1996 LECTURE SERIES

Unlocking the Secrets of Memory

Dr. Larry R. Squire
Department of Psychiatry and Neurosciences
University of California, San Diego
January 31, 1996

One of the most fascinating problems now confronting science and one of its ultimate challenges is to understand the biology of behavior and mental life. The processes of perception, attention, motivation, and thought--and central to these higher functions are the processes of learning and memory.

Learning is the process by which we acquire new knowledge about the world and memory is the process by which that knowledge persists across time. Memory is the ordinary result of learning. All forms of animal life including humans have two fundamental ways to adapt to their environment: biological evolution and learning.

On the one hand, humans and other animals have inherited, through millions of years of evolution, many behaviors important for survival. For example, reproductive behavior, machinery for approach and withdrawal and the ability to perceive the world through various sensory modalities. On the other hand, animals also inherit the ability to adapt or change as the result of experiences that occur during an individual lifetime. The experiences we have modify our nervous systems and we can later behave differently as a result. This capacity gives us the ability to learn and remember.

The Capacity to Learn
Most of what we know about the world is not built into our brain at birth, but is acquired through experience. The names and faces of our friends and loved ones, algebra and geography, politics and sports, and the music of Mozart and Beethoven. As are result, we are who we are in good part because of what we learn and remember. In addition, humans have the unique ability to amplify the effects of learning by communicating what they have learned and in so doing, to create cultures that can be transmitted from generation to generation. In this sense, memory is not only then the precious record of personal experience which gives us a personal autobiography and a sense of identity, it's also a major vehicle for behavioral adaptation and a powerful vehicle for social progress. The capability to learn is so highly developed in humans that the rate of cultural change far outstrips the pace of biological evolution. In fact, the size of the human brain does not appear to have increased significantly since our species first appeared in the fossil records several hundred thousand years ago. All these animals, all these brains have the capacity to learn.

Structure of the Brain
One notes the gradual increase in the size of the brain with evolution of the more recent vertebrates being moved from frog, pigeon, rat, monkey, cat, chimpanzee, and human. Nevertheless, one finds inside the brain the same namable areas of the brain and as a first approximation, the same patterns have conductivity. I think if you looked, we might be more impressed with the similarities than with the differences. The human brain is actually 10 times the volume of the monkey brain, and in this context it's worth mentioning as well that the chimpanzee is a protected species and none are being import or are being used in neuro science research, but the monkey has been and currently is a precious and valuable experimental animal for research and neuro science.

All these brains consist of a wide variety of individual cells called neurons and they look, you might think, rather like trees and like flowers. All vertebrae brains contain these same basic cell types and of course the cellular molecular biology of these individual cells is much the same from species to species, one manifestation of the universality of biology. So we think that differences between species, for example, differences in cognitive ability and differences in memory ability, must be due to the numbers of neurons that the different brains have and to the details as to how they're connected to each other. So to provide a sketch of our current understanding about the biology of memory and memory in the brain, I'd like to consider just three points.

First, the brain is organized such that memory is in part a distinct and separable cognitive function that can be fruitfully studied in isolation from perception and other intellectual abilities. Structures including the hippocampus are essential for our ability to lay down in memory an enduring record of experience and damage to these specialized brain structures including the hippocampus causes amnesia, a neurological condition.

Second, memory is not a single faculty of the mind, but it's composed of multiple separate systems. That is, there's more than one kind of memory, not just in a semantic or a philosophical sense, but in the specific biological sense that different kinds of memory have different brain organizations and depend on different brain systems. The major distinction that I'll be drawing is the distinction between conscious knowledge about facts and events. That is, what we ordinarily mean when we use the term memory, and various non-conscious kinds of memory abilities that support skill learning and habit learning and other ways of interacting with the world that reflect dispositions based on experience. Finally, new technology, including magnetic residence imaging or MRI, is now revolutionizing the study of neuro science and is providing direct anatomical information about memory in living human subjects. At the same time though, the bulk of our current understanding about the anatomy of memory continues to come from the kind of systematic work that is only possible in experimental animals including the monkey.

The brain is a highly specialized and differentiated structure organized so that separate regions simultaneously and in parallel carry out computations on separate features or dimensions of the external world. For example, in the case of visual modality, the analysis of visual pattern, analysis of visual location, the analysis of color, and the analysis of movement are all carried out separately and in parallel. The visual processing is organized so that visual processing in the brain begins at the back, where information is received from the eye and then continues forward through many stations, both in series, and in parallel so that as one moves forward in the brain, one can distinguish what are sometimes referred to as two streams of information processing of ventral or lower stream that proceeds along here and makes it's way to the front here in this area called TE, and a dorsal or upper stream that proceeds along here and finds its ways into the priorital cortex in an area called TG.

