COVER STORY: SEPTEMBER 13, 1999 VOL. 154 NO. 10

When everything is going right, these different systems work together seamlessly. If you're taking a bicycle ride, for example, the memory of how to operate the bike comes from one set of neurons; the memory of how to get from here to the other side of town comes from another; the nervous feeling you have left over from taking a bad spill last time out comes from still another. Yet you are never aware that your mental experience has been assembled, bit by bit, like some invisible edifice inside your brain.

	ALSO IN TIME
	Brainteaser The ethics of engineering minds Pills An industry devoted to memory supplements Viewpoint The discovery disproves two common fallacies

And brain researchers might never have picked up on the fragmentary nature of memory without their studies of people whose memory has been damaged by illness or injury. The most celebrated such individual is H.M. In 1953, when he was 27, he had drastic brain surgery to cure severe epilepsy. The operation cured his epilepsy, but removing parts of his brain's temporal lobes, including a structure called the hippocampus, destroyed his ability to form new memories. H.M., who is still alive, has a reasonably good short-term memory. Once introduced to a visitor, he will remember the person's name and other information while a conversation lasts. But if the visitor leaves and returns, H.M. has no memory whatsoever of having met the person. In fact, H.M. has no permanent memory of anything that happened after his surgery. As far as he's concerned, it's still 1953, and that old man looking back at him from the mirror bears only a passing resemblance to the young man he knows himself to be.

That sort of impairment has convinced scientists that the medial temporal lobe and hippocampus are key in transforming short-term memories into permanent ones, and also that permanent memories are stored somewhere else; otherwise, H.M. would have lost them too.

But a remarkable experiment performed in 1962 by Canadian psychologist Brenda Milner proved that H.M. can form new memories of a very specific sort. For many days running, she asked him to trace a design while looking in a mirror. As far as H.M. knew, the task was a brand-new one each time he confronted it. Yet as the days wore on, his performance improved. Some part of his brain was retaining a memory of an earlier practice session, a so-called implicit--rather than explicit, or consciously remembered--memory. People who suffer from Alzheimer's disease exhibit the same sort of behavior--and it's the medial temporal lobe that is first affected by this devastating disease.

In patients with Huntington's disease, it's the part of the brain called the basal ganglia that's destroyed. While these victims have perfectly intact explicit memory systems, they can't learn new motor skills. An Alzheimer's patient can learn to draw in a mirror but can't remember doing it; a Huntington's patient can't do it but can remember trying to learn. Yet another region of the brain, an almond-size knot of neural tissue known as the amygdala, seems to be crucial in forming and triggering the recall of a special subclass of memories that is tied to strong emotion, especially fear. The hippocampus allows us to remember having been afraid; the amygdala evidently calls up the goosebumps that go along with each such memory.

These are just some of the major divisions. Within the category of implicit (a.k.a. nondeclarative) memory, for example, lie the subcategories of associative memory--the phenomenon that famously led Pavlov's dogs to salivate at the sound of a bell, which they had learned to associate with food--and of habituation, in which we unconsciously file away unchanging features of the environment so we can pay closer attention to what's new and different upon encountering a new experience.

Within explicit, or declarative, memory, on the other hand, there are specific subsystems that handle shapes, textures, sounds, faces, names--even distinct systems to remember nouns vs. verbs. All of these different types of memory are ultimately stored in the brain's cortex, within its deeply furrowed outer layer--a component of the brain dauntingly more complex than comparable parts in lesser species. Experts in brain imaging are only beginning to understand what goes where, and how the parts are reassembled into a coherent whole.

What seems to be a single memory is actually a complex construction. Think of a hammer, and your brain hurriedly retrieves the tool's name, its appearance, its function, its heft and the sound of its clang, each extracted from a different region of the brain. Fail to connect a person's name with his or her face, and you experience the breakdown of that assembly process that many of us begin to experience in our 20s--and that becomes downright worrisome when we reach our 50s.

It was this weakening of memory and the parallel loss of ability to learn new things easily that led Princeton molecular biologist Joe Tsien to the experiments reported last week. "This age-dependent loss of function," he says, "appears in many animals, and it begins with the onset of sexual maturity."

What's happening when the brain forms memories--and what fails with aging, injury and disease--involves a phenomenon known as "plasticity." It's obvious that something in the brain changes as we learn and remember new things, but it's equally obvious that the organ doesn't change its overall structure or grow new nerve cells wholesale. Instead, it's the connections between new cells--and particularly the strength of these connections--that are altered by experience. Hear a word over and over, and the repeated firing of certain cells in a certain order makes it easier to repeat the firing pattern later on. It is the pattern that represents each specific memory.

How this reinforcement happens was a puzzle for much of this century, until 1949, when Canadian psychologist Donald Hebb came up with a related notion: since most memories consist of a group of disparate elements coming together--the hammer again--something more must be happening than just an electrical signal in one brain cell setting off a response in another. Something in the brain must be acting as a "coincidence detector," taking biochemical note that two nerve cells are firing simultaneously and coordinating two different sets of information.

PAGE 1 | 2 | 3