These live dendrites in the brain are covered with slender-necked spines that help to accurately sum all the incoming nerve signals.

When neuroscientist Rafael Yuste likens the human brain to a computer, he is zeroing in on its breathtaking simplicity, a circuitry whose logic, he postulates, is the stuff of first-grade math.

Yuste, an HHMI investigator at Columbia University, studies the cortex, the seat of perception, memory, and language. He and his colleagues treat the cortex in the same way engineering students are taught to treat a device of unknown function: "We take apart the box, look at the wires and transistors, and try to identify the logic of its circuitry."

The potential of this reverse-engineering approach is evident in Yuste's recent investigation of an item in the cortical "parts list" called the dendritic spine.

In most cortical neurons, these tiny knob-like features are liberally scattered over the surface of dendrites—the projections emanating from neuronal cell bodies. When one neuron sends a signal to another, the impulse moves from its cell body, through its axon, to its axon terminals, across a gap, or synapse, to the head of a spine on the receiving dendrite and then to the cell body of the neuron.

The fact that most cortical neurons are covered with as many as 20,000 dendritic spines suggests the spines' importance in processing impulses. Yuste's postdoctoral work revealed that spines are containers for calcium, which controls the strength of the neuron-to-neuron connection. But in a series of papers published over the last two years in Proceedings of the National Academy of Sciences, Yuste provides evidence, using slices of mouse brains, that spines also serve an electrical function, perhaps even more important in the neuron's processing of incoming nerve signals.

Working with Roberto Araya, a postdoctoral associate in Yuste's lab, and Kenneth Eisenthal from Columbia's chemistry department, Yuste started with evidence that nerve signals can be transmitted between neurons that lack spines. While such neurons are comparatively rare in the cortex, their ability to function without the help of spines—still able to marshal the calcium associated with nerve transmission—suggested that "nature doesn't need spines to accomplish this."

Unable to believe that ubiquitous spines were superfluous, the researchers pursued a hunch that spines had an undiscovered electrical role. Using lasers, they "turned on" individual spines to mimic the arrival of an incoming nerve input. Then they measured the voltage generated in the cell body of the same nerve cell, finding that the amount of current delivered through the spine to the soma was inversely proportional to the length of the slender neck supporting the head of the spine.

Photo: Yuste lab

	PAGE 1 OF 2	Continue