Learning and Memory: How it Works and When it Fails

[Music] Stanford University. Good evening, everybody. Great to see you here again this evening. I'm pleased to say that I won't have to exit tonight to go to the pharmacy. I think I'm doing better, but uh, self-medication goes a long way. So, how many of you remember the topic of two weeks ago? I'm glad you're here tonight. Uh, so tonight's topic is on an issue that all of us are concerned about. I think I'm increasingly concerned about it as time goes by, and that is learning and memory. I remember, um, all too longingly as an undergraduate, having a kind of photographic memory and being able to, uh, really envision every page in my organic chemistry text. That's long gone, uh, along with most of the other things that come with aging. Uh, but here we are tonight to reminisce and better understand the process. And, uh, uh, I think we're in great hands tonight because we have Dr. Frank Longo, who is going to be our speaker. Dr. Longo is actually a California boy. He, uh, grew up, uh, in Southern California in San Diego, where he, uh, went, um, and did his undergraduate work as well as, uh, his graduate work as MD and PhD at the University of California in San Diego. Then he moved North, bypassing Stanford for reasons that I've never understood, um, to UCSF, and, uh, he did some great things there and was very active in what is, in fact, a wonderful neurology program. And then took a detour East, um, to the University of North Carolina, where he became chair of the, uh, Department of Neurology and did a fantastic job, uh, in developing that department. And I know it, uh, well, because I visited him there, um, to begin the process of recruiting him back to Stanford, and he did so well there that we wanted him to join us, which he did in 2006, um, to be our, uh, chair of Neurology. Among his, um, many talents and interests is the broad field of Neurology, but he has also, in his own research, been very focused on a disease which increasingly, um, takes a toll on, uh, all of us in the human condition, which is dementia. And his laboratory has been very focused on developing novel small molecules to try and, uh, deal, um, with this very challenging issue. But tonight, we've asked him to speak even more broadly, um, than that, and to talk about the whole issue of learning and memory, um, when it really works and when it doesn't. And all of us know there are times when it doesn't, but we'll celebrate tonight on the time that it does. So thank you.

[Applause]

Well, thank you, Phil, for that kind introduction. Uh, the topic tonight is very broad. This is a topic that could be approached in your philosophy course, could be approached in a psychology course, could be approached in a neurology course, could be approached in a psychiatry course, could be approached in a criminal law course. This is a massive, uh, topic. But what we'll do tonight is synthesize some of the key parts, some of the key foundations of what memory is and what learning, uh, consists of, uh, so that hopefully by the end of the session, you'll have a good idea of what it is and we'll be able to apply it to a number of fields. Now, there are broad definitions of memory and learning, and, uh, during the evening, I think you'll become an expert on it. Uh, there is no official definition. Memory is generally regarded as the ability to record new information into the brain and have it stick, basically, if you want a technical, uh, definition. Um, and learning and memory, the meaning is overlap quite a bit. Learning is the ability to retain, to take that information and respond differently, uh, to a given situation. Obviously, a key, uh, requirement for survival, not only amongst us and other mammals, but a wide range of animals, all the way down to animals consisting only of a few cells. Um, so that's the, the basic, um, uh, range of, of the definitions we're, we're dealing with. And of course, memory has been of great interest for a long time, um, probably from the very beginnings, long before written history was possible. People recognize the importance of having a good memory to survive, to pass on traditions, etc. In fact, the Greeks, um, by the time that we got to the Greeks, um, they had very sophisticated thoughts and knowledge about memory. And where does the word, uh, come from? This is the Greek goddess Mnemosyne. Nemine, and I underlined the "mos" to remember, to really put the emphasis on "mos." Nemine, and she, uh, really had the power of memory and could bestow this power on humans if they were fortunate. And she was one of the more intellectual Greek gods. In fact, she got together with Zeus, uh, nine nights in a row, in fact, and had nine children, the nine Muses. And how that happened, we haven't, we can't explain that biology yet, um, but this was, uh, quite an intellectual family, and each, each Muse had, uh, a talent. And these are the nine talents that go, um, along there. And it was recognized that having an exceptional memory would be fundamental, would be a foundation for having any of these, uh, talents. So their mother was really the, uh, source of that. Now, the other gods appreciated her skills also, so did the oracles. This was the, this is the Oracle of Trophonius. And people went to him to get wisdom, and, uh, sometimes they were happy with what he said, and sometimes they weren't. Now, if they were happy with what he said, they would drink, uh, from the spring of Mnemosyne, shown right here. They found it in Greece, and by drinking from that spring, one would remember what the Oracle had said. Uh, if one didn't like what the Oracle said, one would drink from the spring of Lethe and forget everything. Um, and these were, um, key, um, endeavors because even today, we're working on creating, uh, something that will do this, and we'll, we'll get to that. Uh, there are many things that will do this, um, that that already, um, exist. Um, and of course, uh, many brilliant people have had something to say about, uh, memory. We can look at Cicero: "Sweet is the memory of past troubles." And I thought that matched somewhat to Dante's quote: "There is no greater sorrow than to recall happiness in times of misery." So you just can't quite get it right. Um, and of course, Enrico Fermi, the famous particle physicist, said, "If I could remember the names of all these particles, I'd be a botanist." Um, so clearly, he had a lot of other cognitive skills that made up for that, that memory, uh, challenge. But pretty much every famous person has an insight on memory, and I imagine everyone in this room has considerable insight and memory. Let's jump right in and organize our thinking about memory.

