Functional Magnetic Resonance Imaging (fMRI)

In September 2002 the ULG Cyclotron Research Centre equipped itself with some high performance functional Magnetic Resonance Imaging (fMRI) hardware. In this way it significantly expanded its capacity to explore the human brain). This equipment, entirely devoted to research, benefits from a very high magnetic field (3 Tesla against 1.5 for classical installations).

At the beginning of the 1970s, Paul Lauterbur of the University of Urbana, in the United States, first gave birth to Magnetic Resonance Imaging (MRI). What is the principle behind it? The supply of morphological images by utilising the magnetic properties of one of the major components of biological tissues: the hydrogen nucleus. The fMRI is the worthy descendant of magnetic resonance imaging. It came to life in 1991, following work carried out by Jack Belliveau and his team at the Massachusetts General Hospital for the needs of a study into the visual system. During this first study, the technique employed required the injection of a paramagnetic tracer, gadolinium. There is much about this procedure that reminds one of techniques used by PET.

In being diluted into the blood, gadolinium changes the local magnetic properties of tissues, allowing one to visualise the brain's active regions. In actual terms, the functional Magnetic Resonance Imaging (fMRI) equipment must be seen as an enormous magnet. It was one year later, in 1992, following work carried out by Ken Kwong and Jack Belliveau’s team, that the fMRI was revealed in its present form, that is to say in a form that no longer required the injection of no matter what radioactive tracer. Thence revised and fine tuned, it is based on real time observations of the blood's volume, flow and oxygenation.

In the study of the brain it exploits, in a more precise way, a dropping in the concentration of deoxyhaemoglobin in the flow downstream of the activated neurons. In the cerebral regions which become activated the influx of blood and the oxygen it carries exceeds transitorily local energy needs. As a consequence one witnesses the massive arrival of oxyhaemoglobin, a combination of haemoglobin and oxygen, and, on leaving, a relative reduction in deoxyhaemoglobin. The latter, which is paramagnetic, reduces the MRI signal. Also, in the activated regions, the fall in the concentration of this deoxyhaemoglobin leads to an increase in the signal as the concentration of the substance which reduces the signal has diminished.

Today functional Magnetic Resonance Imaging (fMRI) has established itself as the technical benchmark in studies of the functioning of the human brain, and in particular cognitive functions. De facto, it offers undeniable advantages. One of them resides in its innocuousness. The radioactive tracers used in positron emission tomography are harmless in the doses they are administered. However, repeated experiments on the same subject are ruled out because of the cumulative effects of radiation exposure. FMRI gets around this obstacle, as it can participate in experimental protocols excluded in the use of PET.

Another crucial advantage: functional Magnetic Resonance Imaging (fMRI) is capable of showing spatial and temporal features in excellent resolution. The resolution of cerebral activation maps obtained by PET is of the order of 5 millimetres (2 to 3mm for new generation models). It is distinctly higher in fMRI – of the order of a millimetre. Obtaining a functioning image of the brain by positron emission tomography takes a minute and a half. With fMRI a cross-section of the brain is obtained in 60 milliseconds and the total volume of the whole brain in 2 seconds.

With functional Magnetic Resonance Imaging the limits are now physiological, not technological. The arterioles that irrigate the neurons take 5 to 6 seconds before opening in response to neuronal activation and around 20 to close after the discharge which follows. The only way of managing to depict neuronal activities in fine detail is to couple fMRI, a haemodynamic technique whose spatial resolution is millimetrical, with very high resolution electro-encephalography (EEG), spatially imprecise but temporally very precise.

The principle is clear, but putting it to work is more arduous. Because, dealing with magnetic fields and electrical currents, EEG causes distortions on fMRI images, which for their part lead to alterations in electro-encephalographic recordings. 


© 2007 ULiège