The introduction of four new computerized techniques in medicine has provided scientists with the ability to directly observe and measure the activity of the living human brain as it responds to various stimuli. Visual imaging techniques may provide the necessary tools for the understanding of many central nervous system diseases including that of mental illness. Currently, psychiatry is one of the fields at the leading edge in applying visual imaging techniques to study brain function. Metabolic measures of regional brain activity may provide clues to the identification of relevant brain abnormalities not provided by indirect analysis of whole brain function, identify individuals vulnerable to mental illness, and help in the development of more specific and effective treatments for psychiatric illness.
The measurement of brain activity has been a challenge to scientists for many years. The progression from electroencephalography (EEG) to cerebral blood flow studies, nuclear magnetic resonance (NMR) and positron emission tomography (PET scan) has significantly improved the physician's understanding of brain function. Another brain imaging technique, computerized axial tomography (CAT scan) is discussed elsewhere in this issue (Weinberger and Wyatt).
Historically, one of the earliest methods used to measure brain activity was the electroencephalogram (EEG). It is a noninvasive technique which allows for the measurement of spontaneous patterns of electrical activity as well as cortical activity in response to specific stimuli. One of the major limitations in the use of EEG in psychiatric research has been the difficulty in quantifying large amounts of data recorded from multiple electrode sites. The development of spectral analysis on small and inexpensive laboratory computers brought quantitation to the clinical level, but the high volume of numerical results made neuroanatomical correlations difficult to see. Recent developments have introduced techniques for two-dimensional mapping of the electrical activity of the brain. This method allows for color-coded EEG activity maps and is presented visually on a representation of the cortical surface. Thus far, EEG topographic analysis appears to be compatible with results from other brain activity studies.
Before the advent of current visual imaging techniques it was necessary to rely on conventional radiography which used x-ray opaque dyes (cerebral angiography) or air to replace ventricular fluid (pneumoencephalography). Although these methods yielded interesting data, the detailed study of brain function was difficult because structures often overlapped and had similar densities. In 1948, Kety et al1 ingeniously used a combination of nitrous oxide (a freely diffusable tracer) and arteriovenous measurements to obtain the first data on total cerebral blood flow and metabolism in the living human brain. While Kety and co-workers found no differences between patients with schizophrenia and other subjects, they suggested the possible future value of a regional brain approach. Kety's recognition of the importance of cerebral metabolism and suggestion of the importance of regional approaches provided the impetus for the brain imaging techniques used today.
The next major development, almost two decades later, was the replacement of nitrous oxide with radioactive tracers (eg, "Xenon) which produce a two-dimensional view of the brain and enable investigators to observe and measure cerebral blood flow using external detectors and cathode-ray-tube images. With this technique, Ingvar and Risberg were able to demonstrate for the first time regional differences between patients with schizophrenia and normal controls as well as possible changes incerebral blood flow in various experimental conditions.2
The most recent technical advancement to evolve is the PET scan (positron emission tomography), which has made possible the regional approach suggested by Kety. A brilliant discovery was made by Sokoloff et al3 who introduced methods for quantifying cerebral glucose metabolism. Combined with positron tomography, this techinque provides the means for three-dimensional functional mapping of the living human brain.
This article will review the latest advances in four visual imaging techniques including 133Xe blood flow techniques, EEG and evoked potential topography, nuclear magnetic resonance (NMR) and positron emission tomography (PET scan).