Now, memory has been around since before the Greeks, but it's fascinating to me that it wasn't until Atkinson and Shiffrin, who were here at Stanford in the mid-60s, were the first, really, to organize a working model of memory that has persisted and that people could work with. Now, since then, there have been many modifications of this model, but this is generally regarded as the first formal model of memory, even though we've been thinking about memory for thousands of years. And here's basically how they organize it. Basically, all of us have a number of sensory inputs: visual, auditory, uh, motor and movement, olfaction, gustatory, etc. And in previous lectures, I know that you're now experts on visual and auditory input anyway. These, um, inputs, um, come in through our sensory systems to various parts of the cerebral cortex that you've probably seen, and they end up in a, a register, a sensory register, and only for very brief times. For example, if one looks at something, the, uh, vision might be there for half a second or a second. It's a very fleeting existence. An auditory input might last longer, four or five seconds. For example, one example that's used to illustrate the auditory sensory register is as follows: One might have a, a teenager sitting in front of a television set, engrossed in watching whatever's on television, and the parent might be talking to that teenager, talking to them about whether they've done their homework, etc., etc. And the teenager has completely excluded what the parent is saying. Finally, the parent gets upset and says, "You know, I don't think you've heard a single thing I said," and is very angry. And suddenly, the teenager sits up and can remember, in this sensory buffer, about the last five seconds. And the teenager will say, "Of course, you were just talking about my math homework," if that's what came out in the last five seconds. And that's our, we've all experienced that, that's our sensory register. If we're not attending to something, these are fleeting registrations of the world around us.

Now, if we provide some attention in that case, the parent finally got the attention of the teenager. If we can add some attention to the mix, we can move these things into our short-term memory. Now, in this stage, things can last up to about a half a minute, seconds to a half a minute. We can hold on to these things. Um, an example might be if somebody gave us seven digits, like a phone number. If we weren't paying attention at all and somebody next to us said seven digits, it wouldn't make it into our short-term memory. We at least have to pay some attention. But if we pay attention, it'll get in there. It only lasts a few seconds, um, and then if somebody, a few seconds later, said a number, a phone, a second phone number, it might displace the first one. So we're vulnerable to displacement in the short-term memory. Now, we have this, uh, 7+2 meaning. And interestingly, that the short-term memory seems to be able to hold up in most people about seven objects or seven things to remember, plus or minus two would be the range. So that's why seven-digit phone numbers for most of us, uh, we can handle. If we go beyond that, it becomes difficult, and I'll show you some, some data regarding that. But that's characteristic of short-term memory. Now, once something's in short-term memory, we've got it there for a few seconds. Um, if it's important to us, we'll want to commit it to long-term memory. We have to do something that's generally consists of some kind of rehearsal, which leads to the term consolidation. And if those two things occur, we have a chance of moving it into long-term memory. And when something is moved into long-term memory, we're actually changing the connections in the brain. We're creating new patterns of neurons firing, and we'll look at that, um, in more detail. Um, one feature of long-term memory is it's thought to be somewhat limitless. No one's really defined a, a limit on long-term memory. At least now, we have limits on our short-term memory capacity, but long-term, in a practical, um, sense, seems to be somewhat, um, unlimited, and it can theoretically last for 100 years or as long as we live, with some decay, um, with aging. Now, once we get something to long-term memory, of course, we may want to retrieve it, and that process of retrieval has been studied, uh, considerably. We'll get to that. Interestingly, we bring it back into the short-term memory mechanisms to finally, um, be able to recognize it.

We can layer on top of this original model, and this is the central executive. This is the even higher order parts of our brain that are that are managing and running all this. And this is, uh, the part that's thought to give us the ability to plan, to execute, to organize, to even organize how we might use our memory, um, and, and ultimately conscious thought. And this central executive is considered to have two interesting loops: one is the phonological loop, i.e., our inner voice when we're thinking, we're using that inner voice to think about things, or when we're thinking, we're using a visual sketchpad, so-called inner eye. These two loops are a major part of the central executive, and it's the central executive that allows us to look down, so to speak, on these memory functions and actually be conscious of this, "Gee, how do I get this to my long-term memory? What kind of tricks can I use to get it to my long-term memory?" So it's not an automatic, unconscious part necessarily, although there are automatic parts of memory, uh, that we'll get to. So anyway, this is just a basic model of organizing what people had probably seen for thousands of years, but it allows us to think about it and modify and test it and look for the biology behind it.

