During the past 2 centuries, clinical investigators have been correlating disordered function with abnormalities found in different parts of the brain. Normal functions are deduced from the effects of destructive lesions. Experimentation with animals provides more precise information about the ways that populations of neurons are interconnected. If the same connections are seen in a variety of mammalian species, it is reasonable to suspect that the human nervous system is similarly organized.
Imaging the Nervous System
Since the 1970s, methods have been available to make images of the living human brain that are almost as accurate as the observations of a pathologist. Therefore, it is possible to record symptoms and physical signs and identify the affected parts at the same time. Images can also be made that show regions of the brain with increased metabolic activity associated with sensory, motor, or mental tasks.
The techniques used for obtaining structural and functional information about the living human brain are summarized in Table 4-1. Notes follow on the methods that provide the most anatomical and functional information.
A plain radiograph of the head or spinal column provides hardly any information about the normal anatomy of the brain or spinal cord. In adults, displacement of the calcified pineal gland (see Chapter 11) can reveal displacement of midline structures. With contrast media, the images are more informative. In angiography, a radiopaque liquid injected into one of the carotid or vertebral arteries shows the branches of these vessels and, 1 or more seconds later,
the veins. Some normal angiograms are shown in Chapter 25. The chief value of this technique is for detecting arterial disease (occlusion, stenosis, aneurysm) or displacements of blood vessels by lesions such as tumors. Computed tomography (CT) has replaced the older techniques of pneumoencephalography and ventriculography (see Table 4-1).
TABLE 4-1 Neuroimaging Methods
This application of radiographic imaging is based on scanning the head with a narrow, moving beam of radiographs and measuring the attenuation of the emerging beam. The density readings from thin “sections” (tomograms) of the head are processed by a computer to generate an image whose brightness depends on the absorption values of the tissues. The technique is valuable in clinical diagnosis because the density of many cerebral lesions is greater or less than the density of normal brain tissue.
To avoid irradiation of the eyes, the “axial” plane of the sections imaged by CT is oblique, being somewhat closer to horizontal than to coronal. Special neuroanatomical atlases are available in which CT scans are compared with photographs of slices of the brain cut in the same plane.
Nuclear Magnetic Resonance Imaging
This imaging technique was developed from nuclear magnetic resonance (NMR), a physical method used in chemical analysis. In a strong magnetic field, the nuclei of atoms absorb radiofrequency energy. The absorbed frequency is characteristic of the element and of the immediate molecular environment of its atoms. In diagnostic magnetic resonance imaging (MRI), a frequency is chosen that is absorbed mainly by the nuclei of the hydrogen atoms of water. The patient's head is put into a magnetic field and irradiated with the radiofrequency radiation for protons. The measured energy absorptions are integrated in a computer, which generates a series of images of sections through the head. The sections may be reconstructed in any plane. Horizontal sections (parallel to the plane passing through the anterior and posterior commissures), as well as sagittal and coronal sections, are commonly presented. The reconstructed slices are typically 4 or 5 mm thick.
Images are commonly prepared in three ways, taking advantage of different components of the NMR signal. A T1-weighted image emphasizes the difference between central nervous tissue (brighter) and other fluids and tissues (dark) and gives some discrimination between gray matter (brighter) and white matter (less bright). A T2-weighted image emphasizes the cerebrospinal fluid (CSF; bright) in the subarachnoid space and ventricles, providing crisp anatomical resolution but with poor contrast between gray and white matter. A proton density image emphasizes the difference between gray matter (bright) and white matter (darker). Examples of images of the brain are shown in Figures 4-1, 4-2, and 4-3. Later chapters, especially Chapter 16, explain the anatomy displayed in these images. NMR data can also be processed to show the larger blood vessels (MRI angiography), but the images show less detail than conventional radiographic angiograms.
The advantages of MRI are that no potentially harmful radiation is used, and the anatomical resolution is greatly superior to that obtainable with radiographs. Bone and flowing blood are invisible in MRI images. Gray and white matter and CSF have different densities, and it is sometimes possible to identify regions of white matter that contain degenerating axons. A special contrast medium (a gadolinium compound) can be introduced into the circulation to reveal regions of the brain with abnormally permeable blood vessels, which often occur at sites of disease. The chief disadvantage is that MRI is a slow process, requiring about 1 hour in contrast to a few minutes for a CT scan.
Functional Nuclear Magnetic Resonance Imaging
Increased blood flow and oxygen usage accompany neuronal activity. In the course of NMR imaging, blood oxygen level-dependent (BOLD) signals can be collected that relate to oxygen concentration in the tissue being examined. Locally, high levels of metabolic activity can be translated into signals of high intensity in an image, thus rendering prominent any parts of the brain that are more active than the surrounding regions. Fine anatomical resolution and an indicator of function are obtained without the use of radiographs or radioactive isotopes. This technique is being
used extensively for studies of cerebral metabolism in normal mental and physical activities, and it may become important in the diagnosis of diseases.
