From StudyingMed

Jump to: navigation, search

This learning activity is not yet finished -- we still have to improve it to reach our stringent standards. Please help out!

This article needs its relevant images to be added. Please upload an image and include it!

Lecturer afk. See Soon's notes.

Official Notes (from 2010)

Learning Objectives

This lecture will cover the major imaging techniques currently in use in neuroradiology. It will outline the physical basis of each technique, including current limitations. Selected examples of each approach will be given. It is important to understand the uses and limitations of the available approaches in order to understand the choices that must be made in selecting one or combinations of these technologies. These choices are dependent on the information sought for diagnosis.

This is a large and complex topic and it is patent that further reading and study is necessary if any in depth understanding is to be achieved. This lecture serves as an introduction.

The techniques covered include:

  • Ultrasound
  • CT (X‐ray computed tomography)
  • PET (positron emission tomography)
  • MRI & MRS Magnetic Resonance Imaging and Spectroscopy


In physics, the term “ultrasound” refers to acoustic frequencies greater than 20 kHz (20,000 Hz). The terminology refers to the fact that frequencies in this range are not generally able to be detected by the human ear, although some animals, such as bats can hear frequencies up to 100 kHz.

Diagnostic medical ultrasonography typically uses frequencies in the range of 2‐18 MHz. The choice of frequency is a compromise between spatial resolution (higher frequencies (7‐18 MHz) give higher resolution) and tissue penetration (depth), with lower frequencies (2‐6 MHz) penetrating further into tissue.

Ultrasound technology developed out of SONAR, RADAR and the ultrasonic metal flaw detector constructed to test the hulls of ships during WWII. I refer those interested in the history of it to http://www.ob‐

The technique is mostly used to image soft tissue. Ultrasound frequencies are used for imaging because they can be focused into small, well‐defined beams that can both probe into tissue and interact with that tissue.

Most modern ultrasound machines use pulsed sound waves which are generated using the “transducer” of the machine. These contain piezoelectric elements. Piezoelectric elements have the property that they vibrate at a specific frequency when an electrical current is applied to them. Therefore, to change the frequency of the ultrasound device, you must change the transducer.

When pulses of sound are applied to the body two things may happen. One is that the sound dissipates within the tissue and is lost. The other is that the sound waves are reflected by internal body structures and send echoes back to the surface, where they can be detected by the transducer.

There are several properties of the reflected sound waves which can be used to determine information about the tissue through which they have travelled. These include:

1)Velocity: The velocity of sound waves is not determined by their frequency but by the medium through which they have travelled, according to the relationship Velocity = √E/ρ

Where E is a factor related to the stiffness or elasticity of the tissue, and ρ is the density of the tissue.

Approximate velocity of sound in various tissue media

Medium Velocity (m/s)
Fat 1450
Water 1480
Soft Tissue 1540
Bone 4100

2)Amplitude: in ultrasound refers to the range of pressure excursions experienced by the tissue. It also relates to the degree of tissue displacement caused by the pulse. This also leads us to attenuation of the pulse, or reduction of its amplitude due to absorption of its energy.

Material Attenuation coeff (dB/cn MHz)
Fat 0.66
Water 0.002
Soft Tissue 0.9
Bone 20
Muscle 2.0

From this is can be seen that lung tissue and bone, for example, are not particularly suited to ultrasound.

The various properties of the ultrasound echo are then used to reconstruct the image. Ultrasound machines can be operated in several different modes to give either simple echo information, two dimensional pictures, motion pictures or Doppler information. Doppler is particularly useful in CNS imaging for the detection of vasospasm and other vascular information.

Ultrasound is particularly useful for foetal and pediatric imaging; neural tube defects and CNS malformations can be diagnosed non‐invasively in utero. It also has some important applications in adult CNS diagnosis.


  • Good for delineating interfaces between solid and fluid‐filled spaces.
  • Renders “live” images, so good for ultrasound‐guided interventions and rapid diagnoses .
  • No known long‐term side effects, rarely causes any discomfort to patient.
  • Cheap, flexible, portable, widely available.


  • Has trouble penetrating bone making sonography of adult brain *very limited.
  • Trouble penetrating layers of gas (problems in intestine and lung)
  • Difficulties with depth penetration (e.g. in obese people)
  • Operator dependent (needs skill to be done well)
  • No scout images, so no way to tell exactly which part of body was imaged

X‐ray Computed Tomography (CT)

CT scanning, or CAT scanning as it is sometimes known, uses X‐rays to create a sliced image. The detected x‐rays are reconstructed by a computer to form the image. The word tomography is derived from the Greek tomos (part, or slice) and graphein (to write). Note that computed tomography could refer to any reconstructional technique; strictly speaking then CT scanning in this medical diagnostic sense refers to X‐ray CT.