Another term that we'll hear is so-called working memory. Now, this term is often used interchangeably with short-term memory, um, sort of loosely speaking about memory, although experts will argue that technically working, working memory is not exactly the same thing as short-term memory. Um, working memory, uh, can draw from both short-term and long-term, um, contents. So therefore, it's not exactly short-term memory. And what's an example of working memory? This is the ability to hold something in short-term memory and do something with it. So if I say that, if I give you two numbers, 40 and 13, 40 and 13, you're holding them in your short-term memory. Now, some of you, some of you might be able to move it to your long-term memory. Now, if I say, "Take 40 and subtract 13," you have to do two things. You have to hold 40 there and 13 there. You have to hold them there in the short-term memory. Now you have to do something, do the subtraction. That's your executive function, your ability to manipulate that data that's sitting in your short-term memory there. Of course, we get much more complicated forms of executive function, but that's the basic idea.

Now, another, when we see patients in the neurology clinic, we'll ask them to spell the word "world." So, easy, I can do that: W-O-R-L-D. And then we'll say, "Spell it backwards." Now, this is much more challenging. They have to hold this and bring it out of the long-term memory. "World" is in the long-term memory. Move it into the short-term memory, change letters around. That's executive function. Hold it in executive function. Start: D-L-R-O-W. Much more challenging. And my patients with even very early forms of Alzheimer's disease will have considerable difficulty spelling the word "world" backwards because the short-term memory, this executive function, is quite vulnerable in, in early Alzheimer's disease.

Let's look further at this fascinating entity of, of short-term memory. One way to test it is the digit span. And back in the 1950s, George Miller, a very well-known, um, psychologist, was just, was characterizing short-term memory, giving people sequences of numbers and seeing how many they could remember, i.e., a digit span. And he's the one that arrived at the observation that that most people seem to do pretty well with seven numbers, numbers. So here, looking at his data, we have the number of digits that are given to the person. These are generally done verbally, saying the numbers in a real, even pace: 3, 8, 2, 7, not not grouped together. And looking at the, uh, mean, mean trials to criterion, how many times did they have to hear the number and repeat it to get a perfect score? So when his subjects were given five numbers, only one time to get it right. Six trials. Six numbers, one trial to get it right. Seven numbers, one trial to get it right. But eight numbers, uh, some people, quite a few people, these are averages here, took two trials. Uh, we move up to, uh, nine trials, nine numbers, more trials. So the more digits, more trials are necessary. And we seem to be, um, uh, peaking at about seven or eight trials. Once you get beyond that, it doesn't help much, and people really can't get much, very few people can get beyond this range. These were patients with various types of memory disorders, and of course, even for, uh, seven digits, for example, they had to do many trials before they could get that right. So this was a fascinating phenomenon. The term "Magic Seven" came out of a paper that he published in the 1950s. He actually had an interaction with Bell Laboratories, and that is apparently the reason why the seven numbers was picked for our phone number, uh, because we could remember that. Now, how about the area code? Yeah, the area code. They weren't anticipating area codes. I think the whole country might have been the same area code at that point. Um, now, but when we get to the area code, though, it actually gets us to this phenomenon called chunking. If we group things together, they're easier to remember. So by putting that little hyphen there, 384-0738, for my short-term memory, it's almost as if I'm only having to handle two things, and that's a lot easier than handling seven. And that's why that hyphen, uh, was was put there. So that's, I'm using a trick. It's called chunking. I'm putting groups together. If you ask me to remember 14 numbers, I might think that's fairly hopeless, um, but if you ask me to remember seven two-digit numbers, another two-digit numbers, another way of chunking, I could probably handle that. So again, by grouping things, we can force more things into the short-term memory. A number of tricks we can use to to get have better short-term memory. Get to some other, other tricks, um, later.

