This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
It is not long that the physiological and biochemical processes in the living organism are able to be observed through modern technology. This possibility opens up fascinating possibilities in research of the human body, as well as in daily diagnosis. These techniques are usually ones such as PET, fMRI, MRI, CT, or EEG.
This essay explores the way in which physiological processes that underpin two techniques out of the many used. These are PET and fMRI due to their intensity, popularity, and importance of usage for a long time (Huang, Phelps, Hoffman, Sideris, Selin, Kuhl, 1980). The essay will also explore what determines the temporal and spatial resolution of these two methods.
The first PET scanner working on principles of tomography was first described by Ter-Pogossian with his colleagues in 1975 (Ter-Pogossian, Phelphs, Hoffman, et al., 1975). An FDG PET scan was first described by Reivich, Kuhl, Wolf, and others, four years later, in 1979. PET (positron emission tomography) as it is today, enables tomographic imaging of the dynamic region of the radioactivity distributed in the body (Sokoloff, 1985; Ter-Pogossian, 1985). Acuisition of an image takes about 20-40 minutes. PET is categorized as a non-invasive method, even though a radioactive tracer has to be injected in the human body scanned. An atom in a molecule at the place of interest in vivo is changed by an isotope, which is detectable from outside the body (Saha, 1992). The tracer is usually chemically alike with the original atom. These are called positrons. They have equal mass as the electron, but with a positive electrical charge. This ensures that any biochemical processes do not change. Usually the type of molecules which are of interest mostly consist of carbon (C), oxygen (O), and nitrogen (N) which are all of low concentration (hence only observable by PET). A certain type of tracer is injected, depending on the molecule of interest. A carbon atom is replaced with a 11C, an oxygen is with 15O nitrogen atom is replaced with 13N. These radioisotopes have a short half life, and are small enough to pass through cell membrane. After they are emmited, from atom nuclei, they lose their kinetic energy and combine with an electron. For example, to measure regional cerebral blood flow by PET, 15O-labelled water is used. On the other hand, aminoacids, precursors of neurotransmitters, 18F-thymidine which can be built into DNA or enzyme substrates of pharmaceuticals can be used for PET imaging (Kopecek, 2001; Humes, 2005).
PET scanners can detect two co-incidental electromagnetic photos with the energy of 511 keV (kiloelectronvolt). These result from the annihilation process (conversion of mass into energy) of a positron with a nearby occurring electron, such as those mentioned above. If the two photons with a total energy of 1.02 MeV are detected simultaneously or almost simultaneously, and in opposite directions by a dual-headed gamma-camera, it is assumed that they both can be attributed to annihilation at a specific location, along the line of their directions. The travel distance of the free emitted positron in the matter is extremely short (a few mm), the location of annihilation practically equals the location of the site of the radioactive nucleus (Richardson, 2010; Marsden et al., 2000).
18F-labelled deoxyglucose (FDG) is used the most in clinical, as well as research used of PET. It is an analogue of glucose that is involved in a surplus of biochemical processes (Marsden et al., 2000; Kim et al., 2010; Ferrari et al., 2010). An analogue alternative can be used, for example glucose, 2-[18F]fluoro-2-deoxy-D-glucose (Carlson et al., _date_ ;Watabe et al., _date_). Such a molecule, however, must still be able to enter into the physiological processes and follow them in a way which is identical to the original molecule, such as being taken up into cells by glucose transporters (Miller et al. 2004).
In the cell of the brain, glucose metabolism provides approximately 95% of the ATP (adenosine triphosphate), which is required for brain glucose transport. ATP is used in various amounts in various regions, dependednt on momentary need. FDG is a good indicator to assess the ATP-dependent function, such as glucose in-take of individual structures the brain (Phelps, 2000). This is done by usine an FDG model (Huang et al., 1980). FDG is, therefore, widely used in encology imaging (Reivich, 1985 ; Phelps, 2000; Miller et al, 2004).
The neurophysiology of the brain, used by PET, looks at brain region perfusion and regional metabolism of glucose in the brain. Neuro-chemical PET analysis allows the measurement of the density of neuroreceptors, receptor binding after a pharmacological intervention, activity of endogenous neurotransmitters, or enzymatic activity (Kopecek, 2001).
