A fundamental aspect of scientific experimentation is the identification of a control or baseline state against which the condition of interest can be compared. In functional brain imaging studies that use positron emission tomography (PET) or functional magnetic resonance imaging, task-induced increases in regional brain activity during specific goal-directed behaviour are commonly observed by comparison between a specific experimental task and a control task. The difference between the two tasks is usually considered to represent the brain process of interest and it becomes manifest as a regionally specific and task-dependent increase in brain activity. However, task-induced decreases in regional brain activity have also been observed even when the control task consists of lying quietly with eyes closed or the passive viewing of a stimulus. Interestingly, these decreases in brain activity frequently seem to be task-independent, with little variation in location across a wide variety of tasks. The question then, is what do these decreases in brain activation represent? Are they simply unrecognized increases in activation that are present only in the 'control' state? By this account, the 'control' state is just another task with its own activation patterns. Or do such decreases in brain activity represent decreases from a true baseline state? This issue centres on the question of what is the baseline state of the brain.

Marcus Raichle and colleagues used PET to measure a variety of metabolic and circulatory relationships between blood flow and oxygen consumption in the brain. The key measure was the oxygen extraction fraction (OEF) — the fraction of the oxygen available to the brain that is used by the brain. They found a remarkable spatial uniformity in the OEF across the brain when measured in the resting state and propose that this measure might define a baseline state of brain activity.

None of the areas that are regularly observed to show decreases in activation during goal-directed behaviour showed an OEF that was significantly different from the hemispheric mean OEF. This suggests that these brain areas were not activated in the resting state. However, areas within extrastriate visual cortex did exhibit a significant increase in OEF relative to the hemispheric mean, suggesting that these areas are deactivated in the resting state. This interpretation is supported by the observation that areas of extrastriate cortex increase their blood flow when the eyes were opened, suggesting that the baseline state in these regions is observed when the eyes are open rather than closed. The authors propose that the baseline state can be measured as the OEF and, therefore, decreases in brain activation represent true deflections from a baseline state rather than a return to baseline from an unrecognized increase in the control state. So, these areas might be temporarily suspending specific functions that are specific to the baseline state during this period. Although more work is needed to develop these findings, they might have profound implications for functional neuroimaging and our understanding of the processing within the brain. This paper and two others by the same group that appear in the same issue of Proceedings of the National Academy of Sciences begin this exploration.