Now, let's switch over to long-term memory. And probably one of the best ways to describe the phenomenon of long-term memory is just to look at, see how we test it in the neurology clinic. Um, one of the classic scales is the Wechsler Memory Scale, and I'll briefly discuss some of the main parts of that scale. I'm not going through all of the parts. This is a very large scale. This, this, um, testing a person with this takes several hours. Um, basically, we start with orientation. Um, we ask the, the, uh, patient or the subject basic orientation questions: What's the day, the month, the time, the year? Where are you? Where do you live? How did you get here today? And we're really assessing language, communication, because if there's an issue with language, communication, orientation, delirium, it's really no longer becomes a valid memory test. Then we move on to other things. Logical memory. We'll read two short stories. We'll wait about 20 minutes, and in, in the clinic, waiting 15 or 20 minutes is our way of getting to long-term memory. Now, as you all know, short-term memories up to about 30 seconds. Anything beyond that, so we're talking minutes, hours, days, weeks, months, years, be enters into the wide range of long-term memory. So we're checking long-term memory. We wait about 20 minutes, somewhat arbitrary. We'll ask them to retell the story, and they, we'll ask several yes or no questions. So I happen to have one of these stories here. You can see what this is like. This will be a short story, um, and I, I'll give you a friendly warning ahead of time: Remember everything I'm saying about this. Um, you're in our clinic now. Here's the story: "Two semi-trailer trucks lay on their sides after a tornado blew a dozen trucks off the highway in West Springfield. One person was killed and 418 others were injured in the Wednesday storm, which hit an airport and nearby residential area. The governor will ask the president to declare the town a major disaster area." Now, we don't have time in the lecture tonight to, we could come back 20 minutes from now. Um, now, if I let you sit there silently for 20 minutes, you can maybe try to rehearse this, try to get some of it to long-term memory, or I can apply a distractor, and that means I'm going to keep you busy doing something else for 20 minutes, that makes it much more challenging. So there are many ways, many ways of doing this. So we'll omit the 20 minutes. I get, I just threw in a small distractor. Now, I'll start with a question, and these are yes/no questions. And if we had the buttons where people could press yes/no, I'd love to take a poll of the audience, but just to give you a feel. Question number one: Was the story setting in West Palm Beach? Okay. Was the storm on Wednesday? Yes. Wow, good. Did the storm hit an airport and a shopping mall? No. Okay. Was someone killed in the storm? Yes. Okay. So almost every, or perhaps everybody here, managed to move these things into their long-term memory, perhaps because I gave you a warning. And we'll get into that, how other factors feed into the ability to move things from short, uh, to long-term memory.

Okay, what can we do? Uh, what would we do next? The next thing we would do are provide photographs of 24 faces. And these are not famous people, people that we haven't seen before. We would again have our 20-minute delay or so, and then present 48 faces, 24 of them being the prior ones, and then 24 new faces, and we'd ask a simple question: "Have you seen this face before?" And we'd have a score of how many they got right. Then we would try verbal paired associates. We'd give eight-word pairs, and then 20 minutes later, we'd just give one word and say, "What went with it?" And they would be nonsensical pairs. So one pair we give is "truck" and "arrow." Now, of course, a memory trick would be to picture a truck with an arrow going through the windshield. Okay? Now, my executive function is what helped me work my memory in that way. So I'm getting more out of my memory by imposing my executive function on it. Family pictures, these aren't our own families, but just other families. There are scenes where they can be interpreted, and they are shown, and then at some time later, we ask people to recall the information that was in those pictures. Word lists. We give 12 semantically unrelated words, wait 20 minutes, and then ask to see how many words people can repeat. I mean, there are, and for each age range, all of this is age-adjusted, we would have an expectation of how many can repeat. And we'll typically give the word list four times. So here's a word list. Ready? See how many you can remember: Chimney, lemon, bus, pants, pair, train, window, hat, cellar, apple, shoe, boat. Okay, now, in the real life, we'll let you do this three more times. We'll go through it four times, really give you a good rehearsal. This is the intake now. So we're giving you a good shot at intake, an acquisition, but because what we're really testing is retention, the, the long-term memory. Now, was anybody using any particular tricks during that? Yes. Well, I first of all, I put the lemon down the chimney, but then I began to see that was going to work for the whole thing. So I began to group things: items of clothing, fruit, and places in the house. So that's, uh, two approaches, and that's what I tried to do too. One is to create a sequence: a lemon went down the chimney and ended up, um, on a bus, uh, where the bus driver was only wearing pants. Um, but you have to be quick in making the story up. Um, now, if this were, um, written in front of you, it'd be easier to make up that story, and that's association technique. The other, of course, is grouping. If there are two or three, uh, fruits or foods, group them together. So that's good. You really came up with two of the, the principal techniques. Um, if you don't have a technique, this, this can be very difficult. Any other techniques? Yes. Sketching. Just visual sketching. Sketching. Okay. Have visual sketching. That's another one. And we'll get to a famous, uh, person that was really great at memory that Ed, something like that. Visualization. Okay. Uh, visual reproduction. We'll show the subject five designs. They'll get to see each design for 10 seconds, and then we'll come back and have them draw as many of these designs as they can. Uh, spatial span, a series of, of numbered blocks in a board. I'll show you that. And finally, the digit span. You saw examples of doing that in a forward order. Most people can get seven digits forward, and then if we give them digits, they can do five backwards. So it's seven forward and five backwards, uh, is a rule of thumb. Um, so we did the word list. Um, here are, this is the spatial blocks. So what the examiner will do, these are real blocks in a real board. They'll touch, without saying anything, blocks in a certain order: 1, 7, 4, 3. They'll wait a few minutes. The last, the subject to touch the same blocks in the same order. That's not too difficult. And then they'll surprise them, say, "Okay, now touch them backwards." So a little bit of memory plus some executive function there. And all of these things can be scored. We did the short story. The Stroop test is, it gets a little bit beyond memory. This is more executive function. This is a, a fun test that, that I like to show people. Has three parts. We can all do it right now. The first thing you'll do, and we can do this silently to ourselves, but typically this would be done out loud and scored. First, we'll just look at the color and go down the list and just say the color to yourself: Green, red, yellow, blue, etc. So every, going to try it. Just say them to yourself. You could probably do more than one per second. Okay, not too difficult. Part two: Read the word to yourself. Red, green, blue. You could probably do well more than one per second. Go ahead. And if this were done out loud, we would confirm the person's not colorblind. We, we would confirm that they can read, um, English well, and that's, that's sometimes an issue. Of course. Now, we'll get to the hard part. Part three. Now, as quickly as you can, name the ink color. Red, blue, green. That's pretty good, but you're only doing about one a second. I think you've slowed down considerably. I hope that, um, I hope back here you weren't saying red, green, blue. Probably not. I think people had considerably more difficulty. Why is that? It's because the, the, the language part, the reading part, is so well-entrained in our brains, that's dominant here, and that's really the, the overriding part. Saying the color, of course, we know what red looks like. It's not quite so dominant. So we're taking that, that somewhat automatic relationship and using our executive function to reverse it. Okay, but it's requiring our executive function to do that. And in early Alzheimer's and early memory disorders, that executive function is lost, and people will take forever just to get through that first row. They'll make many mistakes because they've lost their executive function. People that lose their executive function, very likely that when we took their history, which we would have done by this point, can't plan, can't prepare a meal, can't multitask, can't get through the day, etc. But this is the way we, we validate, um, that history.