The energy metabolism of the adult human brain is almost completely dependent on glucose. Glucose is mainly used by synapses which trigger neuronal activity. There is, normally, a correlation between the cerebral blood flow and metabolism. In the regions where there is a high density of synapses , such as in the cortex, perfusion as well as metabolism are naturally higher than, for example in the white matter (Schaller, 2005; Jueptner, 1995).
Active neurons consume the largest part of energy in the synaptic region (mainly in pre-synaptic processes), and this is also where the increase of metabolic rate and increase in glucose consumption occurs. With the increasing metabolism rate, perfusion increases, as well as response to the increasing energetic demand. This indicatres that the glucose metabolic rate gives a direct information, while perfusion data gives a indirect information on the rate of neurotransmission. Perfusion or glucose metabolism, in that case, is an indicator of the afferent synapse activity (Kopecek, 2001; Jueptner, 1995). The measurement includes also the component of perfusion (Miller et al., 2004; Phelps, 2000; Humes, 2005; Novak, 2009). The calculation of biological proceses is called pharmacokinetic modeling (Burger, Buck, 1997).
It has been recvealed that , where the FDG accumulation is high, the glucose uptake and metabolism is high too. The glucose metabolism is closely linked to physiological activity of tissues including individual structures within the brain.
FDG is nowadays most often used in PET imaging for practical reasons, such as a short half-life of 15O. PET with FDG can also cover PET perfusion studies using 15O-labelled H2O because many pathological conditions as well as activation processes in the CNS influence both glucose metabolism and oxygen utilisation. PET also is able to measure FDG retention per volume of tissue (Marsden, 2000; Novak, 2009).
Neurophysiological methods can be divided into rest and activation studies (Kopecek, 2001).
PET scans have spatial resolution (an ability of an image-forming device to distinguish small details of an object) and temporal resolution (the precision of a measurement to time) points, which should be mentioned.
Spatial resolution of PET is given by the resolution capacities of the gamma camera and the detectors, as well as influences from the source of 511 keV photons (the tissues of the patient).
Resolution of the devices is mostly limited by the dimensions and material of the detectors by their distance, and by the post-processing capabilities that process the detected photons (e.g. elimination of partial-volume effect etc.). Scintillation detectors have a spatial resolution of around 1 mm.
In addition to that, there are other components which come from the tissues which may affect spatial resolution. One of these is, for example the distance that a positron is able to move before it collides with an electron. This is effect the spatial resolution. Patients body mass may also cause a decrease in spatial resolution, as well as false coincidence coming from different places in the tissue. Also low signal-to-noise ratio and relatively poor statistics in the activation images may adversely affect the spatial resolution (Grabowski et al., 1996).Another occurrence where spatial resolution may decrease is the patient moving during the image acquisition.
There is debate between what is the resulting spatial resolution of PET. Kopecek (2001) claims it is between 3-6mm, while other sources claim it is somewhere between 1-10mm (Novak, 2009). Grabowski (1996), and Jueptner (1995), however point out that spatial resolution of PET is better than that of SPECT but much worse that that of fMRI.
In case of measurements of biological processes, resulting temporal resolution consists of the temporal resolution given by the speed of biological processes that lead to a measurable signal, as well as the temporal resolution of the measuring device itself. In PET, both these components play a role.
Temporal resolution of dual-head gamma camera and the detectors is very, but in principle depends on the width of the coincidence window. Gamma cameras, however, must sample and join together all scintillations. It takes over 30-60 seconds to get an acceptable count statistics. (A trade-off between the amount of radioactivity administered to the patient, and the desired image quality). The slowness puts measures outside the range of real-time detectability (Frackowiak et al.. 1994; Grabowski et al., 1996).
In contrast to spatial resolution, the effective temporal resolution of PET is much lower (45-60 seconds) which is too low to investigate biological neuronal mechanisms of the brain in real-time (Dhond, 2007; Kopecek, 2001; Grabowski et al., 1996). This, however, does not mean that cognitive processes cannot be studied indirectly, by experimental PET designs of repetitive performance of a task (Grabowski et al. 1996). In fact it is done very well.
