How Can We Make BOLD Contrast Bolder?

Functional magnetic resonance imaging (fMRI) based on blood oxygenation level–dependent (BOLD) contrast is a noninvasive technique that offers an unprecedented opportunity to explore the neuronal basis of human cognition, perception, and behavior. Despite extensive studies over the last decade, we still have only a rudimentary understanding of the relationship between the BOLD fMRI signal and the underlying neuronal activity. Understanding this relationship is important if we are to discover methods to increase the sensitivity of what is now a technique with relatively low sensitivity, which has particularly limited the application of fMRI in the evaluation of cognitive and psychiatric disorders.

It is now generally accepted that the BOLD fMRI signal changes in response to activation stimuli are related to the increase of regional cerebral blood flow (rCBF), which alters the relative local concentrations of oxyhemoglobin and deoxyhemoglobin. However, little is known about the process by which focal neuronal activity triggers the increase in rCBF. A complex system involving vasoactive substances, nitric oxide, neurotransmitters, and intrinsic factors has been postulated. In addition to this local coupling of rCBF with neuronal activation, CBF is also sensitive to factors acting globally; these include perfusion pressure, the partial pressures of CO₂ and O₂ in the cerebral circulation, and other biologic and pharmacologic variables.

In view of these global perfusion regulatory systems, the question of whether the increase in rCBF during neuronal activity is linked to the baseline cerebral blood flow arises. The clarification of the relationship between the baseline cerebral blood flow at rest and the increase in rCBF as a result of focal neuronal activity may lead to a better understanding of the processes underlying the BOLD response. It also has substantial practical relevance for the intersubject comparison of BOLD signal intensity during task activation. The dependency of the rCBF associated with a neural activation on the prevailing cerebral blood flow has been the focus of many functional neuroimaging investigations.

For BOLD fMRI, changes in global cerebral blood flow and rCBF are indirectly related to changes in cerebrovascular oxygenation, but this fact accounts only for a small fraction of the measured signal intensity on T2*-weighted images. Therefore, the physiologic basis for the current analytical models of the BOLD signal is incomplete. This lack is reflected by the often-inconsistent reports on the dependence of local activation–related BOLD signal changes on the baseline BOLD signal modulated by various approaches such as hypercapnia, hypocapnia, and vasoconstrictive or vasodilatory drugs. The report of the animal study by Morton et al (1) in this issue of the AJNR represents another effort to clarify the relationship between the local activation–induced BOLD signal changes and globally modulated baseline BOLD signal amplitude.

In brief, the results from previous research are complex and sometimes confusing. The effect of global cerebral modulation (either vasodilation or vasoconstriction) on the local activation–induced BOLD response can be either up- or down-regulating, depending on the experimental setting. On the basis of the current understanding of the BOLD effect, the activation-induced BOLD signal changes are dependent on both the rCBF response and the cerebral metabolic rate of oxygen. If the activation-related rCBF response is assumed to be independent of the prevailing CBF level, as the direct perfusion results from arterial spin-labeling fMRI studies (2) implicate, the observed variations in the BOLD response under conditions of global cerebral vasodilation and vasoconstriction can only be attributed to changes in the oxygen consumption. Quite possibly, CO₂-, O₂-, and drug-induced global modulations of vasomotor tone may lead to different functional-metabolic couplings of tissue during task activation. However, the precise mechanisms underlying the effect of each agent on the BOLD response is not yet well understood.

In this study, Morton et al (1) found that the systemic administration of theophylline markedly increases the BOLD response in rats. The findings from this study are similar to the recent results of a study about the use of caffeine as a contrast booster for BOLD fMRI studies in human subjects (3). Given the fact that caffeine and theophylline belong to the same methylxanthine family of drugs and that both drugs act as vasoconstrictors in the brain, the two drugs are likely to have the same mechanism regarding the booster effect on BOLD contrast. It was suggested that the enhanced BOLD sensitivity was possibly due to the increased concentration of deoxyhemoglobin at the resting state that results from the reduced CBF during vasoconstriction. This interpretation is conceivable, but it fails to explain why vasoconstriction induced by indomethacin (4), for example, attenuates the BOLD response. The neuroexcitability effect of these drugs offers another plausible explanation for the resetting of the coupling between cerebral blood flow and energy metabolism.

A full understanding of why and how BOLD fMRI signal is related to the underlying neuronal activity requires a great deal of further research. However, this issue is very important because the ultimate success of fMRI depends on the establishment of a specific relationship between the fMRI signal and neuronal firing activity. The characterization of the relationship between local BOLD response and the prevailing cerebral blood flow level modulated by chemical agents is a promising approach that can potentially elucidate the cascade of processes that trigger rCBF change in response to task activation. Despite an incomplete understanding of the underlying mechanisms, the study by Morton et al reveals a significant increase in the BOLD response in rats by using systemically administered theophylline. This finding indicates that the development of BOLD contrast boosters to overcome the low sensitivity of fMRI technique might be possible; this improvement may make fMRI more applicable in clinical medicine.

• Copyright © American Society of Neuroradiology