Clinical Neuroanatomy, 27 ed.

CHAPTER 22. Imaging of the Brain

Brain imaging provides essential diagnostic information and is very useful for research on the brain. Images of the skull, the brain and its vessels, and spaces in the brain containing cerebrospinal fluid can aid immeasurably in the localization of lesions. In concert with physical examination and history, imaging studies can provide important clues to diagnosis, and often permit a definitive diagnosis. In emergency cases, images of unconscious patients may be the only diagnostic information available.

Computed tomography (CT), magnetic resonance imaging (MRI), and other similar imaging methods are usually displayed to show sections of the head, the sagittal, coronal (frontal), and horizontal (axial) planes are commonly used (Fig 22–1).


FIGURE 22–1 Planes used in modern imaging procedures.


Skull x-rays provide a simple method for imaging calcium and its distribution in and around the brain when more precise methods are unavailable. Plain films of the skull can be used to define the extent of a skull fracture and a possible depression or determine the presence of calcified brain lesions, foreign bodies, or tumors involving the skull. They can provide images of the bony structures and foramens at the base of the skull and of the sinuses. Skull x-ray films can also provide evidence for chronically increased intracranial pressure, accompanied by thinning of the dorsum sellae, and abnormalities in the size and shape of sella turcica, which suggest large pituitary tumors. Skull films are sometimes used to screen for metal objects before beginning MRI of the head.


Cerebral Angiography

Angiography (arteriography) of the head and neck is a neurodiagnostic procedure used when a vessel abnormality such as occlusion, malformation, or aneurysm is suspected (Figs 22–2 to 22–12; see also Chapter 12). Angiography can also be used to determine whether the position of the vessels in relation to intracranial structures is normal or pathologically changed. Aneurysms, arteriovenous fistulas, or vascular malformations can be treated by interventional angiography using balloons, a quickly coagulating solution that acts as a glue, or small, inert pellets that act like emboli.


FIGURE 22–2 Angiogram of the aortic arch and stem vessels. Normal image. 1: Brachiocephalic artery; 2: common carotid artery; 3: left subclavian artery; 4: right vertebral artery. (Reproduced, with permission, from Peele TL: The Neuroanatomical Basis for Clinical Neurology. Blakiston, 1954.)


FIGURE 22–3 Left internal carotid angiogram, early arterial phase, lateral view. Normal image (compare with Fig 22–4).


FIGURE 22–4 Schematic drawing of a normal angiogram of the internal carotid artery, arterial phase, lateral projection. The numbers refer to vessels shown in Figures 22–4 and 22–6. (Redrawn and reproduced, with permission, from List C, Burge M, Hodges L: Intracranial angiography. Radiology 1945;45:1.)


FIGURE 22–5 Left internal carotid angiogram, arterial phase, lateral view. Normal image (compare with Fig 22–6).


FIGURE 22–6 Schematic drawing of a normal angiogram of the internal carotid artery, arterial phase, frontal projection. (For significance of numbers, see Fig 22–4.) (Redrawn and reproduced, with permission, from List C, Burge M, Hodges L: Intracranial angiography. Radiology 1945;45:1.)


FIGURE 22–7 Right internal carotid angiogram, arterial phase, anteroposterior view. Abnormal image.


FIGURE 22–8 Vertebral angiogram, arterial phase, right lateral view. Normal image. Arrows indicate posterior choroidal arteries.


FIGURE 22–9 Vertebral angiogram, arterial phase, anteroposterior view, with head flexed (Towne position). An aneurysm is present, but the pattern of the vessels is normal.


FIGURE 22–10 Left internal carotid angiogram, venous phase, lateral view. Normal image. (Compare with Figs 12–9 and 22–11.)


FIGURE 22–11 Schematic drawing of normal venogram in lateral projection, obtained by carotid injection. Superficial veins are shaded more darkly than the sinuses and deep veins. (Redrawn and reproduced, with permission, from List C, Burge M, Hodges L: Intracranial angiography. Radiology 1945;45:1.)


FIGURE 22–12 Digital subtraction angiogram of the neck vessels, oblique anterior view. Open arrow shows small sclerotic plaque; closed arrow shows large plaque.

Angiograms consist of a series of x-ray films showing contrast material introduced into a major artery (eg, via a catheter in the femoral) under fluoroscopic guidance. Arterial-phase films are followed by capillary and venous-phase films (see Figs 22–6 to 22–10). Right and left internal carotid and vertebral angiograms may be complemented by other films (eg, by an external carotid series in cases of meningioma or arteriovenous malformation). The films are often presented as subtracted, that is, as reversal prints superimposed on a plain film of the skull.


