LONDON - People who suffer a stroke affecting the hippocampus of the brain can usually recall events that happened before their illness, but may lose the ability to form new memories.

Whatever it is that takes place in the hippocampus, it allows us to remember our names, who is dear to us and where we live - as well as what we had for breakfast this morning.

But what happens in the hippocampus that allows memories to form? Neuroscientists have been puzzling over this question for a long time. Now, a new piece of research is causing them to reconsider an old hypothesis about how memories are generated. The finding could lead to new pharmaceutical approaches to treat the memory loss that often accompanies old age, or to ameliorate the early stages of Alzheimer's disease.

The hippocampus receives nerve impulses from the sensory cortex of the brain. The impulses are its window to the outside world, so to speak. Each nerve impulse that arrives at a synapse in the hippocampus causes the release of neurotransmitter, which acts on a target cell and produces a response. The magnitude of this response is a measure of the efficacy or weight of that synapse.

If all synapses had the same synaptic weight all the time, it would be very difficult to imagine how new information could be stored in the brain. But this information could be stored if neural activity - the response to events in the world outside - caused changes in synaptic weight, and if these changes were then maintained.

Neuroscientists have been trying to prove this hypothesis for a long time, with the help of an experimental phenomenon in rats known as long-term potentiation. This occurs when implanted electrodes are used to deliver brief high-frequency impulses to the hippocampus. The result is changes in synaptic weight which can last for days or weeks. But does this long-term potentiation account for the storage of new information, and thus the ability to learn?

Researchers have shown that when drugs that block the receptors of chemicals involved in the induction of long-term potentiation are infused into the hippocampus of rats, learning is impaired. But although such studies suggest that long-term potentiation is involved in learning, they do not prove it.

Another way to gain evidence for the hypothesis would be to increase the synaptic weight of each synapse to its maximum, by providing maximum stimulation. This strategy, known as saturation of long-term potentiation, would make it impossible for the subtle manipulations in synaptic weight that are thought to underlie learning.

Earlier Results Couldn't Be Replicated

During the late 1980s, one team of researchers led by Bruce McNaughton and Carol Barnes had data that appeared to support this theory. In two papers, they described how they repeatedly induced long-term potentiation via an electrode inserted into the hippocampus of rats. When they tested the animals' memory, they found that memory was impaired in those animals that had been stimulated, whereas it was not in control animals.

But when several groups of researchers tried to replicate the results, they failed. So did McNaughton and Barnes when they tried to repeat their own experiment.

The hypothesis began to fall out of favor. However, a new report in this week's Science puts the theory firmly back in the limelight. Edvard Moser, associate professor of biological psychology at the Norwegian University of Science and Technology, in Trondheim, together with Richard Morris at the Centre for Neuroscience at the University of Edinburgh, in Scotland, and colleagues present data showing that if long-term potentiation is saturated, learning is impaired. The title of their paper is: “Impaired spatial learning after saturation of long-term potentiation.“

Moser told BioWorld International: “After several groups failed to replicate the finding by Barnes and McNaughton, nobody really believed in this approach.“ But Moser and his colleagues decided to investigate strategies that would improve the saturation of long-term potentiation, believing that the previous studies had not saturated a great-enough proportion of neurons.

“We started out believing that we would find that long-term potentiation is related to learning, but it has been a four-year struggle and we have been through all kinds of emotions, from surprise to depression, so it was a big relief when we saw that it finally worked,“ he said.

Long-Term Potentiation Related To Memory

The group began by removing the hippocampus on one side to reduce the volume of nervous tissue that needed to be saturated. They also used two electrodes, each with two poles, to stimulate the perforant path of the hippocampus, so that the bundle of fibers was straddled by four peripheral electrode poles. The electrodes were used to stimulate the bundle in all possible permutations.

Moser and his colleagues then trained the rats to find a platform hidden in a pool of water (the water maze test). Rats whose memory is impaired will subsequently swim randomly around the pool, whereas unaffected animals concentrate their swimming in the region of the hidden platform.

In addition to the four electrode poles inserted in the bundle of nerve fibers, a third electrode was inserted with its end at the center of the bundle. After the rats had been trained, this third electrode was used to check whether more long-term potentiation could be induced in the perforant path. This procedure formed a check on whether long-term potentiation was saturated or not. If it was, then no further potentiation would be possible.

“We showed that some animals - those which did not show any more long-term potentiation following stimulation with the third electrode, and which we believed therefore were more or less saturated - failed to learn,“ Moser said, “whereas, in those which showed more long-term potentiation, we showed that they were still able to learn.“

Tim Bliss, head of the division of neurophysiology at the U.K.'s National Institute for Medical Research in Mill Hill, told BioWorld International the finding “strengthens the widely believed but unproven hypothesis that the experimental phenomenon of long-term potentiation is actually related to memory.“

Writing in Science, in an article titled “The saturation debate,“ Bliss concludes: “Despite [certain] caveats . . . the new data clearly demonstrate that even partial saturation of synaptic weights in a distributed neural network will degrade its function.“

The search is on for drugs to enhance long-term potentiation, as it seems likely that these may affect learning ability, Bliss suggested. “It is thought that, because long-term potentiation lasts such a long time, gene activation is involved, and researchers are looking for genes which are turned on by neural activity,“ he added. “You can imagine that it might be possible to compensate for the loss of memory associated with aging, or that which occurs due to death of cells in early Alzheimer's disease, by increasing the ability to potentiate the synapses that are left.“

Moser and his colleagues plan to find out if it is possible to reverse the learning impairment of animals with saturation of long-term potentiation, once the potentiation decays. They also will be looking at the roles different types of cells play in potentiation of synaptic transmission during learning. *

Venture Development Firm