The ability to measure the effects of local alterations in blood flow, blood volume and oxygenation by nuclear magnetic resonance has stimulated a surge of activity in functional MRI of many organs, particularly in its application to cognitive neuroscience. However, the exact description of these effects in terms of the interrelations between the MRI signal changes and the basic physiological parameters has remained an elusive goal. We here present this fundamental theory for spin-echo signal changes in perfused tissue and validate it in vivo in the cat brain by using the physiological alteration of hypoxic hypoxia. These experiments show that high-resolution absolute blood volume images can be obtained by using hemoglobin as a natural intravascular contrast agent. The theory also correctly predicts the magnitude of spin-echo MRI signal intensity changes on brain activation and thereby provides a sound physiological basis for these types of studies.

The oxygen extraction ratio (OER) of a tissue describes the interplay between oxygen delivery and consumption and, as such, directly reflects the viability and activity of any organ. It is shown that OER can be quantified using a single magnetic resonance imaging observable, namely the relaxation time T2 of venous blood draining from the tissue. This principle is applied to study local OER changes in the brain on visual stimulation in humans, unambiguously demonstrating a mismatch between changes in blood flow and oxygen metabolism on activation.

The hypothesis was tested that hypoperfused brain regions, such as the ischemic penumbra, are detectable by reductions in absolute transverse relaxation time constant (T2) using magnetic resonance imaging (MRI). To accomplish this, temporal evolution of T2 was measured in several models of hypoperfusion and focal cerebral ischemia in the rat at 9.4 T. Occurrence of acute ischemia was determined through the absolute diffusion constant D(av) = 1/3 TraceD, while perfusion was assessed by dynamic contrast imaging. Three types of regions at risk of infarction could be distinguished: (1) areas with reduced T2 (4% to 15%, all figures relative to contralateral hemisphere) and normal D(av), corresponding to hypoperfusion without ischemia; (2) areas with both reduced T2 (4% to 12%) and D(av) (22% to 49%), corresponding to early hypoperfusion with ischemia; (3) areas with increased T2 (2% to 9%) and reduced D(av) (28% to 45%), corresponding to irreversible ischemia. In the first two groups, perfusion-deficient regions detected by bolus tracking were similar to those with initially reduced T2. In the third group, bolus tracking showed barely detectable arrival of the tracer in the region where D(av) was reduced. We conclude that T2 reduction in acute ischemia can unambiguously identify regions at risk and potentially discriminate between reversible and irreversible hypoperfusion and ischemia.