Okay, so let's organize what we've learned about memory. We have long-term and short-term memory, but we have several types of long-term memory. So we're delving deeper now into long-term memory, and these categories are critical for understanding memory in us and our society, as you will see. So long-term memory can be divided up into two main categories: declarative and non-declarative. Declarative, that's our conscious memories that we can consciously think about, consciously use, and we're consciously aware whether we remember them or not. And our declarative memory can be divided up into events or facts. Events are also called episodic memory. These are events in our life, uh, when we graduated from high school, when we got married, um, events that we remember happening, uh, where we were when something happened. Where were we when Kennedy was shot? Or the more modern, where were we when 9/11 happened? Etc. Our, our personal autobiographical journey through life. Facts are facts about the world. Who was the 21st president? Um, what's the capital of Indiana? Etc. They aren't personal to us, facts about about the world. We either remember them or we don't. Um, this includes remembering what an object is for, how to work a remote control, etc., or a laser pointer, um, or remembering a language. Now, of course, um, these are intertwined. One can remember a fact that 9/11 happened but entirely forget where we were. So, so this, this, this episodic and semantic memory are somewhat, um, intertwined, but they do tend to happen in separate parts, um, of the brain, and they're, they're distributed. And we'll get more into the brain structures, the medial temporal lobe, and the diencephalon. We, we'll point those out.

The non-declarative memory is memory we're not directly conscious of. Yes, and feel free to ask questions, by the way. The episodic memory sometimes, or at least for me, tied to, we, we'll get to that, but that's a great question. Why is episodic memory tied to a smell? And when we look at the neuronal networks, uh, that'll make sense, but it's basically coming into the hippocampus, which we'll look at. We have inputs from the olfactory region, inputs from the part of our cortex that, uh, have that memory, and they can reinforce each other in an associative way. But smell is actually one of the most powerful stimulants of pulling out memories. Procedural memory is one of my favorite parts. Motor skills are easiest to think about. They also include cognitive skills. Um, so if we think about tennis, Roger Federer has amazing procedural skills, and he is not thinking about them consciously. Um, in fact, if you talk to any elite athlete, um, when they're at their peak, um, they're not thinking about what they're doing, other than higher order strategies. But so if Roger Federer ever stopped and thought about the technique of his forehand and thought about what his coach told him about his forehand when he was six years old, it would foul up his forehand entirely. Um, these are procedural skills are in the striatum and cerebellum, the parts of the brain that are not really involved, um, with, with, with consciousness. The perceptual representation is another form of non-declarative memory. This consists of priming. Um, so we can prime certain parts of the memory, and this is tested, uh, using words and partial words. So we can show a person that has complete amnesia for events and facts. We can show them the word "table," and then come back a few minutes later and show the letters T-A-B and ask them how quickly they can finish that. Now, they have no memory that they were shown the word "table," but if they've seen "table" before, they'll put it in that, uh, very quickly. If we didn't show them "table" a few minutes ago, they might put, uh, they might spell anything. So their memory was primed, but they have no consciousness of it. It's not a direct conscious memory, and that's called priming. Classical conditioning is Pavlov's dogs. Um, so when he brought the dogs meat, he rang a bell, and then after a while, he just rang the bell, and the dog salivated just as much as if he brought the meat. So something in the dog's brains, quote, remembered, uh, that the bell was associated with meat, but it wasn't necessarily conscious thought. And that happens on the human level too, involving the amygdala, um, and the cerebellum. And finally, we can have pure reflex paths that aren't conscious. Habituation, for example, if we present, say, an animal with a tone, um, and initially, the first time the animal hears it, it'll be concerned. But if we just have that same tone occurring many times, after a while, it's not concerned about it at all. It's not necessarily a conscious thing, just it's habituated. And that's another form of non-conscious memory. We'll spend most of the evening on declarative memory. It, it has really fascinating ties to society that we'll get to, but this is a more recently really well-formulated area, non-declarative, uh, memory.