FIGURE 4-1 Sagittal T1-weighted nuclear magnetic resonance image of the normal brain. Note that compact bone and flowing blood are not visible. Many other neuroanatomical features can be seen, including the paracentral lobule, the fornix, a mamillary body, the cerebral aqueduct, the pons, and the medulla. Compare this image with Figure 1-3. CS, calcarine sulcus; Cb, cerebellum; G, genu of corpus callosum; M, marrow in parietal bone; S, scalp. (Courtesy of Dr. D. M. Pelz.)
FIGURE 4-2 Three coronal magnetic resonance images of a plane that contains the insula, lentiform nucleus, internal capsule, and head of caudate nucleus. (A) T1-weighted image: T, trunk of corpus callosum; L, lateral ventricle; S, septum pellucidum. Three coronal magnetic resonance images of a plane that contains the insula, lentiform nucleus, internal capsule, and head of caudate nucleus. (B)T2-weighted image: C, cingulate sulcus with callosomarginal artery; L, lateral ventricle. M, middle cerebral vessels in subarachnoid space; S, superior temporal gyrus. (C) Proton density image: A, amygdaloid body; C, head of caudate nucleus; IC, internal capsule; L, lentiform nucleus. (Courtesy of Dr. D. M. Pelz.)
FIGURE 4-2 (continued)
Two types of BOLD signals can be exploited. In T2* BOLD functional MRI (fMRI), the most frequently used type, the signal is due to deoxygenated hemoglobin, and “activation” is comonly
imaged in large veins as well as in nervous tissue. More accurate localization is achieved with Hahn spin-echo (HSE) BOLD fMRI, which detects movements of water into and out of red blood cells, with much stronger signals arising from capillaries than from larger vessels. The spatial resolution of HSE BOLD fMRI is about 0.1 mm, but this technique is available only in certain research institutes because it requires a stronger magnetic field (7 to 9 Tesla) than that ordinarily used (1.5 Tesla) for MRI.
FIGURE 4-3 A proton density nuclear magnetic resonance image in a horizontal plane through the level of the insula. C, head of caudate nucleus; E, external capsule; F, forceps frontalis; L, lentiform nucleus; LV, lateral ventricle; T, thalamus; V, primary visual cortex. (Courtesy of Dr. D. M. Pelz.)
FUNCTIONAL MAPPING WITH RADIOACTIVE TRACERS
Structure can be related to function by mapping the distribution of a metabolically significant substance that has been labeled with a radioactive isotope.
Regional Cerebral Blood Flow
Although the flow of blood through the whole brain does not change much, transient but conspicuous local increases in flow are associated with the activity of neurons. To monitor regional cerebral blood flow, a radioactive tracer such as 133Xe is introduced into the blood, and the intensities of the emitted gamma rays are measured by an array of detectors at the surface of the patient's head. The intensity of the radiation at any point varies with the vascular perfusion of the underlying tissues. The method is used for examining different parts of the cerebral cortex. A computer integrates the measurements of radioactivity and provides anatomical images of the functioning areas. Clinicians use this method to identify cortical regions in which the circulation is inadequate, and research with normal volunteers provides evidence of functional localization in the cerebral cortex.
Regional cerebral blood flow can also be studied by single-photon emission computed tomography (SPECT) using 133Xe or 99Tc as the tracer and by positron emission tomography (PET) using [15O]carbon dioxide. SPECT and PET provide sets of reconstructed sections, thereby providing information about both the cerebral cortex and the interior of the brain.
Single-Photon Emission Computed Tomography
Each disintegrating atom of an ordinary gamma-emitting isotope emits one photon. The SPECT technique makes sectional maps based on the
uptake into tissue and subsequent dispersal of radiolabeled compounds that have been introduced into the blood. Regional blood flow is represented as variations of intensity in the resulting images. The images have low resolution (2-3 cm), but they are obtained in only a fraction of the time needed for a PET scan and at much lower cost.
Positron Emission Tomography
Positrons are emitted by certain radioactive isotopes, of which 15O, 13N, 11C, and 18F are the most useful. A positron is immediately annihilated when it encounters an electron and two gamma-ray photons are emitted. Detection of these pairs of photons enables the computation of sites of concentration of the isotope, which is incorporated into a metabolically significant compound. For example, [15O]water can indicate blood flow, and [18F]fluorodeoxyglucose is taken up by cells as if it were real glucose. The images of slices of the brain built by PET scanning are based on such functions as blood flow, uptake of a glucose analog, metabolism of a neurotransmitter precursor, or the binding of a labeled drug to receptors on the surfaces of cells. The anatomical resolution of PET (5-10 mm) is superior to that of a cortical blood flow or a SPECT scan, but it is inferior to that attainable with CT (2 mm) or MRI (0.5-1.0 mm).
