Disciplines Converge In Probe Of Memory And Learning

Humans and sea snails have a lot in common when it comes to learning and memory. Indeed, neuroscientists have found that little has changed at a cellular level since we departed evolutionarily from these mollusks. And this is just one of the recent findings that has brought neuroscientists to the edge of translating the molecular biology of nerve cells into an understanding of how humans first obtain and then retain information, sound, and images throughout the 70, 80, or even 100 years of a lifetime. A marriage between molecular biology and psychology may soon be in the offing. "Learning and memory is popping up in all kinds of places, not just psychology, where you expect it, but in cell biology, in physiology, in anatomy, in computation," says neuroscientist Thomas Carew, chairman of psychology at Yale University. "This tremendously important and interesting field is now seen as a legitimate topic for someone in molecular biology. Ten years ago, if someone who's interested in gene expression said, `Let's study learning,' you'd say, `You're on a different wavelength!' But now good molecular biologists, good psychologists, good physiologists all have something to contribute to the issue." Keeping track of those contributions has its challenges, says Charles Stevens of the Salk Institute in San Diego, a highly cited researcher in the area of long-term potentiation (LTP). "I have to read not only the molecular biology and the electrophysiology literature, but I have to keep up on the psychology literature, too. Everybody who works in learning and memory does this sort of thing," he says. When the gargantuan Society for Neuroscience, which over the past decade has grown from 7,982 members to 19,716, descends upon Anaheim, Calif., on October 25-30, learning and memory will occupy 22 of the annual meeting's sessions. And that's not counting related sessions, including five devoted to LTP, one of the hottest and most controversial explanations of where memory resides. "LTP occurs when you stimulate a neuron with a certain frequency, and the synaptic connections get stronger," explains Gary Lynch at the Center for the Neurobiology of Learning and Memory at the University of California, Irvine. Stronger connections mean more of a neurotransmitter being released at a synapse, as well as an increase in the number of receptors on the postsynaptic neuron. "It's been shown that you can, in a few tens of milliseconds, change the strength of synaptic contacts in a way that will last for weeks or months," says Lynch. "That, of course, sounds like memory." Stevens says he has had to overcome earlier reservations about the validity of LTP as a mechanism for storing information. "I was, personally, skeptical about whether LTP actually had anything to do with learning and memory until fairly recently," he says. "The first piece of evidence before was, we didn't have any alternatives. You tend to believe something when you don't really have anything else." And LTP, at least superficially, had all the right properties: It was long-lasting, quickly induced, and synapse-specific, he adds. Plus, researchers had shown that when you block NMDA receptors, thus blocking LTP, rats are unable to learn how to negotiate a maze. But because no one could tease apart the roles of NMDA and LTP, this last piece of evidence didn't conclusively tie LTP to memory. Evidence from a collaborative study Stevens's lab has recently done with a group headed by Susumu Tonegawa at the Massachusetts Institute of Technology has eased his skepticism. Their much- publicized work with mice that lack a gene encoding the a- calcium-calmodulin kinase II (CAM kinase) showed that the genetically altered mice couldn't learn to find their way out of a water maze (A.J. Silva, et al., Science, 257:201-5, 1992). The mice's synapses were normal and the receptors were normal. The only thing missing in the altered mice, says Stevens, was CAM kinase, which has been implicated as one of several kinases that has a role in creating LTP in a nerve cell. The other experiment that Stevens says has converted him to an LTP believer was carried out by Larry Squire and Stuart Zola- Morgan (L.R. Squire, S. Zola-Morgan, Science, 253:1380-6, 1991). In experiments with monkeys, they showed that animals without the hippocampus portion of their brains can't learn a specific task. However, if researchers teach the animals the task, wait a month, and then remove the hippocampus, the monkeys can still perform the task perfectly well. "This is suggestive that what's happening is the memories are being stored in the hippocampus for the first couple of weeks, which is about how long LTP lasts, and then they're getting moved to some other place," says Stevens. "All of these experiments are adding weight on the side of LTP actually being the storage mechanism. "That's pretty neat, because although there are a lot of gaps, we now have an explanation for learning that goes right from the molecular level, up to the cognitive level." Several labs have provided evidence that different types of memory are stored in neural circuits of mammalian brains in different areas. A process separate and parallel to LTP, called post-titanic potentiation (PTP), has been found in the hippocampus. PTP appears to last for a few minutes, compared to LTP's duration of many