So we can take these conceptual parts of memory and begin to move them into the anatomy of the brain. And we'll be discussing quite a bit the hippocampus. This is the temporal lobe down here, and within the temporal lobe, toward the middle of the brain, we have the hippocampus, and that's really where input comes, um, to form memories. So basically, we have the neocortex, or you just say cortex for short. This is really well-evolved in mammals, especially well-evolved, um, in humans. Uh, the hippocampus goes all the way to reptiles, dinosaurs, and birds. It's a much older, uh, part of the brain. Anyway, we might get visual input coming into our occipital cortex. We might, uh, get touch, uh, sensation coming into our parietal cortex. We've got the frontal cortex, um, that's managing all this. These things, it's a two-way street. Come into the hippocampus, get processed, and go back out for long-term storage. Hippocampus really regulating short-term memory but being necessary for long-term memory. So we've got encoding, things coming in to the hippocampus. We have storage, things being sent out to various parts of the cortex. Retrieval, the ability of getting these long-term memories out of the cortex, back to the hippocampus, and back into consciousness. We'll go into more detail in all of those areas, but this is just a quick overview.

Yes, how do you deal with a memory that touches many parts of this, like playing the piano? Yeah, there's tactile, there's encoding, there's, yeah, that's an important question. How do we deal with memory that's dealing with many parts of this, for example, playing the piano? It's, it's tactile, there's music memory, there's all kinds of memory, and it's really an important concept because in the old days, people thought the piano lessons were just in one place, and people will joke, "If you slip in the operating room, there go the piano lessons." But, uh, as, as your question implies, different components are the memory are in different regions. So the memory is really distributed. This is the idea of distributed networks, distributed throughout the brain. So if there was brain damage in one place, we might lose a component of that skill, but not the entire skill. And we'll get to that more later, but that's a real fundamental question in, in memory. And forgetting is a broad term, but it can imply problems anywhere along the line, not being able to store something in the first place, or having it in storage and not being able to retrieve it. And there's plenty of debate as to, well, if you can't retrieve it, did you forget it or not? Um, and it gets into more of a philosophical debate. So things that we may never remember again might still be in our brains till the day we die, but what good does that do us unless we can retrieve it? Um, so the hippocampus, we'll use this term quite a bit. It, it comes from the Latin term. So this is a real human hippocampus dissected out, and of course, the Romans and probably even the Greeks had a separate term for it, but the Romans thought that this piece of tissue here resembled, um, what their, a seahorse. And the Latin term for seahorse is hippocampus. And this is a hippocampus. That image that we can attain by modern-day imaging of a living human, um, and so you can see the three ways of thinking about, uh, the hippocampus. This is another way of looking at the brain in terms of different, uh, memory functions. We've got the semantic memory that we talked about. Much of this is in the, the temporal lobe. The hippocampus is tucked away on the, the middle side of the temporal lobe there. We talked about procedural memory, that ability to hit a forehand in tennis would be using the cerebellum and the, the striatum, the basal ganglia. Another, the striatum's part of the basal ganglia, the putamen part of the basal ganglia. These two areas are where our motor, uh, procedural memory are located. Finally, our working memory is really the prefrontal cortex is, um, much of it here. The ability to take things in our short-term memory and do something is prefrontal cortex.

So we'll look at some case histories of amnesia. But before we do that, we really need to understand what amnesia means. So amnesia is typically often, um, associated with some kind of event. This could be a traumatic brain injury, could be electroshock therapy for depression, could be a cardiac arrest, any kind of event that's damaging or causing injury to the hippocampus, this, this center for mediating formations of memory and for managing long-term memory. So this event happens at a certain point along a timeline. So when this event happens, um, the person will have what we call an anterograde amnesia that might last days, weeks. In this example, it lasted for two months. This means for the two months following the head injury, they weren't able to lay down new memories moving into the future. They couldn't lay down new memories. It lasted two months. After two months, they regained the ability to lay down new memories. They had two months of anterograde, i.e., moving forward, amnesia. Now, they may have also had a retrograde amnesia, i.e., moving backward from the time of the accident. They couldn't remember anything for the prior six months leading up to the accident, but before that, they could remember events. So that would be retrograde amnesia. So these are two entirely different things. So these injuries damage the hippocampus. The hippocampus is quite vulnerable, and it's located in that medial part of the temporal lobes, particularly vulnerable area for many events. And so that's why we're often dealing with anterograde or retrograde amnesia. We, and different patients have different, uh, mixtures of these things, but two basic kinds of, of amnesia.

