n the 1860s and 1870s Paul Broca, the great French neuro-anatomist, showed that damage to a particular area of the brain prevented people from producing language, both spoken and written. It was one of the first demonstrations of an idea now taken for granted, that the brain practices division of labor, with different parts specialized for different tasks. Over the next few years Broca used thermometers on the skull to measure temperature changes associated with the increased blood flow brought on by particular mental tasks, thus mapping mental faculties to cerebral real estate. A hundred and thirty years on, Broca’s intellectual heirs claim the #6 spot in the Biology Top Ten.

The modern equivalent of Broca’s thermometers is functional magnetic resonance imaging. fMRI scanners detect changes in the availability of oxygen. Nikos Logothetis and his group at the Max Planck Institute for Biological Cybernetics in Tübingen, Germany, examined the underlying neural activity that creates the signal detected by fMRI. They did so by developing a new type of magnet that allowed them simultaneously to record fMRI signals and various direct measures of neuron activity in monkeys. The monkeys sat in an MRI scanner looking at a checkerboard pattern. At the same time electrodes picked up activity from individual cells in the visual cortex of the brain, which is responsible for the primary processing of visual input.

Crucially, the researchers could distinguish between action potentials and local field potentials. Action potentials are the all-or-nothing spikes that transmit nerve impulses over longer distances. They normally occur immediately after the presentation of a stimulus and represent the output of the nerve cells. Local field potentials vary much more slowly and are generally part of the input process, particularly the process by which neurons integrate the inputs from several different sources. Logothetis’s work shows that the fMRI signal, which is known as the blood-oxygen-level-dependent (BOLD) signal, reflects input to an area rather than outputs from it. Some researchers had previously assumed that the BOLD signal was a manifestation of spike activity rather than signal processing.

By recording neuron activity at the same time as fMRI signals, Logothetis, who shared the 2003 prize for medicine awarded by the Louis-Jeantet Foundation in Switzerland, has shown that the signal-to-noise ratio of the neural signal was at least 10, and often 100 times greater than that of the fMRI signals. This implies that analysis of fMRI signals probably underestimates the actual amount of brain activity called for by a given task. "A certain degree of caution is called for when interpeting mapping studies," the paper advises, "particularly when precise localization of activity is required."

While this is interesting in itself, Marcus E. Raichle, a pioneer of fMRI at Washington University in St. Louis, points out in a companion commentary in the same Nature issue (page 128) that the work of the Logothetis group has forced a reappraisal of the links between brain activity and energy metabolism. Despite representing only 2% of the body’s mass, the resting brain uses up to 20% of the body’s oxygen supply. The oxygen is used to break down glucose, in order to supply energy to the neurons. But when an area of brain becomes active its blood flow and glucose consumption are much greater than the increase in oxygen consumption would suggest. This is because the first flush of energy is supplied by the rapid anaerobic breakdown of glucose, which does not need oxygen and which thus results in a brief increase in the supply of available oxygen. This is the blood-oxygen-level-dependent signal that fMRI detects.

The main excitatory neurotransmitter is the amino acid glutamate, which carries a signal from one nerve cell to another across a synapse. Once the glutamate has been released it must be removed to clear the way for another signal if needed. An adjacent cell, an astrocyte, absorbs the glutamate and converts it to glutamine before recycling it back to the neurons. It does so by the anaerobic glycolysis of glucose derived from the blood and, at times of sudden brain activity, from a glycogen store within the astrocytes. The BOLD fMRI signal represents an increase in astrocytes processing glutamate after an episode of excitatory stimulation.

Note: The presence of only one human-genome paper at the top of the list is not significant. Despite being published in the same week, the two monumental papers that have exerted a lockhold on #1 and #2 were actually indexed by Thomson ISI in different bimonthly periods. That means that one drops out this time for being too "old" while the other remains behind. By next issue both will have exceeded the Hot Papers two-year limit.