In a major breakthrough in brain science, a team of Swiss researchers has captured a picture of what goes on between brain cells when they form a memory.
Reconstructed from electron microscope data, the unprecedented images are a mighty step in the field of neuroscience. The brain remains one of the least understood areas of bodily research, and memory one of its most mysterious processes.
The discovery could help scientists develop treatments of brain disorders. But it also holds significance for one of the hottest areas of computer science today: information processing through so-called neural networks.
The brain has more to teach engineers about this subject than their textbooks do, said Chris Wilkinson and Adam Curtis, neuroelectronic scientists at the University of Glasgow in Scotland, in a recent review in Physics World.
"The human nervous system and brain form what is undoubtedly the most powerful, complex, and least understood signal-processing network in the universe," they said.
This is complicated stuff
Indeed, the complexity of the human brain is daunting. It has 100 billion cells called neurons, and each of these can link with surrounding neurons in as many as 10,000 ways. But when it comes to memory, only a few dozen links, called synapses, make the critical connections.
It has taken neuroscientists a long time to find those connections, let alone figure out how they lock in memory.
The new images, made by D. Dominique Muller and colleagues at the Universities of Bern and Geneva in Switzerland and published yesterday in the journal Nature, show one long-suspected mechanism.
They confirm, for the first time, that neurons lock in memory by strengthening the structural connections between them.
Specifically, they show how two rat brain neurons start out with one connection - one synapse. Within an hour of the original memory-forming stimulus, another of the suction-cup-like synapses has appeared.
It's the discovery of this doubling, which strengthens the memory linkage, that the Swiss announcement calls "a breakthrough in the understanding of brain functioning."
Neurobiologist John Lisman at Brandeis University in Waltham, Mass., calls that work "the clearest evidence to date that new synapses do form over time."
Neuroscientists believe memory works by strengthening synapses, he says. Yet he cautions that structural change is only one possible way to do it. Chemical effects may also be involved.
"The bottom line," Dr. Lisman says, "is that there would seem to be many ways to make a synapse stronger. We haven't yet established any of those mechanisms."
But with these new images, he adds, there is "a clear indication that part of the process is structural." Now, he says, scientists can extend that exploration by teasing out exactly what is going on when synapses double - what chemical process occurs and what genes are involved.
Like neurobiologists, computer engineers are looking at how the brain works. They know how neurons perform as electronic "devices." And they also are beginning to understand how the brain works as a system. But the Scottish scientists point to "a critical gap" in this knowledge.
What is memory and how is it stored? What are the "syntax" and "grammar" of the messages? These are some of the key questions that computer scientists have not been able to answer.
The Swiss images bring the answer to the first question a little closer.
The Swiss work is part of an ongoing memory-research program pursued by a global network of neuroscientists organized under the Human Frontier Science Program.
Nations of the G-7 economic group started the program a decade ago to encourage collaboration among the world's leading scientists.
In their memory research, the Swiss scientists worked with teams at Kumamoto University School of Medicine in Japan and at Oregon Health Sciences University in Portland.
Those collaborations laid the groundwork for the imaging study reported yesterday. In that study, an electron microscope recorded changes in rat brain tissue as neurons were stimulated to activate mechanisms involved in memory and learning. The Swiss team then used those images to reconstruct the three-dimensional models.
(c) Copyright 1999. The Christian Science Publishing Society