Brief Introduction to FMRI Physiology

Contrast generation in Magnetic Resonance Imaging

Image intensity observed in MR images is determined by various tissue contrast mechanisms: proton density, T1 and T2 relaxation rates, diffusive processes of proton spin dephasing, loss of proton phase coherence due to tissue magnetic susceptibility variations and in-flow of blood plasma protons. Two dominant tissue contrast mechanisms have functional sensitivity in MR imaging and are produced via haemodynamic responses. Precise changes in brain activation or metabolism are not directly observed, but the effects of local increases in blood flow and microvascular oxygenation on one or more of the above mentioned MR mechanisms can be mapped as a change in raw image intensity.

One mechanism depends upon the fact that the microvascular MR signal on T2 and T2* weighted images is strongly influenced by the oxygenation state of the blood. The rate of loss of proton spin phase coherence is a measure of T2 and local magnetic field homogeneity (T2*); this can be modulated by the presence of intravoxel deoxyhaemoglobin. Recent data shows that the observed T2* is dependent on the presence of blood deoxygenation and that deoxygenated haemoglobin is a "blood oxygenation level dependent" or "BOLD" effect that can be observed by noninvasive MR imaging at high magnetic fields.

The BOLD imaging technique does not measure tissue perfusion or flow directly, however, because over 70% of the brain's blood lies within the microvascular capillaries and venules, the measurement of the magnetic susceptibility-induced T2* signal loss is thought to most reflect the regional deoxyenation state of the venous system. In addition, proton perfusion and diffusion through changing local gradients modulated by changing oxy-/deoxyhaemoglobin levels has a direct impact on the observed T2 relaxation times, which is another mechanism of tissue contrast generation. Amongst these various mechanisms, the T2* effect is larger by factors of 3 to 10 and is the dominant and most widely-studied mechanism employed in fMRI.

In short, the response to a local increase in metabolic rate is increased delivery of blood to the activated region. Such a change in haemodynamics produces small alterations in T1, T2 or T2*, which can be visualized as a change in MR image intensity (approx. 1-10%).

What generates magnetic field susceptibility?

The presence of any substance in a magnetic field alters that field to some extent. Certain metal elements such as gadolinium and dysprosium have an inherently high magnetic moment relative to water or air, and experience pronounced polarization when placed in a magnetic field. The degree of this effect is referred to as the "magnetic susceptibility". The iron in blood haemoglobin is a superb inherent magnetic susceptibility-induced T2*-shortening intravascular contrast agent found in every tissue. It is therefore used as a local indicator of functional activation because oxygenated arterial blood contains oxygenated haemoglobin, which is diamagnetic and has a small magnetic susceptibility effect. It does not, therefore, significantly alter the regional magnetic field and does not greatly affect tissue T2*. Deoxygenation of haemoglobin produces deoxyhaemoglobin, a significantly more paramagnetic species of iron due to the four unpaired electrons, and this species disturbs the local magnetic field, B0, in a region of tissue leading to the large observed magnetic susceptibility effect. The balance of spatial and temporal alterations in local concentrations of deoxygenated to oxygenated iron affects the local observed T2* by causing fluctuations in magnetic susceptibility. Arterially delivered blood consists mostly of oxyhaemoglobin, however, as HbO₂ passes through the capillary bed, the local concentration of deoxyhaemoglobin (Hb) increases and often predominates. Therefore, a T2* gradient can exist across the vascular tree from a diamagnetic HbO₂-rich environment (with a longer relative T2*) to a more “paramagnetic” Hb environment with a shorter T2*.

The local T2* critical in fMRI contrast is thus determined by the balance of deoxygenated to oxygenated haemoglobin in blood within a voxel, which in turn is a function of local arterial autoregulation or vasodilation. By increasing the flow of oxygenated blood or reducing oxygen extraction to a region in the brain an increase in local, intravoxel T2* occurs which in turn leads to an increase in image intensity. An increase in oxygenated arterially delivered blood in response to local activation will result in more oxygenated iron in the capillary and venous vascular beds, thereby creating a relatively longer regional T2* and an image intensity increase. It also reflects a decrease in deoxyhaemoglobin content, i.e. an increase in venous blood oxygenation and a longer effective T2*.

The image intensity for a given voxel in the brain can therefore significantly increase if more oxygenated blood enters this region and fills the venous bed. This assumes, however, that cortical activation causes local vasodilation which is not accompanied by a significant increase in oxidative metabolism. It should be remembered that local image intensity increases will also be dependent on differences in haemodynamic (blood volume, flow and oxygenation) and vessel architecture (radii, orientation, vascular openness).

To illustrate this haemodynamic induced T2* change in the vascular bed the primary visual cortex will be used as an example. During photic stimulation a simple presentation of a flashing lights is given during an image acquisition series. The MR-observable T2* is affected by the balance of HbO₂ to the more paramagnetic Hb existing in the capillary and venous beds. This balance produces a gradient in the local magnetic field and a potent tissue contrast mechanism because the large surface area of the capillary bed amplifies the long range effects on the magnetic field. The photic stimulation produces rapid neuronal activation, which in turn increases cerebral blood flow (CBF), cerebral blood volume (CBV), and oxygen delivery. As CBF increases more than CBV, oxygen delivery quickly exceeds slight increases in local oxygen needs owing to the activation. Increases in local CBF in the arterioles and small arteries that occur rapidly are said to be uncoupled to local metabolism. The net effect is a surplus in the amount of oxygentated haemoglobin delivered to any activated voxel. As the delivered oxygen exceeds local demands, the capillary and venous beds fill with a larger ratio of oxygenated to deoxygenated haemoglobin compared to when the cortex was at rest. This larger amount of diamagnetic oxyhaemoglobin will mean less effect from the field-altering deoxyhaemoglobin, a longer T2*, and to an increased signal on the T2*-weighted images. The actual volume of haemoglobin in the brain is quite small (a few percent), however, the T2* effects extends for microns beyond the vascular bed, because magnetic susceptibility is a long-range effect. This leads to approximately a 1%-10% T2*-induced image intensity increase for a typical cortical activation task. This can be observed from intensities measured from T2*-weighted MR images or from a simple subtraction of images acquired at rest from those acquired during task.

The paramagnetic properties of deoxyhemoglobin have been known for over 50 years, however, the realisation of its usefulness in understanding brain function has only been made possible in the past decade.
© Copyright 1998, Irene Tracey
(Summary figure courtesy of Peter Jezzard):