. These findings and others confirm that the primary somatosensory and motor areas are organized as two adjacent narrow bands extending from ear to ear across the top of the cortex. The maplike relation between the parts of the body and the somatosensory and motor areas of the cortex has been known in detail since Penfield and his colleagues plotted these areas by electrically stimulating the cortex. The map resembles a distorted homunculus with an enlarged head pointing toward the temporal lobe, an enlarged hand and thumb in the middle and a reduced foot at the top, reaching the inner side of the hemisphere.
Contractions of voluntary muscle also activate the premotor cortex in the upper part of the frontal lobe. Such activation always involves both hemispheres and as far as we can tell is located in the same area regardless of whether the mouth, the eyes, a hand or a foot is moved. When the detectors of our apparatus are placed above the subjects head, the most marked change in regional cerebral blood flow occurs close to the midline and involves a region of the premotor cortex on the inner part of each hemisphere called the supplementary motor area. This area is known to play a roIe in complex motor tasks of all kinds, including speech. Nevertheless, to see the supplementary motor area and the surrounding premotor cortex so consistently and massively activated during voluntary movements somewhat surprised us. We found that activation of the supplementary motor areas was more marked during dynamic muscle movements, such as

p 57

operating a typewriter, than it was during steady muscular contractions. For this reason, and because of supporting evidence in the scientific literature, we have concluded that the upper premotor cortex, including the supplementary motor area, is involved in the planning of sequential motor tasks. Here it is relevant to mention a recent experiment on the nature of voluntary movement. In both Copenhagen and Lund we have studied the difference between the pattern of regional cerebral blood flow that appeared when a simple sequence of finger movements was being performed and the pattern that appeared when the subject was merely thinking about performing the sequence. With suitable instructions the subject could perform the movement mentally in the correct temporal sequence while keeping the hand perfectly still; the imagined movement activated the supplementary motor area. When the sequence of movements was actually performed, the hand-finger area of the primary motor cortex and the related areas of the somatosensory cortex also became active. These findings suggest that the supplementary motor area is a programmer of dynamic movement, whereas the primary sensory cortex is the controller and the primary motor cortex is the executor. We have investigated speech processes in detail. Here we were impressed to find that both the right and the left hemispheres become active in much the same manner. As we have mentioned, listening to simple words activates the auditory cortex in both hemispheres. Speaking aloud activates three more areas, namely the face, tongue and mouth areas of the somatosensory and
motor cortexes, the upper premotor cortex in both hemispheres (which includes the supplementary motor area) and Broca's area in the lower rear part of the left frontallobe and the corresponding part of the right frontal lobe. Reading aloud adds activation of the visual association cortex and the frontal eye fields as well as the primary visual cortex (although the last is only inferred, since it is not made visible in our technique), Thus seven discrete cortical regions are simultaneously active, forming a Z-like figure on the surface of each hemisphere. It was interesting to note the difference between reading aloud and merely reading: reading in itself does not activate the mouth areas of the somatosensory cortex or the motor cortex or the auditory areas, although the five other areas are active. Studies of the effects of brain damage

p 58

on speech have revealed that destruction of Broca's area in the left hemisphere results in motor aphasia, that is the loss of the ability to speak more than simple words but not the loss of the ability to understand spoken and written language. Destruction of the corresponding area in the right hemisphere, however, has no discernible effect on speech. We were therefore surprised to observe that this part of the right hemisphere was active during verbalization, suggesting that it makes some contribution (albeit a nonessential one) to the final synthesis and mobilization of speech. In Copenhagen, Borge Larsen, working with one of us (Lassen), analyzed his observations further to see if some slight differences in the blood-flow response of the two hemispheres during speech could be discerned. Although for ethical reasons he could not measure the flow in the right and left hemispheres of the same subject, Larsen's results suggest the following differences: in the left hemisphere an increase in flow is usually seen in the mouth area and the auditory cortex separately, whereas in the right hemisphere the two often form one confluent active region. Moreover, the supplementary motor area in the left hemisphere is usually more active during speech than the one in the right hemisphere. The analysis of cortical activation during reading illustrates, that a complex task is carried out by several circumscribed cortical regions brought in to action in a specific pattern. This system is analogous to a computer program in which different subroutines are brought into play depending on the problem to be solved. In general our results confirm a conclusion reached by the late A. V. Luria of Moscow State University on the basis of his neuropsychological analyses of patients with brain damage: "Complex behavioral processes are in fact not localized but are distributed in the brain, and the contribution of each cortical zone to the entire functional system is very specific." Early in our studies of various forms of brain activation we recognized that many of our patients in the conscious waking state showed not only local increases in the blood flow of specific regions of the cortex but also an increase in the blood flow of the cerebrum as a whole. Jarl Risberg and one of us (Ingvar) found that psychological testing with simple routines of recall and reasoning causes in addition to localized changes a significant overall increase in the cerebral blood flow of about 10 percent. This general increase in blood flow appears to be distinctly related to the subject's effort in performing the task, because it is absent when simple tasks are performed but is evident when the subject shows signs