In FDG studies, temporal resolution is influenced by the speed of biological pathway of. The adjustments of regional cerebral blood flow to activation or deactivation of the neurons occur within seconds. This delay contributes to the decrease in temporal resolution of FDG PET studies (Jueptner et al., 1995).
There is, however, even a stronger factor affecting temporal resolution in PET studies using FDG. FDG cannot be metabolised, neither can be released from the cell readily. It takes some time for it to decay within the cell, so once accumulated within the cell, its presence will block the detection of the further physiological processes, such as neuronal inhibition, which also increases glucose and FDG uptake (Jueptner 1995). It will do this at least until its activity is decreased by the natural positron-decay of 18F.
The other technique which is used greatly, mentioned in the introduction, is functional magnetic resonance imaging. It produces images of brain activity, and even maps of neuronal activation in different locations of the human brain. In cognitive neuroscience research field, various fields of brain activity can be studied by fMRI, such as vision (Tootell et al., 1998), emotion (Phillips et al., 1997), language (Binder, 1997), or for example memory (Fletcher et al., 1997). In addition, fMRI has its role in clinical applications such as in stroke, brain tumours and various psychiatric disorders (Howseman et al., 1999).
The first fMRI experiments were done using a gadolinium (Bellieau et al., 1991). It was soon discovered, that another excellent endogenous contrast called deoxyhaemoglobin is much better. Measurement of blood oxygenation level (BOLD) eventually proved to be a very sensitive indirect marker of neuronal activity (Howseman et al.,1999; Sutton et al., 2009) and is now the most commonly used method (Ogawa, Lee, Kay, & Tank, 1990; Turner, Bihan, Moonen, Desper, & Frank, 1991; Bandetti, Wong, Hinks, Tikofsky, & Hyde, 1992).
The source of the fMRI contrast using the BOLD technique is changes in T*2 relaxation due to changes in deoxyhaemoglobin. Deoxyhaemoglobin is paramagnetic and therefore influences the local proton relaxation by creating microscopic inhomogeneities of the magnetic field. So, by optimisation the signal dependence on deoxyhaemoglobin concentration it is possible to image blood oxygenation levels and its changes quite rapidly (Sutton et al., 2009; Tieleman et al., 2009; Howseman et al., 1999).
The BOLD signal changes indirectly measure the neuronal activity in the region very close to where the is coming from. The relationship between the signal and the underlying neuronal activity is very complex.
Any increase in the activity of neurons (especially at synapses) means increased needs of energy, which is obtained from glucose by oxidative phosphorylation. The beginning of neuronal activation, especially of the processes at synapses, is followed by increase in oxygen utilisation and also by increased blood flow. This means there are two main components that will underpin the changes in the BOLD signal. There is, however a phenomenon of dissociation in the intensity of the tissue oxygen utilisation and the perfusion: When the local synaptic activity increases both of them increase but the earlier increases much less than the latter. Therefore, the blood from the active brain tissue eventually contains relatively more oxyhaemoglobin and relatively less deoxyhaemoglobin than in the parts of the brain that are not activated. Oxyhaemoglobin is diamagnetic. It influences the local proton relaxation by creating microscopic inhomogeneities of the magnetic field. These can be detected by the T*2 relaxation changes.
Since deoxyhaemoglobin itself increases the speed of T*2 relaxation (and decreases T*2 time) the relative lack of it in the venous blood will lead to less increase in the T*2 relaxation rate and to relative increase in the T*2 time. This leads to an increase in the BOLD signal, as compared to the other parts of the brain. The increase in the BOLD signal is quite small ââ‚¬" on the range of <1% up to several %ââ‚¬â„¢s. In the activated regions of the brain first a short increase in deoxyhaemoglobin occurs which is replaced by immediate increase in oxygenated haemoglobin because of the increase in local perfusion (Sutton et al., 2009; Tieleman et al., 2009; Hlustik et al., 2008; Chlebus et al., 2005; .Bandettini, 2009; Howseman et al.,1999).
The haemodynamic response function can be divided into several phases (Chlebus et al., 2005, Howseman et al. 1999, Bandettini 2009).