Positron-emitting isotopes have half-lives ranging from 2 minutes for 15O to 2 hours for 18F, during which time they must be made, incorporated into suitable compounds, and administered to the patient. This technique can, therefore, be used only in hospitals equipped with a cyclotron and a laboratory for rapid radiochemical syntheses. The images obtained by PET, some of which display the distributions of neurons that use or respond to particular synaptic transmitters, can be more informative to physicians than the purely anatomical images obtained with CT or MRI.
Methods for Investigating Neural Pathways and Functions
Clinicopathological correlations and functional imaging techniques show which parts of the brain and spinal cord are used for particular purposes, but they do not provide much information about the ways in which neurons, with their long axons, communicate between different parts of the nervous system.
In histological material from normal animals, it is seldom possible to follow an individual axon from its cell body of origin to the distant site in which it terminates. The small diameters and curved trajectories of axons, together with the fact that different pathways commonly occupy the same territory, make the direct tracing of most connections impossible. It is, therefore, necessary to use experimental methods to determine the connectivity of the many groups of neurons in the brain and spinal cord. Results obtained in laboratory animals, especially cats and monkeys, may be applicable to human brains. This transfer of data from animals to humans is usually justifiable when there are no major differences between the connections found in diverse groups of animals. Some injuries and diseases in the human nervous system can cause degeneration of particular tracts of axons. Postmortem examination of the degenerated fibers provides valuable information about human neural connections.
NEUROANATOMICAL METHODS BASED ON DEGENERATION
Until the introduction of methods based on axoplasmic transport, fiber tracts were traced by staining fibers undergoing Wallerian degeneration (see Chapter 2) after the placement of a destructive lesion at a selected site in the central nervous system (CNS) of an animal. The oldest staining method for anterograde degeneration is the Marchi technique, which selectively stains degenerating myelin with osmium tetroxide in the presence of an oxidizing agent. The course of a tract can be followed in sections taken at appropriate intervals (Fig. 4-4). The archi technique does not show the unmyelinated terminal branches of the degenerating axons, but it is the only method that can nevertheless give useful results when applied to human postmortem material. Silver methods, which can show degenerating unmyelinated axons and synaptic terminals, were much used for laboratory animals until about 1975. These methods are not suited to the human nervous system because the degenerating axons are demonstrable
only during a critical period of 4 to 8 days after placement of a lesion. Degenerating axonal terminals can also be recognized in electron micrographs.
FIGURE 4-4 Section of the third cervical segment of a human spinal cord. The patient died 9 days after an injury that damaged the dorsal roots of the second, third, and fourth cervical nerves on the right side, together with the dorsal part of the right lateral funiculus in segment C2. The tissue was processed by the Marchi method, and degenerating myelin can be seen in entering fibers of the third right cervical dorsal root (A), branches of fibers derived from dorsal roots C3 and C4 in the lateral part of the dorsal funiculus(B), and descending corticospinal fibers in the lateral funiculus (C).
NEUROANATOMICAL METHODS BASED ON AXOPLASMIC TRANSPORT
Research methods based on degenerating axons were replaced in the 1970s by much more sensitive techniques that reveal both the cells of origin and the sites of termination of axons. In these procedures, a tracer substance is injected into a region of gray matter. The tracer is taken up by axonal terminals or neuronal cell bodies (or both) and transported within the cytoplasm. Retrograde tracers accumulate in the cell bodies of neurons whose axons end in the injected region. Anterograde tracers enter cell bodies and are moved into the presynaptic terminals at the destinations of the axons. A tracer may be a radioactively labeled amino acid, a fluorescent dye, a histochemically demonstrable enzyme (notably horseradish peroxidase [HRP]), or a protein that has been chemically linked to a fluorescent dye or HRP.
Some hydrophobic fluorescent compounds, most notably a cyanine dye called DiI, enter the lipid domains of cell membranes, including the neuronal axolemma, and then diffuse in the plane of the membrane. This happens even in dead tissue, allowing the tracing of fiber tracts from a site of application of the dye. Diffusion within the axolemma is slow: several months are needed to trace axons over distances of less than a centimeter. This method has been applied to human postmortem material but has not yet yielded new neuroanatomical knowledge.
TRANSSYNAPTIC TRACING OF PATHWAYS
Certain viruses are used for experimental neuronal tracing because they replicate within neurons, are transported within the axon, and are passed from one cell to another at synapses. These viruses can be modified to make the cells that harbor them synthesize a histochemically detectable enzyme, or the viral protein may be stained immunohistochemically. Transsynaptic
transfer of viruses occurs naturally in some diseases, notably rabies.