So let's look at our first case study, patient HM. Now, HM is probably the most famous patient in all of neuroscience, and perhaps one of the most famous patients in all of medicine. He unfortunately had a normal life up to, I, I believe it may have been late teens, early 20s, when he had a trauma, and he ended up having post-traumatic epilepsy. Epilepsy is not uncommon after head trauma. Many of the troops coming back from Iraq, surviving injuries they otherwise wouldn't survive, but having head trauma, are now coming back with epilepsy. And at that time, we didn't have most of the epilepsy drugs that we have now. And so in the 1950s, the field of neurosurgery came up with the idea. They, they thought that the epilepsy originated in the medial temporal lobe, that's where the hippocampus is, and that by resecting the medial temporal lobe, they could remove the source of the seizures, and the patient wouldn't have seizures. And for patients with severe seizures that were disabling, they couldn't work, couldn't function, the seizures could lead to death, um, it was, that was justified at the time that this fairly dramatic neurosurgery, uh, was justified. And so this patient HM had the procedure in the mid, uh, 1950s, and after the procedure, his, um, neurosurgeon, Dr. Scoville, a neurosurgeon in Connecticut, uh, realized that this patient, uh, was no longer able to form new memories. His life had been entirely disrupted. And Milner was a, a psychologist who worked with Dr. Scoville to characterize the patient. She actually studied this patient for some 30 years, uh, after this, and hundreds of papers have been published by her and many other people on patient HM. And so he had a normal short-term memory. He could do that seven-digit span, uh, pretty easily, uh, but he had severe anterograde amnesia. He could not, he could move, bring things into a short-term memory, but he couldn't move them to the long-term memory. He could not acquire stable long-term memory. He had about three years worth of retrograde amnesia moving back. He couldn't remember anything for the three years prior to the surgery, but before that, he had those. And so that retrograde amnesia is often graded. And by graded, we mean the, if we go back in time, the timeline right preceding the injury is most affected, and as we move back, early in our years, less and less affected, and finally, we might get five and 10 years earlier, and we can recall those events normally. So that's what we mean by a graded retrograde amnesia. That grading is important to know about when we get to the idea of fake amnesia, and we'll get to that.

So, patient HM taught the world many things. Believe it or not, prior, uh, to the publication of this paper in 1957, people just mixed memory in with all the other intellectual functions, and people couldn't separate memory from any other intellectual function. It was just one big, complicated black box, intellectual function. We knew it was in the brain. We had some idea about which parts of the brain might be important for the broad term of intellectual functions. It sounds obvious now, but people had no idea if memory was a specific intellectual function that had its own parts of the brain and its own mechanisms. The fact that this patient was pretty good at doing, had a lot of other intellectual functions independent of memory, really provided the first sense that memory is somehow separate from other intellectual functions. The medial temporal lobe, uh, was seen to be realized, uh, was realized with this patient to be critical for new memories, but that old memories, especially those prior to three years for this patient, were stored beyond the injured medial temporal lobe. And finally, one surprise was that this patient still had good visual motor learning, which is a non-declarative form of learning. The way this learning is often tested, it's interesting. People learn how to read, um, when something's held up to a mirror. It's backwards. The words are backwards, and when you first try to read that, it's difficult. You really stumble through it. But if you tried a few times, for a few days, you get better and better at it, and after a while, you can read normally, uh, through a mirror. So that, that's a learned skill. So when they tested patient HM, they thought that he wouldn't be able to learn how to do this because learning requires memory, and he had lost all the forms of memory that they had tested so far. But when they tested this, they were shocked because he acquired this as well, in fact, better, uh, than average. So that's a procedural memory, and it was unaffected, um, by this injury. So it separated out. That's where the idea was born that some memories are declarative, they really depend on this medial temporal lobe, but the non-declarative memories, to a large extent, reside in other areas. So fundamental things that we take for granted today were really derived from 30 years of publications of this one patient, HM.