CT uses a computer and a rotating x‐ray device to create detailed, cross‐sectional images, or slices, of organs and body parts. The CT scanner itself is a circular, rotating frame with an x‐ray tube mounted on one side and a banana‐shaped detector mounted on the other. A fan‐shaped beam of x‐rays is created as the rotating frame spins the x‐ray tube and detector around the patient. For each complete rotation, one cross‐sectional slice of the body is acquired.

CT was invented By Sir Godfrey Hounsfield, based on heoretical work by Allan McLeod Cormack. The two shared the Nobel prize for Medicine in 1979.

CT is mostly used in the head to image infarcts, tumours, bone trauma, calcification and haemorrhages.


  • Improvement over more traditional x‐rays – no interference from foreground structures out of the field of view.
  • Fast


  • Improvements in CT scanning have actually increased total radiation dose, both to individual scans and in the general population since the uptake of CT technology has taken off.
  • Needs radiation.
  • Spatial resolution has improved but not as good as MRI

Positron Emission Tomography (PET)

PET requires introduction into the body of a radionuclide which emits positrons (anti‐electrons). The radionuclide is incorporated into a biologically active molecule which then makes its way into the brain following injection. There it decays, emitting a positron. This positron will travel a short distance (< 1mm depending on the nuclide) before losing enough energy that it can interact with an electron. This interaction results in the annihilation of both the electron and the positron, causing emission of a pair of gamma rays which travel in approximately the opposite direction. These are detected by gamma cameras placed around the head (PET cameras). From this, the spatial position of the annihilation reaction can be calculated.

PET radionuclides are typically nuclei with very short half‐lives. Consequently they need to be made using a cyclotron, a device which accelerates charged particles in a spiralling pathway with ever increasing radius until they hit their target. Generally this collision creates secondary particles, such as the radionuclides of interest for PET.

PET is of particular use in neuroimaging. Its particular strength lies in its ability to target particular molecules. This means that in addition to its use in studying blood flow and metabolism (with 18FDG or with H2 18O), it can also be used to study binding of neurotransmitters on a receptor‐specific basis, or to identify the presence of molecules such as the Aβ fibrillar protein, found deposited in Alzheimer disease plaques (so‐called PiB‐PET imaging, using the PiB ligand).

PET is often combined with CT or with MRI to improve the spatial resolution and to provide a brain against which to co‐register the PET scans. When combined with CT the total radiation dose can be as high as 27 mSv (in a 70 kg person) which is high enough to require proper justification for the risk vs benefit.

If you are interested the safety aspects of radiation exposure, a good recent review of the topic is Wall et al. (2006) What are the risks from medical X‐rays and other low dose radiation? British Journal of Radiology 79, 285‐294.


  • Can use it to study certain neurotransmitters which can’t be measured any other way.
  • Chemically sensitive


  • Involves administration of (sometimes very) expensive radioactive ligands, with short half‐lives.
  • Limited ability to restudy same person in short period of time, and not generally applicable to children unless benefit outweighs risk.
  • Poor spatial resolution compared to, say, MRI
  • Data not available in real time, due to computer reconstruction although this is improving rapidly

Magnetic resonance imaging

This relies on the principle that certain nuclei possess a property known as “spin”, which causes them to behave like little unbalanced bar magnets. As a result, when they are placed in a magnetic field, they align themselves with respect to that field and precess in that field. Under these circumstances, if one bombards them with a radio signal at the right frequency, a small percentage of them will absorb some of this radio frequency energy and, in relaxing back to their former energy state, they will emit radio waves of their own. The frequency of the radio waves they emit is a property of their immediate chemical environment.

This emission of radio waves by nuclei with “spin” in a magnetic field is the underlying principal behind nuclear magnetic resonance imaging (MRI) and nuclear magnetic resonance spectroscopy (MRS). [When referring to the human applications of imaging and spectroscopy, the word “nuclear” is routinely dropped due to the apparent tendency of the general public to associate the word “nuclear” with radioactivity.]

MRI produces a spatial map of a tissue. As mentioned earlier, the emitted radiofrequency of a particular nucleus is dependent on its immediate chemical environment. It is also dependent upon the strength of the local magnetic field.

[This is the Larmor equation which relates the emitted frequency , to the strength of the magnetic field B through a constant , known as the magnetogyric ratio.  is constant for any given nucleus, so if you look at a simplified version of the Larmor equation, viz   B (1) you will see immediately that  will vary with B.]