CT, also called computed axial tomography (CAT), affords the possibility of inspecting cross sections of the skull, brain, ventricles, cisterns, large vessels, falx, and tentorium. Since its development in the 1960s, the CT scan has become a primary tool for demonstrating the presence of abnormal calcifications, brain edema, hydrocephalus, many types of tumors and cysts, hemorrhages, large aneurysms, vascular malformations, and other disorders.

CT scanning is noninvasive and fast. Although it has a high degree of sensitivity, its specificity is relatively limited. Correlation with the clinical history and physical examination is an absolute requirement. In the case of a subarachnoid hemorrhage, for example, although a CT scan may quickly localize the areas containing blood (Fig 12–21), additional CT images (Fig 12–22), magnetic resonance imaging, or angiography is often required to determine whether the cause was an aneurysm or an arteriovenous malformation.

The CT scanning apparatus rotates a narrow x-ray beam around the head. The quantity of x-ray absorbed in small volumes (voxels [volume elements, or units]) of brain, measuring approximately 0.5 mm2 × 1.5 mm or more in length, is computed. The amount of x-ray absorbed in any slice of the head can be thus determined and depicted in various ways as pixels (picture elements) in a matrix. In most cases, absorption is proportional to the density of the tissue. A converter translates the numeric value of each pixel to a gray scale. Black-and-white pictures of head slices are then displayed, with black representing low-density structures and white representing high-density structures. The thickness of the slices can vary, from 1.5 mm to 1 cm. The gray scale can also vary; although a setting at which brain tissue is distinguished best is commonly used, in some cases bone, fat, or air need to be defined in great detail.

A series of 10 to 20 scans, each reconstructing a slice of brain, is usually required for a complete study. The plane of these sections is the orbitomeatal plane, which is parallel to both Reid’s base plane and the intercommissural line used in stereotactic neurosurgery (Fig 22–13). Usually, a “scout view” similar to a lateral skull roentgenogram is taken with a CT scanner to align the planes of sections (Fig 22–14). With the modern technology now available, each scan takes only a few seconds. Examples of normal and abnormal CT scans are shown in Figures 22–15 and 22–16.


FIGURE 22–13 Schematic image of the zero horizontal and coronal planes. The line between anterior and posterior commissures parallels Reid’s base plane.


FIGURE 22–14 Lateral “scout view” used in CT procedure. Superimposed lines represent the levels of the images (sections). Line 1 is at the level of the foramen magnum; line 4 is at the level of the infraorbitomeatal plane.


FIGURE 22–15 CT image, with contrast enhancement, of a horizontal section at the level of the thalamus. Normal image. Compare with FIGURE 13–5.


FIGURE 22–16 Representative examples of CT images. (Courtesy of GP Ballweg.)

CT scanning of the posterior fossa may provide only limited information because of the many artifacts caused by dense bone. Images reformed by a computer from a series of thin sections allow visualization in any desired plane, for example, midsagittal (see Fig 6–15) or coronal. Coronal sections are often extremely useful for structures lying at the base of the brain, in the high convexity area, or close to the incisura. Detailed examination of orbital contents requires planes at right angles to the orbital axis.

Tissue density can change pathologically. Areas of hyperemia or freshly clotted hemorrhage appear more dense (Fig 12–19); edematous tissue appears less dense (Fig 12–14).


Magnetic resonance imaging (MRI) is widely used for noninvasive visualization of the brain and spinal cord. This imaging method depicts protons and neutrons in a strong external magnetic field shielded from extraneous radio signals; no radiation is used.

The spatial distribution of elements with an odd number of protons (such as hydrogen) within slices of the body or brain can be determined by their reaction to an external radio frequency signal; gradient coils are used to localize the signal (Fig 22–17). The signal of every voxel is shown as a pixel in a matrix, similar to the CT technique. The resolution of the images is comparable to, or exceeds, that of CT scans, and with MRI an image of the brain or spinal cord in any plane can be obtained directly; no reformation is required. Bone is poorly imaged and does not interfere with visualization of nervous tissue; thus, MRI is especially useful for imaging the spinal cord and structures within the posterior fossa.


FIGURE 22–17 Schematic representation of MR imager and its components. (Reproduced with permission from deGroot J: Correlative neuroanatomy of computed tomography and magnetic resonance imaging, 21st edition. Appleton & Lange, 1991.)