Electroencephalography (EEG) is a technique which measures brain electrical activity by recording electrical impulses from electrodes placed on the scalp. This methodology can be used to assess spontaneous activity as well as responses to specific stimuli. The EEG can be divided into four major rhythms. The alpha rhythm isa 10 cycle/ second pattern which frequently appears in the posterior parietal and occipital regions of the brain in normal individuals resting with their eyes closed. This rhythm is blocked when the eyes open. Delta and theta rhythms which are slower than alpha are associated with drowsiness. Finally, the fast beta rhythm is associated with mental activity. The proportion of each of these brain rhythms is very sensitive to psychoactive medication.4'5 Delta rhythms (associated with drowsiness) have been reported to be more prevalent in the frontal lobes of drug-free schizophrenics as compared to normal controls consistent with decreased blood flow in the frontal cortex of schizophrenics.6'7 The potential for decreased frontal lobe function as a trait marker for schizophrenia is also being explored by PET8 and is a potential for CT sulcal atrophy assessment, NMR scanning and regional evoked potential studies. Regional differences were also reported in the study of affective illness.9
NUCLEAR MAGNETIC RESONANCE IMAGING (NMR)
Nuclear magnetic resonance (NMR) is based on the reception of radio frequencies emitted by nuclei of different elements in the brain following application of a magnetic field.10'" In 1952, Drs. Felix Bloch and Edward Purcell won the Nobel prize for developing NMR spectroscopy techniques. The development of computercoded visual displays of NMR data almost three decades later has revived interest in this potentially powerful diagnostic tool. In a recent review, Pykett described the NMR technique.'2 Atomic nuclei generate small magnetic fields dependent on their spin, properties. The nonzero spin nuclei are made up of those nuclei having odd numbers of protons or neutrons. The NMR technique is capable of detecting the magnetic behavior of the nuclei by measuring the net effect of all the magnetic moments of the nuclei in a sample of material. Early clinical results at Massachusetts General Hospital suggest that NMR is good at detecting dead tissue, local anemia caused by mechanical obstruction of the blood supply, malignancies and degenerative diseases. The soft-tissue analysis has been reported to be superior to that of the xray techniques. A limiting factor for medical practice is that some parts of the body appear to be more amenable to NMR analysis. For example, the head can be kept relatively still while the beating heart may require special high-speed imaging techniques or synchronization of the data over a series of cardiac cycles. Although no psychiatric studies have been completed, this technique potentially allows for the measurement of cell energy metabolism and NMR-sensitive labeled pharmaceuticals. However, the distributions of NMR-sensitive nuclei other than protons in the water are in low concentrations, and current technology has allowed for high resolution only of gray and white matter of brain structures. The potential of the NMR as a technique to identify specific biological markers has yet to be investigated, but it conceptually appears that it might be valuable in the identification of functional cell nuclei.
BLOOD FLOW TECHNIQUES
Cerebral blood flow is closely related to cerebral metabolism (oxygen consumption and glucose use). Regional cerebral blood flow can be studied with the133 Xenon washout techniques. The isotope is injected into the internal carotid artery or breathed as a mixture with oxygen and air. A battery of detectors arranged over the scalp picks up the time course of the appearance and disappearance of the radioactivity. Analysis of this clearance curve yields estimates of blood flow rate under each detector. With intracarotid injection and detectors spaced at I cm intervals, resolution may be in the range of 2 cm. Inhalation techniques yield less resolution but avoid possible hazards associated with carotid injection. Recently, single photon emission tomography has been introduced using l33-Xe or 123-iodoamphetamine'3~!5 which yields a slice image with resolution somewhat less than that achieved with PET scanning but is still capable of visualizing basal ganglia and other major cortical structures.
In some of the first studies on possible blood flow differences in psychiatric patients, Ingvar and Franzen injected 133Xe into the internal carotid artery of 31 schizophrenic patients and ten neurologically normal male alcoholics.16 Thirty-two detectors placed around the outside of the scalp were used to measure the appearance and disappearance of the radioactivity. Analysis of these data yielded estimates of regional blood flow beneath each detector. Subjects were tested in a resting state, usually with their eyes closed. Results indicated that schizophrenic subjects had decreased blood flow in the frontal regions of their brains as compared to controls. More complex tasks revealed that although postcentral blood flows increased equally in both schizophrenics and controls, there was a consistent hypofrontal function in schizophrenics.17 These studies of Ingvar and co-workers emphasize not only the changes in blood flow with task but also the variations between schizophrenics and controls. In several experimental paradigms, Ingvar and Phillipson used "motor ideation" techniques to measure changes in regional blood flow.18 They measured regional blood flow during three conditions. Subjects were initially at rest, then imagined a slow rhythmic clenching of their right hand, and finally, rhythmically squeezed a small rubber balloon. During the rest period the highest flow values were found in the frontal regions. When the subjects were asked to imagine moving their hand, there was a marked increase in the premotor and frontal regions. However, when the actual movement took place, the cerebral blood flow pattern markedly changed to a localized increase in the hand area of the primary motor cortex. Thus, brain areas associated with the planning and organization of behavior and revealing task-related blood now changes are the same ones with low flow in patients with schizophrenia.