days, suggesting that long-term and short- term memory can be processed separately and simultaneously in the same neuron, says Carew. His research group at Yale made an important discovery that brings this notion down to the cellular level. Colleagues Nigel Emptage and Laura Stark will provide evidence at the neuroscience meeting that the prevailing idea that long-term memory is simply the deep-storage version of the same short-term memory is too simplified. In a paper published late last year (A.R. Mercer, et al., Science, 254:1811-13), the group reported that a drug that blocks the actions of serotonin also blocks short-term changes in the amount of neurotransmitter released in sensory neurons of the sea snail Aplysia. They recently found that even with the short-term enhancement blocked by the drug, long-term changes are still present. "The mechanism of short-term memory exists for about an hour and then decays back to baseline," Carew explains. "By three hours there's no evidence for memory, and you might think, `Gee, the synapse is back to baseline; it no longer retains the memory.' But that very same synapse, 24 hours later, shows clear synaptic facilitation." Carew doesn't know whether the short-term and long-term changes are PTP and LTP, but he says "there is a parallel at the conceptual level." Because this long-term facilitation has been found only at the synaptic level, Carew urges that his findings not be overstated. "We don't know if this is how it works at the organismic level," Carew says. "Learning is not simply a change in one synapse; learning is a change in the behavior of an organism. Memory is not a change in one synapse; memory is a reflection of learning as expressed by an animal at a later time," says Carew. "So we have predictions based on our cellular observations, and now we're currently examining those predictions on a behavioral level. "That's what you do. You not only keep going deeper and deeper, more basic, from circuits to synapses to the genome, but you go backward and you say, given these cellular observations and these molecular observations, what can they predict for us about behavioral outcome? And it's that backward and forward give-and- take that makes the field move forward." Daniel Alkon, chief of the Neural Systems Laboratory at the National Institute of Neurological Disorders and Stroke, says he is particularly excited about new findings that have relevance to Alzheimer's disease. Several years ago, Alkon and his colleagues showed that potassium channels are regulated during learning by CAM kinase and by protein kinase C. Beginning with the hypothesis that this disease has a primary effect on memory storage, his group began to look at potassium channels similar to those that changed in snail and rabbit neurons for days or even weeks after associative learning. "Together with Gary Gibson at Cornell, we asked, `Is it possible that Alzheimer's involves a derangement of the same potassium channel regulation that we had implicated in memory storage?' " says Alkon. The answer is yes. They've looked at fibroblasts from Alzheimer's patients, normal age-matched controls, and normal younger people. An analysis of the potassium channels and a separate study of the calcium signals that depend on the integrity of the channels both showed that "a certain potassium channel effectively is absent," says Alkon. "I say `effectively' because we can still allow for the possibility that it might be silent and just not functional." The hypothesis that they're now testing is that potassium channel function in general has some subtle differences in the way it is regulated by kinases, for example, in Alzheimer's patients, and that only under conditions of age and vulnerability, perhaps, does that dysfunction become apparent. "This may be a more subtle and perhaps a more causal kind of change that subsequently leads to the more destructive lesions and, ultimately, symptomatic appearance of an Alzheimer's patient," Alkon suggests. So where is this research leading? For one thing, it's leading not just to excitement among scientists, who see the new molecular biology as a way of finally getting a handle on how people think and remember, but also to an upsurge in interest from the general public. To address the public's interest, Alkon has written a book, aimed toward an educated lay audience, called Memory's Voice, due to be published in November by Harper Collins of New York. Alkon also sees learning and memory research leading straight to clinical applications: "The future is the clinic. It's being able to offer people who are recovering from a variety of diseases assistance in relearning, or in making available old memories that are buried." Dominick Purpura, dean of Albert Einstein College of Medicine in New York and editor of the journal Brain Research, says that with so many questions left to answer, no one supposes that learning and memory research will cool off or stop growing any time soon. "Everybody will be looking for agents to potentiate or get involved with the kinases and the receptors, and out of this will come some considerable biotechnology industry stories, looking for memory molecules," he says. "It's only around the corner." Laurel Joyce is a freelance writer based in Stanford, Calif.

Disciplines Converge In Probe Of Memory And Learning

Interested in reading more?

Become a Member of