MRI can also be used to directly and noninvasively evaluate the flow of blood within medium and larger arteries and veins, with no need for intravenous injection of a contrast agent. This makes MRI particularly useful in cerebrovascular studies.

The sequence of radio frequency excitation followed by recording of tissue disturbation (echo signals) can be varied in both duration of excitation and sampling time. The images obtained with short time sequences differ from those obtained with longer time sequences (Fig 22–18). A normal MR image is shown in FIGURE 22–19; other MR images, both normal and abnormal, are found in Chapter 12 and elsewhere throughout this text.


FIGURE 22–18 MRI of horizontal sections through the lateral ventricles. Normal images. A: Image obtained with a short time sequence; the gray-white boundaries are poorly defined, and the spaces filled with cerebrospinal fluid are dark. B: Image obtained with an intermediate time sequence. C: Image obtained with a long time sequence; the white matter is clearly differentiated from gray matter, and the spaces filled with cerebrospinal fluid are white.


FIGURE 22–19 MRI of a horizontal high section through the head. Normal image.

Magnetic resonance angiography (MRA) uses a water proton signal to provide images of the cerebral arteries and veins. The method does not require the catheterization of vessels or the injection of radiopaque substances and is thus safer than traditional angiography.

The MRI process is relatively slow and takes more time than CT scanning. However, it provides high-quality images of the brain and spinal cord, and is safe for patients who have no ferromagnetic implants. The increasing sophistication of MRI technique (eg, with the use of contrast agents) has broadened its clinical usefulness. Although the distribution of water (hydrogen protons) is used for diagnostic purposes in patients at this time, experimental work with phosphorus, nitrogen, and sodium is under way. MRI provides a primary method of examination, especially in cases of suspected tumors, demyelination, and infarcts. As with CT scanning, successful use of MRI for accurate diagnosis requires correlating the results with the clinical history and physical examination.


In MRI, signals collected from water protons are used to assemble an image of the brain. The resonance signals collected by the computer using nuclear magnetic resonance can also be used to measure the levels of several dozen compounds that are present in the brain, including lactate, creatine, phosphocreatine, and glutamate. Magnetic resonance spectroscopy is routinely used as an experimental tool that provides a noninvasive means of measuring the levels of various molecules within the brain. Magnetic resonance spectroscopy can be used to study the human brain and may be useful in the diagnosis of various neurologic disorders and in studies on putative therapies for diseases affecting the nervous system.


By varying the magnetic field gradients and pulse sequences, it is possible to make MRI sensitive to the rate of diffusion of water within various parts of the brain; this is called diffusion-weighted imaging (DWI). DWI permits the visualization of regions of the brain that have become ischemic within minutes after the loss of blood flow (Fig 22–20).


FIGURE 22–20 Cerebral infract shown by diffusion-weighted imaging (DWI). On the left, a conventional MRI (T2- weighted image) 3 hours after stroke onset shows no lesions. On the right, DWI 3 hours after stroke onset shows extensive hyperintensity indicative of acute ischemic injury. (Reproduced, with permission, from Warach S, et al: Acute human stroke studied by whole brain echo planar diffusion weighted MRI. Ann Neurol 1995;37:231.)


Neuroanatomy is being revolutionised by functional MRI, which uses MRI pulse sequences that can also be adjusted to produce an image sensitive to local changes in the concentration of deoxyhemoglobin. When neural activity occurs in a particular region of the brain, there is usually an increase in oxygen uptake, which triggers increases in cerebral blood flow and cerebral blood volume. These increases lead to a local reduction in the concentration of deoxyhemoglobin. Changes in deoxyhemoglobin concentration are thus related to the level of neural activity within each part of the brain. By measuring deoxyhemoglobin levels and comparing them when the brain is in a resting state and when it is involved in a particular activity, functional MRI (fMRI) can provide maps that show regions of increased neural activity within the brain. fMRI can be used, for example, to detect changes in brain activity associated with motor activity (eg, tapping of the fingers), sensory activity (ie, stimulation of a sensory organ or part of the body surface), cognitive activity (eg, calculation, reading, or recalling), and affective activity (eg, responding mentally to a fearful stimulus). Examples are shown in Figures 15–17 and 22–21.


FIGURE 22–21 Example of functional MRI showing increased perfusion of the motor cortex in the right hemisphere associated with rapid finger tapping of the left hand. Upper left: Relative blood flow map in the transverse (horizontal) plane during rest. Upper right: Relative blood flow map during rapid tapping of fingers of the left hand (the arrows point to the region in the right hemisphere where the increased blood flow response is seen). Lower right: Subtracting one image from the other provides a “difference image,” which shows a hot spot corresponding to the active region of the cortex. Center:Difference image superimposed on structural MRI, showing that increased perfusion maps precisely to the motor cortex in the anterior bank of the central sulcus of the right hemisphere. (Courtesy of Dr. S. Warach.)


