Mapping brain function with magnetic resonance imaging.

Abstract

Historically, the most compelling argument for the existence of regional specialization of human brain function was presented by Pierre Paul Broca in the middle 19th century. Broca examined a patient who, as a result of a stroke, presented with the problem of inability to speak or aphasia but was otherwise normal. Based on an autopsy performed subsequent to the patient's death, Broca concluded that the seat of the damage was an egg size lesion located in the inferior frontal gyrus of the frontal lobe in the left hemisphere; this general area is now commonly referred to as Broca's area although its precise topographical extent remains somewhat ambiguous. Such lesion studies and, later intraoperative mapping efforts with electrodes have been until now the primary source of our current understanding of functional compartmentation in the human brain. Recent techniques permit the acquisition of such information much more rapidly and with greater spatial accuracy, fueling explosive developments in our investigation of human brain function. For example, the language area first identified by Broca can now be visualized with unprecedented spatial resolution using functional magnetic resonance imaging (NRI) , in data collection times that last only a few minutes. The most significant and revolutionary advance in magnetic resonance imaging in the last several years has been the use of this methodology to non-invasively map areas of increased neuronal activity in the human brain without the use of exogenous contrast agents. Since its initial demonstration (e.g. [1-31), functional magnetic resonance imaging (NRI) has been applied to study a variety of neuronal processes, ranging from activities in the primary sensory and motor cortices to cognitive functions including attention, language, learning, and memory. The majority of NRI experiments are based on the blood oxygenation level dependent (BOLD) contrast [4] which is derived from the fact that deoxyhemoglobin is paramagnetic, and changes in the local concentration of deoxyhemoglobin within the brain lead to alterations in the magnetic resonance signal. In BOLD based NRI, there exists two related but different mechanism underlying the origin of the MRI signal changes coupled to the presence of deoxyhemoglobin; both of these arise because deoxyhemoglobin is compartmentalized within red blood cells in the blood and within blood vessels in the tissue. In the presence of the paramagnetic deoxyhemoglobin, the bulk susceptibility and consequently the magnetic field, of the compartment containing deoxyhemoglobin is different compared to the surrounding environment. Outside the boundaries of this compartment, the magnetic field gradually changes over distances comparable to dimensions of the compartment itself to assume the characteristic value associated with the surrounding environment. This leads to magnetic field gradients immediately outside the deoxyhemoglobin compartment. When these distances are small compared to diffusion distances of water molecules, the magnetic field gradient is dynamically averaged and leads to a decreas in T, signal loss in appropriately weighted MR images. Dynamic averaging is not possible when the distances spanned by the inhomogeneous magnetic field surrounding deoxyhemoglobin containing compartments are large compared to water diffusion distances. In this case, however, static averaging occurs; in other words, different H,O spins within a single voxel experience a different magnetic field and consequently have a different resonance frequency; the signal associated with that voxel represent the static sum of these different signals. This sum is smaller in the presence of magnetic field inhomogeneities when the signal is allowed to evolve for a period before it is sampled. This is referred to as a T,* effect.. It is generally assumed that neuronal activation induces an increase in regional blood flow (rCBD without a commensurate increase in the regional oxygen consumption rate (CMRO,) [51 in which case the capillary and venous deoxyhemoglobin concentrations should decrease, leading to an increase in T,* in tissue around vessels larger than -10 pmeter in diameter and T, in blood itself and around small (-10 pmeter or less diameter) deoxyhemoglobin contain blood vessels. This increase is reflected as an elevation of signal intensity in T,*and T,-weighted MR images, and can be spatially associated with large venous