Now the ventral stream of information which moves towards area TE is concerned with achieving representations about the identification of visual objects; that is, what something is. The dorsal stream of information which moves along up here is concerned with achieving information, achieving representations about where something is in space visually about its location in space and its relationship to other objects and the kind of computations needed to reach out and get to places in space. Now our current understanding about memory is that memory is stored in these same distributed geographically separate locations that are ordinarily engaged in the processing and in the analysis of what it is to be remembered. That is, it's reasonable to say that memories are stored as outcomes of perceptual and cognitive processing.

Now from these areas of processing and ultimately storage, information converges into a number of different target zones or a number of different brain areas. One of those target zones includes the hippocampus, the area I've already mentioned, and this target zone is critical for the formation of stable memories. This area of convergence is critical for converting perceptions into memories.

The human brain is the most complex device known in the universe. It contains some one billion neurons, or approximately the number of stars in our own galaxy. In this temporal lobe that we were just looking at where area TE is on the inner surface of this temporal lobe is this convergence zone that I'm speaking about. In other words, if you went around under here and up on the other side, you would come to this area that we will be talking about which is important for converting perceptions into memories, as well as important for making and forming stable memories and when damage produces a syndrome that's known as amnesia.

I'm going to speak just a little bit about that syndrome because although it's a rather rare occurrence, when it does occur, it occurs in such a pure form that the study of the syndrome has informed us in a number of useful ways about how memory is organized in a normal brain. That is, just as much in biology where one can study errors as a way of learning about normal function, so it is in the case of the brain that one can study certain kinds of disorders and impairments as a way of gaining clues about the organization and the normal function. So in the case of this amnesiac syndrome then, which occurs when there's damage to both sides of the brain in this inner surface of the temporal lobe in the area of the hippocampus, what happens then is that one sees patients that have selective impairment in memory but without any impairment of other intellectual or perceptual functions. What this means in terms of numbers is that these patients will score normally on an IQ test, but these patients are scoring very poorly in a conventional memory test which asks them to learn and retain for later on new kinds of material, either verbal or non verbal material such as words or drawings. Again, this test is normalized for a score of 100 in the normal population.

So you see that whereas performance is normal in this intelligence test, it's very poor indeed on this memory test-- and what this then is telling us is something rather important and fundamental about how the brain has organized its learning and memory functions. To some extent the brain has separated its intellectual and its perceptual functions from its capacity to lay down and memory the records that ordinarily result from engaging in intellectual and perceptual work. These patients who have amnesia as it's called are intact in a number of ways other than just having normal intellectual abilities on an IQ test, they also have very normal short-term memory or telephone digit memory. They can carry on a normal conversation as long as it doesn't last too long or accumulate across too many different topics, they can repeat back a telephone number in an normal way and they can often then join us for dinner and completely appear normal except that they would then afterwards not have a normal record of what has happened.

Now many of you may say, "well we all have this problem and we all do at some extent as we grow older," but I'm talking now about a syndrome, about a condition that is far beyond the ordinary subject matter of conversations in our ordinary experiences in ordinary, in normal aging. The condition can best be described as profound forgetfulness, and it's demonstrated here in this demonstration experiment where control subjects or normal subjects and amnesiac patients are given a short prosee passage to listen to of some three or four lines of text, and they're asked to repeat it back immediately after hearing it and then again after a delay. And what one finds is that the amnesiac patients are very good at repeating back the paragraph, the story immediately, but after some 12 minutes are past, none of the patients is able to remember any of the story at all and the more severe ones will not recall that there was a paragraph even read to them.

Having laid out for you the severity of the memory impairment that's associated with amnesia, I want to emphasize just a couple things about it. First, this is a memory impairment that is first of all severe, and second of all global in the sense that it effects all sensory modality. That is, it doesn't matter whether the patient is hearing the story or reading the story or feeling it out in Braille, the information is a global deficit across all sensory modalities- a deficit that's expressed as forgetfulness. These patients will be, or are unable to live independently. They need supervisory care. They become lost without help and so on. So against that background, I think one of the most extraordinary and insightful advances to be made in the study of memory and the brain which occurred about 10 years ago now, 15 years ago was the discovery that these same amnesiac patients who are so severely impaired in most conventional learning memory tests are nevertheless entirely normal in other kinds of learning and memory tests. It was that finding that led to this idea that there must be then more than one kind of memory. The kind of memory that's impaired in these patients and these other kinds of learning and memory that are entirely intact. So let me just give you a flavor of the kinds of things that these same amnesiac patients who were so disabled in their everyday life, let me give you a flavor of the kinds of learning and memory test that they not only pass with flying colors, they actually perform as well as you and I. They are completely intact.

One example comes from the kind of task in which the subjects are being asked to learn to read mirror reverse print. These are low frequency words. It takes an average subject as much as 60 seconds to make out these words, but you find out that as you practice more and more novelty triplets, you get better and better at it. And after some time, it doesn't take you so long to read off "hypnotic," "apocalypse," and "functional." The finding of major interest is simply how long it takes a subject to answer each triplet and to pronounce each triplet correctly. So these are learning curves, skill learning curves for three kinds of amnesiac patients across three days of testing with 50 trials each day and then a retest at three months. You see that these amnesiac patients, these memory impaired patients are learning this mirror reading skill at an entirely normal rate. In addition, they retain this skill at an entirely normal level three months later, despite the fact that these patients will deny that they've ever done the task before. To them, they're doing it for the first time, and they failed very badly on tests that ask them to remember from a group the words that they've read before.