Studies of patients with depression have reported a decrease in right hemisphere blood flow.19 Additional studies have reported decreased blood flow in manies and depressives.20'21
PET SCAN TECHNIQUES
The exciting development of positron emission tomography (PET scan) has allowed scientists to measure the functional activity of the living human brain. For the first time it is possible to obtain three-dimensional colored mappings of the metabolic activity of brain regions during specific psychological states or tasks. The methodology is based on a triad of technical advances. A mathematical model, originally developed by Sokoloff and associates in 1977, allows quantification of regional brain glucose use from a glucose analogue autoradiography technique.3 The development of small but highly sensitive crystal detection and coincidence counting electronics, together with the mathematical reconstruction techniques similar to those used for CT, allows quantitative images to be generated. Glucose metabolism is an indicator of brain work since cerebral tissues under normal circumstances derive most of their energy from glucose. The amount of glucose metabolized in a specific structure at a certain point in time is related to the amount of cell firing and repolarization activity in that neuronal structure. Preliminary work in animals suggests that glucose metabolism is sensitive to changes in functional activity including sleep, electrical stimulation, olfactory stimuli, circadian rhythms, visual and auditory activity, stress and the administration of various pharmacological agents. Furthermore, these findings are consistent with other measures of brain activity such as blood flow studies and EEG recordings.
PET scan technology is a fascinating combination of several scientific disciplines including physics, computer science, neurology, radiology, pharmacology, biochemistry, psychiatry, and mathematics. Basically, PET scan methodology involves the intravenous administration of a neutron-deficient isotope which contains a proton. Because of the rapid decay of most common elements (oxygen-5, two minutes; nitrogen- 13, ten minutes; carbon-11, 20 minutes), fluorine-I8 has been chosen as the most valuable label as it has a half-life of 1 10 minutes and therefore allows for longer periods of analysis. The proton decays to a neutron and a positron is emitted. Initially, the positron is at a very high energy level but interacts with an electron in a very short distance and spontaneously annihilates. Energy is then released in the form of two high energy photons (or gamma rays) which travel in opposite directions (180 degrees apart). Multiple radiation detectors which are arranged in a circle are activated only when both photons hit detectors 180 degrees apart simultaneously. This technique accurately determines the line segment in which the isotope resided. The data are then analyzed by a computer which constructs three-dimensional color-coded slice images of brain structures demonstrating a quantitative display of biochemical activity of glucose use. Currently, in addition to deoxyglucose, scientists are considering labeling compounds which can be used to study neuroreceptors (ie, protein synthesizer, and labeled psychoactive agents such as neuroleptics and benzodiazepines). Some of these agents include 18F-labeled haloperidol, l8F-labeled spiroperidol, nC-labeled pimozide, and "C-labeled flunitrazepam.
More than 500 patients and normal volunteers have been studied using PET scan techniques in seven operating centers. Thus far, evidence suggests that it is possible to study various disease states and their effects on brain function. In a series of studies in man, glucose metabolism was shown to be altered in response to auditory, verbal, visual and shock stimuli, and during periods of aphasia. Recent findings using PET scan techniques comparing schizophrenics to normals have been very promising. Ina longitudinal study, Farkasetal studied a 45-year-old patient with a history of schizophrenia since the age of 16.21 Although the patient had never had neuroleptic medication before the scan, there was still a 40% decrease in frontal cortex glucose use as compared to controls. When the patient was treated with phenothiazines, there was an apparent return to normal glucose utilization levels.
In our recent report,8 local cerebral glucose use was measured with 18F-labeled 2-fluorodeoxyglucose (FDG) in eight unmedicated schizophrenic patients and six agematched normal volunteers. Subjects sat resting in a quiet darkened room with eyes closed after the injection of an isotope-labeled glucose analogue. Following uptake, we obtained seven to eight horizontal brain scans parallel to a line connecting the outer canthus of the eye with the external auditory meatus. Patients with schizophrenia showed relatively lower glucose use in frontal than occipital cortex than normal controls. These results are consistent with previous blood flow studies which demonstrated decreased blood flow in schizophrenic patients during similar experimental (rest-eyes closed) conditions. Diminished pain responses in schizophrenics and affectively ill patients were also demonstrated in both PET scan and blood flow studies, again illustrating a consistent relationship between the two techniques.