p 59

of struggling with a difficult problem. This finding supports a distinction made by neurophysiologists between the specific and the nonspecific pathways of the brain. The specific sensory and motor pathways arrive and depart from well-defined areas of the cortex that be come active during the reception and interpretation of specific sensory messages or the execution of a specific motor task. At the same time demanding tasks activate larger cortical areas over diffuse pathways that fan out from the reticular formation of the brain stem and the thalamus of the midbrain. Animal studies have shown that in the absence of an activation of the diffuse nonspecific pathways the specific pathways by themselves do not appreciably alter the activity of the brain outside the regions of the brain to which the specific pathways project. Hence it appears that for the brain to "understand" the surrounding world, to perceive its meaning and to take action in difficult tasks the cerebral cortex must be activated not only locally but also totally. A correspondence between the general activation of the cortex and the level of consciousness during the performance of various tasks has been strongly supported by studies of anxiety and pain. A threat to one's body or psychological well-being elicited by a pain stimulus or by strong anxiety provokes a drama tic increase in one's awareness of self and the environment and also causes a generalized increase in cerebral metabolism and blood flow. Stress activates pathways in the brain and also triggers the secretion of the hormone epinephrine (adrenalin) from the adrenal glands, which has a general arousing effect on the body. Bo Siesjö and his coworkers at the University Hospital in Lund recently studied rats waking up from anesthesia after they had been paralyzed with a muscle-relaxant drug of the curare type. The stress of the animals' waking in a state of paralysis gave rise to an enormous increase in the rate of their cerebral oxygen uptake and blood flow; the increase could be diminished, however, by the removal of the adrenals and could be completely abolished with drugs that block epinephrine receptors. In Lund we recently studied the effects on the pattern of human cerebral blood flow of a pain stimulus: an electric shock applied to the skin. (All such experiments were of course performed with the patients' full informed consent and as a basis for clinical diagnostic tests.) When the intensity of the stimulus was low, just above the threshold, the shock was experienced much as a simple touch: when the intensity was high, the shock was experienced as a moderately painful sensation. As one would expect, the threshold stimulus did not measurably alter the mean hemispheric blood flow or oxygen uptake, although there was a small increase in regional blood flow in the front and upper parts of the frontal lobe. Moderate pain stimulation, on the other hand, gave rise to a general increase of 20 percent in the mean hemispheric blood flow and oxygen uptake, and there were localized increases above the mean in the frontal lobes. Thus touch and moderate pain appear to make the brain more aware, or more conscious. These observations support the hypothesis that the general activation of the brain is accompanied not only by an arousal of electroencephalographic activity but also by an increase in cerebral blood flow and oxygen uptake, and that this reaction is related to an increased level of awareness. The global activation of the brain is probably also related to the emotional components of experience, although we have not yet made any systematic observations of this kind. What will cortical mapping be like in the future? The radioactive-Xenon technique (and a noninvasive variant of the technique in which the radioactive isotope is inhaled rather than injected) is a somewhat crude one. The results mainly reflect events in the superficial layers of the cortex and leave out deeper structures; moreover, the time resolution is low (minutes at the most). These limitations, together with recent technical developments, have inspired refinements of our technique. Complex multidetector instruments are now being developed for measuring the distribution of the radioactive isotope inside the head in three dimensions, so that activity in deeper parts of the brain can be analyzed. In addition the application of radioactive isotopes to investigations of brain function has opened up the new field of regional metabolic studies. Louis Sokoloff and Martin Reivich at the National Institute of Mental Health in the U.S. have studied cerebral metabolism on a microscopic scale by injecting a radioactive analogue of glucose into the brain; the rate at which the substance is taken up by nerve cells reflects their functional activity. Such experiments have revealed that the metabolic rate in very small regions of the brain changes in consistent patterns during various activities. For example, the illumination of one eye in monkeys alters the consumption of glucose in the visual cortex in columns of nerve cells less than a millimeter apart; these functional columnar units correspond to those demonstrated with neurophysiological techniques by David H. Hubel and Torsten N. Wiesel of the Harvard Medical School. The sokoloff-Reivich technique is now being modified for clinical purposes. The coming generation of powerful integrative techniques and tools will set the stage for a new type of clinical neurophysiology: an era in which the regional circulatory and biochemical accompaniments to the functions of the human brain can be both precisely localized and measured quantitatively. Such methods will bring us closer to perceiving the intricate patterns of activity that underlie the functioning of the most complex of all biological systems, the human brain.

Return to Introduction