Metabolic Marking Methods
The sugar 2-deoxy-D-glucose, an analog of ordinary D-glucose, enters cells but is not metabolized. Consequently, radioactively labeled 2-deoxyglucose accumulates in the cytoplasm of metabolically active cells and may be detected there by autoradiography. The deoxyglucose method can reveal structures in the brain that are active when a particular system of pathways is in use. Thus, it may be possible to determine which of a multitude of connections demonstrated by neuroanatomical tracing methods are the most important in relation to function.
The catalytic functions of certain enzymes used in the metabolic activities of all cells can be demonstrated histochemically. Cytochrome oxidase is a notable example, and in regions that contain active neurons, the activity of this enzyme is higher than in adjacent quiescent areas. Cytochrome oxidase histochemistry has been used with great success in the demonstration of columns of cells that respond to different visual stimuli in the cortex of the occipital lobe of the brain (see Chapter 14).
Physiological and Pharmacological Methods
Neuroanatomical studies are often supplemented by electrically stimulating neurons and recording the potentials evoked elsewhere. Timing of the response may help to determine the number of neurons, or synaptic delays, that are included in a pathway. Neuroanatomical tracing and electrophysiological experiments are frequently combined with immunohistochemistry to identify neurotransmitters and to ascertain their actions on postsynaptic neurons. Electrophysiological investigations of the human CNS are necessarily more limited in scope than experiments using animals. Nevertheless, a great deal has been learned from observation of the effects of stimulating the cerebral cortex. Such studies are reviewed in Chapter 15.
Several toxic substances are used in laboratory animals as adjuncts to the study of neuroanatomy. For example, nicotine was used a century ago by Langley to block synapses and thus establish their locations in autonomic ganglia. Local injection of kainic acid or ibotenic acid kills many types of neurons without causing transection of passing fibers. These substances are known as excitotoxins because they are analogs of the excitatory transmitter glutamic acid. When an excitotoxin binds to glutamate receptors, there is an unusually long activation of nonspecific ligand-gated cation channels of the postsynaptic cells. Calcium ions diffuse into the neurons and activate proteolytic enzymes that destroy the cytoplasm. The resulting lesion is more selective than one produced by physical methods. Cells that use monoamines as synaptic transmitters are selectively intoxicated by analogs of these substances or their metabolic precursors. Thus, neurons that make use of dopamine or noradrenaline are selectively poisoned by 6-hydroxydopamine, and serotonin cells are similarly sensitive to 5,6-dihydroxytryptamine.
Some toxic lectins (e.g., ricin-60 from the castor bean) and other compounds (notably the antibiotic doxorubicin) are taken up by axonal endings and by injured axons of passage and transported retrogradely to the neuronal cell bodies, where they inhibit nucleic acid and protein synthesis. This strategy, known as suicide transport, produces selective lesions that can provide experimental models of diseases in which certain populations of neurons degenerate spontaneously.
DeYoe EA, Bandettini P, Neitz J, et al. Functional magnetic resonance imaging (FMRI) of the human brain. J Neurosci Methods 1994;54:171-187.
Frackowiak RSJ, ed. Human Brain Function, 2nd ed. Amsterdam: Elsevier, 2004.
Heimer L. Neuroanatomic Techniques. In Heimer L, ed. The Human Brain and Spinal Cord, 2nd ed. New York: Springer-Verlag, 1995:172-184.
Krassioukov AV, Bygrave MA, Puckett WR, et al. Human sympathetic preganglionic neurons and motoneurons retrogradely labelled with DiI. J Autonom Nerv Syst 1998; 70:123-128.
Lukas JR, Aigner M, Denk M, et al. Carbocyanine postmortem neuronal tracing: influence of different parameters on tracing distance and combination with immunocytochemistry.J Histochem Cytochem 1998;46: 901-910.
McLean JH, Shipley MT, Bernstein DI. Golgi-like transneuronal retrograde labelling with CNS injections of Herpes simplex virus type 1. Brain Res Bull 1989;22:867-881.
Purves D. Assessing some dynamic properties of the living nervous system. Q J Exp Physiol 1989;74:1089-1105.
Raichle ME. Functional brain imaging and human brain function. J Neurosci 2003;23:3959-3962.
Rajakumar N, Elisevich K, Flumerfelt BA. Biotinylated dextran: a versatile anterograde and retrograde neuronal tracer. Brain Res 1993;607:47-53.
Rao SM, Binder JR, Hammeke TA, et al. Somatotopic mapping of the human primary motor cortex with functional magnetic resonance imaging. Neurology 1995;45:919-924.
Ugurbil K, Toth L, Kim DS: How accurate is magnetic resonance imaging of brain function? Trends Neurosci 2003;26:108-114.
Vercelli A, Repici M, Garbossa D, et al. Recent techniques for tracing pathways in the central nervous system of developing and adult mammals. Brain Res Bull 2000;51:11-28.