Although the PET scan technique is in its early developmental stages, there appears to be promise in the future that it will have a major impact on the understanding and treatment of mental illness. In fact, there is some hope that PET scan results will have the potential of serving as trait- or state-markers of mental illness and will someday help in the identification of highrisk individuals. For example, it may be possible that abnormal metabolic brain functioning will alert the physician to the potential of developing mental illness.
1. Kety SS. et al: Cerebral blood flow and metabolism in schizophrenia. Am J Psychiatry I948; 104:756-770.
2. Ingvar DH. Risberg J: Increase of regional cerebral blood flow during mental effort, in normals and in patients with focal brain disorders. Exp Brain Res 1967; 3:195-211.
3. Sokoloff L. et al: The [14C] deoxyglucose method for the measurement of local cerebral glucose utilization; Theory, procedure and normal values in the conscious and anaesthetized rat. J Neurochem 1977; 28:897-916.
4. Fink M: EEG profiles and bioavailability measures of psychoactive drugs. in ltil TM (ed): Psychotropic Drugs and the Human EEG. Basel. S Karger. 1974.
5. Saletu B, et al: Classification and assessment of pharmacodynamics of SGD-SCHA 1059 (Binodine) by quantitative EEG and psychometric analyses, in Perris CP, von Knorring L (eds): Advances in Biological Psychiatry. ed4. Clinical NeiiraphvsialogicalA speets of ' Psychopathologie»! Conditions. Basel, Switzerland. S Karger 1980, pp 140-166.
6. Buchsbaum MS. et al: The effect of sleep deprivation on average evoked responses in depressed patients and in normals. Biol Psychiatry 1981; 16:351-363,
7. Morihisa JM, et al: Topographic analysis of computer processed electroencephalography in schizophrenia, in Usdin E, Hanin i (eds): Biological Markers in Psychiatry and Neurology. IVew York, Pergamon Press, 1982, pp 495-504.
8. Buchsbaum MS, et al: Cerebral glucography with positron tomography. Arch Gen Psychiatry 1982; 39:251-259.
9. Shagass C, et al: Relationships between psychiatric diagnosis and some quantitative EEG variables. Arch Gen Psychiatry 1982; 39:1423-1435.
10. Brownell GL, et al: Positron tomography and nuclear magnetic resonance imaging. Science 1982; 215:619-626.
11. James AE. et al: Nuclear magnetic resonance imaging: A promising technique. JAMA 1982: 247:1331-1334.
12. Pykett IL: NMR imaging in medicine. Scientific American May 1982, pp 7888.
13. Cornar D. et al: Brain distribution and kinetics of "C-chlorpromazine in schizophrenics. Psychiatry Res 1979; 1:23-29.
14. Kuhl DE, el al: Quantifying local cerebral blood flow by N-isopropyl-p[l; 'liodoa mphetamine (IMP) tomography. J Nucl Med 1982; 23:196-203.
15. Rush AJ. et al: Cerebral blood flow in depression and mania, in Buchsbaum MS, Usdin E, Bunney WE Jr. et al (eds): Brain Imaging in Psychiatry and Neurology: Positron Emission Tomography and Other Techniques. Pacific Grove, California, Boxwood Press, 1982.
16. Ingvar DH. Franzen G: Distribution of cerebral activity in schizophrenia. Lancet 1974; 2:1484-1486.
17. Franzen G, Ingvar DH: Absence of activation in frontal structures during psychological testing of chronic schizophrenics. J Neurol Neurosurg Psychiatry 1975; 38:1027-1032.
18. Ingvar DH, Phillipson L: Distribution of cerebral blood flow in the dominant hemisphere during motor ideation and motor performance. Ann Neurol 1977; 2:230-237.
19. Mathew R, et al: Cerebral blood flow in depression. Am J Psychiatry 1980: 137:1449-1450.
20. Gustafson L, et al: Cerebral blood flow in dementia and depression. Lancet 1981; 1:275.
21. Silfverskiold P.etaJ: RCBF in affective disorders, in Bushsbaum MS. Usdin E. Bunney WE J r. el al (eds): Brain Imaging in Psychiatry and Neurology: Positron Emission Tomography and Other Techniques. Pacific Grove, California. USA: Boxwood; Synapse, to be published.
22. Farkas T. et al: Regional cerebral glucose utilization in schizophrenia. Presented al the Third World Congress of Biological Psychiatry. Symposium on Cerebral Circulation and Metabolism Related to Psychopathology, Stockholm, Sweden. 1981.