In terms of understanding, there are important issues of what parts of the brain actually are involved in giving the precious capacity for declarative memory. And as I've already suggested, a major component of this work has come not from these occasional cases of individual patients who have become available every decade or so, but rather from the kind of systematic and cumulative work that can only be done with experimental animals. In fact, one could say that around 1900 we first got our first hints that the temporal lobe part of the brain might be important for memory from cases that weren't paid that much attention to. In the 1950s we got our first demonstration that this part of the brain was important for human memory. In the 1980s, we finally were able to establish an animal model of human amnesia in the monkey, and then it became only a matter of time when the problem would fall into place. About 1990 is when we finally were in a place where we could say now we think we know what parts of the brain are important for this function. In these times I can't really understate the important and vital contribution that is being made to neuro science research of all kinds, not just neuro science work on memory, that's being made by responsible research on experimental animals. I might say for example, what did Louis Pasteur, Charles Darwin and Thomas Huxley have in common? They all lost children to illnesses during the ages between 5 and 10 years old. In the 19th Century, as many of you know, losing children to illnesses and viruses and bacteria was a common event in all families, and yet today this is a much rarer event and as a result of research and research involving animals, we can love our children more comfortably. Don't think though that now we live in the modern times and this need is past. This problem is past. But I remind you that we're in the middle of an AIDS epidemic, and cancer and heart disease are terrible illnesses we have very little clues about. We don't have an animal model of Alzheimer's Disease. We don't have any ready way of carrying out systematic and cumulative research on these important diseases, and our need for experimental animals if anything even greater than it used to be.

The monkey as I suggest has been an important part of this journey to learning about the parts of the brain important for memory. The animal model which was established in the early 1980s was to make a surgical removals bilaterally on both sides of the brain in areas of the temporal lobe being driven by our knowledge of anatomy or connectivity and then test the animals on memory tasks of the type that human patients fail, and then ask the question, what parts of these, of the brain are critical for this memory ability.

So how do we test memory in the monkeys? One famous method is a task of recognition memory or multiple choice memory where what happens on the first part of the trial is that the experimenter places an object to find--a raisin-- as a way of guaranteeing attention to other placed objects, and then a screen comes down and the objects are rearranged. On this task, which is referred to as delayed non-matching the sample, what happens is the monkey now finds the raisin hidden under the "novel" object and by choosing the novel object, the animal tells the experimenter that the animal remembers that it was this object that was originally baited with a food reward. So the animal is doing a multiple choice test. Through many, many trials of this using different objects each time, the animal can simply plot out for us, tell us about its level of memory ability.

So for example, here's a case of a group of animals that had rather small lesions in this area of the hippocampus that we've been speaking about and what you see is this typical kind of picture of forgetfulness that is the hallmark, the signature of the human amnesiac condition. That is, across these delays that go from 8 seconds out to 40 minutes, these monkeys with these lesions are rather good when the delay is still short--up to about 60 seconds--but they have difficulty in remembering things that happened after 10 minutes or 40 minutes have passed. And by doing these experiments over and over again with different lesions and different groups being guided by the connectivity of the brain, one works out in the end a sketch, of these structures here, all of which communicate with the rest of the brain. The point then about memory is if the precept of an object in space is to make it into long-term memory, that is so that after you take your mind off of it, after the information is no longer on the sensory surface we have the possibility of revivifying the image of the recollection of the memory. What we think has to happen is that this whole memory system here for declarative memory has to be involved through these connections.

The other thing though that I've tried to emphasize is that this is just one of several kinds of memory, this declarative memory, this conscious memory, and that there are other kinds of memory as well--priming, habit learning, skill learning like mirror reading which depend on other areas of the brain. Declarative memory is, one might say, an imperfect kind of memory. It's subject to error as we all know. It's subject to reconstruction and to distortions. It's subject to having disassociations between the accuracy of our recall and our confidence in how the recall is. We all have the experience of being sure we're right, but finding out that we're wrong about a fact in memory.

Other kinds of memory also exist. Skills and habits and priming and these are acquired, stored and retrieved without this system, and independently of this system of declarative memory, intact and even in amnesiac patients. These kinds of memories one might propose are more fundamental. They're essentially for survival. But in contrast to declarative memory they tend to be reliable and consistent kinds of memory. And these kinds of memories I would suggest give us much of our personality and many non-conscious ways of responding to the world, and in no small part by virtue of the non-conscious status of these kinds of memories. That is, here in these non-declarative, non-conscious kinds of memory arise the dispositions and the habits and the preferences that are not accessible to conscious recollection, but does nevertheless arise from experience and very much influence us and are a part of who we are.