The first phase is the initial dip of the BOLD signal, which lasts about 1 second. It corresponds to the original increase in the usage of oxygen and is in very close relationship with neuronal activity of the location. However, this change is every minute and fMRI with 1.5T is not able to detect it reliably and is more distinct when using high-field scanner.
The second phase is the increase in the BOLD signal (2-5%). This matches to the increase in the flow of oxygenated blood by approximately 50-70%. This actually is the result of the BOLD measurement. The maximal change in BOLD signal occurs 4-9 seconds after the start of the stimulation, after that the intensity of the signal remains almost unchanged during the lasting stimulation (plateau).
The third stage of the HRF is the decrease in the BOLD signal after the end of the stimulus. The synaptic activity as well as the perfusion decreases back to normal. A post-stimulation undershoot follows due to unknown reasons and then the fifth stage occurs, which is the returning to normal phase. It occurs about 20 s after the end of stimulation.
The haemodynamic response to the neuronal activation can vary in timing between different regions of the brain but in principle is similar across the regions.
The spatial and temporal resolution of MRI is better than in PET by more than an order of scale (Lelyveld et al., 2010). MRI has a wider spectrum of contrast mechanisms which PET does not.
FMRI has relatively high spatial resolution, 3 mm or less (Bassett et al., 2009; Howseman, 1999; Chlebus et al., 2005). It is influenced by several things.
Firstly, it is influenced by voxel size. In principle, the bigger voxel the worse the spatial resolution is. Too small voxels, however, would lead to too low signal per voxel and insufficient measurement statistics. The optimal size of the voxel is, therefore, a swap between the two. It is dependent on the region measured, the strength of the magnetic field, the sequences used and the nature of the measured process (Huettel et al., 2009).
The main issue with voxels, however, which has an affect on the main spatial as well as temporal resolution, is that a usual measured voxel (typically 3x3x3 mm = 9 mm3) contains a variety of neurons, glial cells, as well as various vessels with distribution, such as capillaries, venules and larger draining veins. Hence, the signal change measured will only reflect a sum of all the micro-changes, especially those in oxygenation of the blood in different vessels contained in the measured voxel. This menas, if a larger draining vein is captured within or near the voxel, even much more significant influences may disturb the preciseness of the measurement. This is called venous weighting. It may cause a loss of specificity and significant reduce spatial resolution in the BOLD image (Huettel et al., 2009; Howseman et al.,1999; Sutton et al., 2009; Bandettini, 2009).
The fMRIââ‚¬â„¢s spatial resolution can also be effected by field strength. The changes in T*2 relaxation do not come only from the intra-vascular space where the changes in oxygenation occur. Depending on the strength of the magnetic field, also protons in the nearby extravascular spaces are affected (Howseman et al.,1999).
Another influence can be the signal-to-noise ratio. This requires repletion of tasks (Grabowski et al., 1996; Chlebus 2005).
The detectable temporal resolution of fMRI in human brain is in the range of 3-5 seconds in the one region (Howseman et al.,1999; Bassett et al., 2009). In comparison with EEG or MEG, the time resolution of fMRI is considerably worse. (Chlebus, 2005).
The main limiting factor, as explained above, is the disproportion between the rapid course of the neuronal activation processes and the much slower haemodynamic response, which, in addition follows the neuronal processes with some lag-time (in terms of seconds). Moreover, this lag-time can vary between regions. And, it is exactly the haemodynamic response what we measure with fMRI. Since the latter timely covers several individual successive neuronal and vascular events we are unable to make exact inferences about the rapid neuronal processes. (Huettel et al., 2009; Sutton, 2009; Bandettini et al., 2009).
Higher magnetic fields allow to detect the initial dip of the curve. It is believed to be more specific to the activation locality than the BOLD signal itself. So, in some aspects, high field magnet contributes to better temporal resolution (Sutton et al., 2009)
In conclusion, PET and fMRI are functional imaging methods that are able to indirectly measure neuronal activity. They have different physiological processes (what we are measuring) and they measure very different phenomena. Each generates the output signal differently, over different time scales, and with very different limitations and different time and spatial resolution. In both methods, however there is huge potential and it is believed they will keep to be used more and more.