There are many different applications of magnetic resonance imaging, which is a powerful technique. Most human CNS imaging is done either at magnetic field strength of 1.5 Tesla, or, more increasingly, 3 Tesla. There are also at least 30 7 Tesla MRI systems in the world which can study humans, although their use is currently confined to research. 3T is currently the highest field licenced by the FDA for diagnostic purposes.

Structural Magnetic Resonance imaging

In MRI we tend only to look at one particular chemical species. The most common one in MRI is water. Water is present in the brain at about 80 M, so there is plenty of it. Occasionally the CH3/CH2 groups in fat are also used. To get an MRI image of someone’s head, we put them in the magnet then, simply, apply what is known as a “field gradient”. In other words, we vary the magnetic field across the person’s head in a constant fashion. This has the effect of making the water molecules in one part of their head emit a different frequency to water molecules in another part of their head. We can then reconstruct this map of frequencies from all the water molecules to make a picture.

So spatial information can be obtained from the frequency of oscillation of the water molecule (as well as from its phase). Contrast in the image, on the other hand, is obtained from properties of the nuclei themselves including how many nuclei there are (proton density) and properties related to the manner in which the nuclei relax from their excited state (so called T1 or T2 relaxation). These properties, which can be highlighted in each imaging “sequence” depending on the way in which the scanner acquires the information.

Diffusion magnetic resonance imaging

We have seen that the local field to which a proton is exposed will determine the frequency of its resonance. We also know that water molecules will move around by diffusion, much the same way that dye diffuses through a glass of water. If a water molecule moves between the time it is excited by an externally applied radiofrequency pulse and the time we open the receiver and “listen” to the decay frequency, this will mean that the signal from all the water molecules in that particular region of the MRI image will be comprised of different frequencies, rather than the single frequency we would expect if the molecules had NOT moved around. This leads to “dephasing” of the signal, such that the signal is less than we would expect.

There are a number of different imaging approaches which take advantage of this phenomenon. Diffusion‐weighted imaging, which is often used to localize or diagnose stroke in the acute phase, takes advantage of the fact that the apparent diffusion constant of water can change within seconds of an infarct occurring. This is well before the infarct will be apparent on normal T2‐weighted imaging (around 24h) or on CT (up to a week).

Diffusion tensor (or spectrum) imaging uses the restricted diffusion of water by tissue structures to make maps of white matter tracts in the brain. Water is more likely to diffuse along an axon than across an axon as it is restricted from doing so by the axon membrane. This technique is extremely powerful. It is used heavily in neurosurgery to determine the boundaries of tumour resection and to determine which areas of the brain are connected with one another.

Magnetic Resonance Angiography relies on the flow of the water molecules in blood to make maps of the major arteries. It can be used to detect aneurysms and bleeds. A quantitative version of this, arterial spin labeling (ASL), can be used to measure blood flow, blood volume and mean arterial transit times.

Functional magnetic resonance imaging (fMRI)

This approach relies on the fact that blood flow increases to an area of the brain when it is activated. The increased blood flow results in a change in the ratio of oxy‐ to deoxy‐haemoglobin. Haemoglobin contains iron (four molecules of iron per molecule of haemoglobin). The oxidation state of this iron differs depending on whether or not the molecule is also carrying oxygen.

Deoxyhaemoglobin is in the paramagnetic state = Fe3+, while Oxyhaemoglobin is in the diamagnetic, or Fe2+ state.

Paramagnetic iron interferes more with the local magnetic field than does diamagnetic iron. This leads to a change in the intensity of the local water signal. The main functional imaging signal results from the increased blood flow that occurs in an area 3‐4s following activation. This leads to increased oxyhaemoglobin in the area and hence an increase in the local water signal. It is this increase that is detected in the fMRI paradigm.

Diagnostically, fMRI is sometimes used to replace the Wada test for determining on which side of the brain language is localized prior to surgery for epileptic seizure ablation. It is also a heavily used research tool for study of the brain.


It is non‐invasive (where contrast agents are not used). It is extremely safe if operated within guidelines It can be applied to the same patient repetitively It has excellent spatial resolution There are many different investigations that can be done on the same scanner.


It is not a sensitive technique (like PET). SAFTETY: People with metal implants can often not be scanned. This includes pacemakers, devices such as cochlea implants and some medical devices such as stents and other instruments, depending on the type and make. Persons with any metal on their person, such as shrapnel, metal shards from metal working etc, are also at risk and must not be scanned. There is also the possibility of radiofrequency heating of the tissue, although most scanners have strict safety limits on this. It is also possible, with fast switching of magnetic field gradients, to induce peripheral nerve stimulation in some subjects.