As circuitry goes, the human brain has always seemed a doozy.

It contains something like 100 billion nerve cells, or neurons, and each neuron may transmit signals to other neurons at up to 1,000 points of communication, called synapses. So these junctions total in the trillions.

		Related coverage Online graphic: How nerve impulses are transmitted Upsetting conventional wisdom on neurotransmitters

That might seem complex -- and it is -- but Pittsburgh computational neuroscientist Dr. Joel Stiles says that view now appears to "very grossly underestimate" the number of interconnections and, thus, the brain's overall complexity.

A study by scientists in La Jolla, Calif., and by Stiles at the Pittsburgh Supercomputing Center, published Friday in the journal Science, found that the operation of those synapses is not as straightforward as has been depicted in textbooks.

"We have the potential for much more complexity in how the brain works and this complexity underlies what we call personality and behavior," Stiles said.

Nerve signals travel through neurons as electrical impulses, but generally are converted into chemical signals when transmitted from one neuron to another. The conventional view is that this signal is transmitted when one neuron releases chemicals called neurotransmitters, which then cross the tiny gap between the cells at the synapse, and are received by specialized proteins on the surface of the neuron on the other side.

The new study, which combines insights from three-dimensional electron microscope maps of synapses, direct measurement of nerve signals in the laboratory and computer simulations of nerve-to-nerve signaling, suggests that this firing of neurotransmitters is less like a rifle shot than a shotgun blast.

And even that analogy doesn't quite capture it. The computer modeling strongly suggests that most of these neurotransmitters aren't released from a single spot, called the active zone, but from any number of sites along the outer edge of the neuron

This idea that neurotransmitters are released outside the active zone, which has been dubbed "spillover," has been a matter of ongoing debate among neuroscientists, said Vladan Lucic and Wolfgang Baumeister of the Max Plack Institute of Biochemistry in Martinsried, Germany, who wrote a commentary that accompanied Friday's report.

But they concluded the study provides convincing evidence that this spillover effect is occurring, though for now the finding only pertains to ciliary ganglion, the nerve bundle that the scientists studied. The ciliary ganglion, in this case from chicks, connects the brain with the nerves that control the diameter of the pupil of the eye.

The lead investigator, Terrence Sejnowski of the Salk Institute for Biological Sciences and the University of California at San Diego, acknowledged that they can only be sure that the findings apply to the ciliary ganglion.

But because the findings so heavily favor neurotransmitter release outside the active zone, he said, "you can't really trust the traditional textbook view . . . that's taken for granted now."

Sejnowski and Stiles said no one knows what this spillover release is all about. "Perhaps the most surprising thing is that it's the predominant form, rather than the exception," Stiles said, so it likely is important and may send different types of messages than "classical" releases at a synapse's active site.

The researchers studied ciliary ganglia from chicks -- a favorite type of neuron for study because the nerve bundle is reasonably accessible for lab measurements and plays an important physiological role.

The synapses themselves are highly complex, with finger-like projections from each neuron intermingling with each other. This complexity in the past has forced investigators to develop simplified models of the synapses.

But the use of sophisticated computer software allowed the reseachers in this case to draw upon highly detailed 3-D maps of synapses produced by high-voltage electron microscopy scans in the lab of co-author Mark Ellisman at UC San Diego.

Synapse-level measurements of neurotransmitter releases in neurons were made in the lab of UC San Diego biologist Darwin Berg.

These were combined into a highly detailed computer simulation by Jay Coggan and Thomas Bartol of the Salk Institute, using software developed by Bartol and Stiles and an additional modeling tool developed in Pittsburgh by Stiles. The simulations showed that the type of signaling recorded by Berg could not be replicated by neurotransmitter releases only from the active sites.

"Computer simulations based on carefully built, realistic models provide important insights that cannot be obtained by current experimental methods," Lucic and Baumeister wrote. "We hope to see more such studies in the future."

Stiles said this type of computer simulation is likely to play an important role in sorting out the complex signaling that occurs within in the brain and in developing an understanding of the links between brain activity and behavior.

"It'll keep me busy for the rest of my career," he added.