Positron emission tomography (PET) scanning has become a major clinical research tool for the imaging of cerebral blood flow, brain metabolism, and other chemical processes (Fig 22–22). Radioisotopes are prepared in a cyclotron and are inhaled or injected. Emissions are measured with a gamma-ray detector system. It is possible, for example, to map regional glucose metabolism in the brain using fluorine 18 (18F)-labeled deoxyglucose. Images that show focal increases in cerebral blood flow or brain metabolism provide useful information about the parts of the brain that are activated during various tasks. This is another example of functional brain imaging.


FIGURE 22–22 PET scan of a horizontal section at the level of the lateral ventricles. The various shades of gray indicate different levels of glucose utilization.

It is also possible, using PET scanning, to localize radioactively tagged molecules that bind specifically to certain types of neurons. Using this type of technique, it is possible, for example, to localize dopaminergic neurons in the brain and to quantitate the size of the nuclei containing these neurons.

One disadvantage of PET scanning is the lack of detailed resolution; another is that most positron-emitting radioisotopes decay so rapidly that their transportation from the cyclotron (the site of production) can be rate limiting. Some isotopes, such as 18F and gamma-aminobutyric acid, have a sufficiently long half-life that they can be shipped by air. Some, such as ruthenium derivatives, can be made at the site of examination.


Advances in nuclear medicine instrumentation and radiopharmaceuticals have opened renewed interest in single photon emission CT (SPECT) of the brain. The increasing use of investigative agents in conjunction with PET imaging has stimulated the development of diagnostic radiopharmaceuticals for SPECT; these are routinely available to clinical nuclear medicine laboratories. A technetium 99m (Tc-99m)-based compound—Tc-99m-hexamethylpropyleneamineoxime (Tc-99m-HMPAO)—is widely used. It is sufficiently lipophilic to diffuse readily across the blood-brain barrier and into nerve cells along the blood flow. It remains in brain tissue long enough to permit assessment of the relative distribution of brain blood by SPECT in 1.0- to 1.5-cm coronal, sagittal, and horizontal tomographic slices. SPECT studies are especially useful in patients with cerebrovascular disease (Fig 22–23).


FIGURE 22–23 SPECT image of a horizontal section through the head at the level of the temporal lobe. An infarct (arrow) is shown as an interruption of the cortical ribbon. (Courtesy of D. Price.)


Cabeza R, Kingstone A (editors): Handbook of Functional Neuroimaging of Cognition. MIT Press, 2006.

Chugani H: Metabolic imaging: A window on brain development and plasticity. Neuroscientist 1999;5:29.

Damasio H: Human Brain Anatomy in Computerized Images. Oxford Univ Press, 1995.

Detre JA, Floyd TF: Functional MRI and its application in clinical neuroscience. Neuroscientist 2002;7:64.

Greenberg JO: Neuroimaging. McGraw-Hill, 1999.

Kaplan RT, Atlas SW: Pocket Atlas of Cranial Magnetic Resonance Imaging. Lippincott Williams & Wilkins, 2001.

Mills CM, deGroot J, Posin JP: Magnetic Resonance Imaging: Atlas of the Head, Neck, and Spine. Lea & Febiger, 1988.

Oldendorf WH: The Quest for an Image of the Brain. Raven, 1980.

Osborn AG: Introduction to Cerebral Angiography. Harper & Row, 1980.

Senda M, Kimura Y, Herscovitch P (editors): Brain Imaging using PET. Elsevier, 2002.

Tamraz JC, Comair Y, Luders HO: Atlas of Regional Anatomy of the Brain using MRI. Oxford, 2000.

Toga A, Mazziotta J (editors): Brain Mapping: The Systems. Elsevier, 2000.

Toga A, Mazziotta J, Frackowiak R: Brain Mapping: The Disorders. Elsevier, 2000.

Truwit CL, Lempert TE: High Resolution Atlas of Cranial Neuroanatomy. Williams & Wilkins, 1995.

Warach S: Seeing the brain so we can save it: Magnetic resonance imaging as a clinical tool. In: From Neuroscience to Neurology, Waxman SG (editor). Elsevier Academic, 2005.

If you find an error or have any questions, please email us at Thank you!