Yes, did Dr. Milner have to reintroduce herself to this patient? Yeah. No, that's a, a really good question. Did Dr. Milner have to reintroduce herself? And yes, she did. Uh, in fact, I'll show you a number of other cases. I'll show you a video of that, patient HM. Um, every, every few moments, he was starting from scratch. He didn't know who anybody was around him. So anybody around him had to continuously introduce themselves. Yes. What exactly did they do to his brain? Uh, what I should have pointed out here, these are images of his brain. So we're looking at a coronal section. So if I cut my brain facing you, just like this, these are the temporal lobes here and here, and toward the middle, what we call medial, the area shown in red, those are the parts removed by the neurosurgeon, and that red area contains the hippocampus. So the hippocampus was removed, and parts of the temporal lobe adjacent to the hippocampus were also removed. The, the operative report said about 8 cm of temporal lobe was removed, and then they later on, they did MRIs on him, they decided it was only 5 cm. You wouldn't want 5 cm chunk removed from your temporal lobe, um, but that's, that's basically what the injury was. He had no more epileptic seizures, a dramatic decrease in epilepsy, but he still had some seizures, and we still deal with this issue today. We still, uh, surgery still performed on patients with seizures who respond to no medication, and the, and the seizures are disabling, but we have much better technology for controlling today. What's cut out? Yes. Was he aware that he had this condition? Or did, and also, yeah, was he aware of this condition? How did it affect his quality of life? Was he depressed? He had, apparently had remnants of memory and remnants of, in some kind of insight. He knew something happened. He knew that by all, by cooperating with all these doctors like Dr. Milner, that he, he said he was hoping to help other people. He had some of that. So he had some sense that something happened that rendered him, uh, less functional. But apparently, he was not depressed. He apparently functioned as well as he could, not being able to form new memories. I'll show you a video of another, another, another patient that really addresses your questions. Now, so do we ever have a case? Um, so the first patient HM had a lot of anterograde amnesia and a significant retrograde amnesia, and

In most head injuries, most injuries, there's more anterograde amnesia than there is retrograde amnesia. In other words, people have considerable difficulties with most forms of injuries of laying down new memories. It would be unusual for somebody, as a result of an injury, to show up and say, "I have normal anterograde memory, no problem laying down new memories, I.E., pretty much day-to-day normal life, but yet I can't remember the past. I have a pure retrograde amnesia."

Now, this is very common in Hollywood and in movies, and we look at that, it's very common in criminal courts. Um, so it's really important to know, can it happen? If it can, how often? Well, anything in neurology can happen, at least rarely. Here is a rare case of what we call dominant retrograde amnesia, where following the injury, it was the retrograde amnesia that was much worse than the anterograde.

This was a 41-year-old gentleman who fell off a horse, completely normal prior to falling off the horse. By age 45, four years later, by then his wife couldn't take it anymore. She brought him, uh, to physicians and reported that he had lost, and she didn't use these terms, basically what she was dealing with, that he had lost most of his episodic and autobiographical knowledge of his life. He just couldn't recall much of anything prior to that injury in his life, unless he went to the severe, uh, remote past of childhood. His difficulty forming new memories, I.E., the anterograde amnesia, was relatively minimal. So clearly an atypical pattern.

And sure enough, when they looked at his scans, the, the hippocampus, here's the temporal lobe. The hippocampus is in the medial, I.E., toward the middle temporal lobe. The damaged part is here where I've put the red circle, and these are just two different levels. His injury led to damage primarily of the lateral, not medial, lateral temporal lobe, somewhat of an atypical injury pattern, largely sparing the hippocampus and the critical medial part of the temporal lobes. So that might explain why his anterograde, I.E., ability to form new memories, was relatively spared, and it was this unusual pattern of having more difficulty with past memories. The larger parts of the temporal lobe are, are critical for dealing out, digging out those past memories.

Yes, could you define what you mean by damage? Well, in this case, with head damage, there would be bleeding, be a contusion in the brain, which is another word for bruise. Um, with head injury, the small blood vessels in the brain rupture, so we have blood in the brain tissue. The neurons are, are, are moved around by sheer force, and sometimes their axons, the connections between the neurons, actually break off the, the, uh, neurons. So the neurons are no longer able to communicate. Those would be the two primary mechanisms of brain injury with head trauma.

Yes, could someone in that situation learn, learn what his past had been and sort of recovered in that way? To a limited, very limited extent, um, yes, but there wouldn't be enough connections to other things. It'd be a very artificial way of learning, and they'd still be functionally impaired.

Was there another question? Yes. Um, are you going to get into later how memory is laid down in the neurons, similar to, yes, memory in a computer? Yeah, the question is, how are memories laid down in neurons? Can we compare that to a computer? We'll get to that. So we're gradually working our way down to the cellular or molecular level, starting with the big picture.

Interestingly, uh, when they were testing his episodic memory, they were asking him, in semantic memory, they asked him a question, um, "Where is Mount Kilimanjaro?" And apparently for most people, it's, it's a difficult question. There might be some people here who've climbed it, and most of us know it's in Africa. But what country it's in, and it's largely in Tanzania. I guess it's right on the border with Kenya there, but it's accessed apparently largely through Tanzania. And it's interesting, his wife, you know, let the physicians know that he'd actually climbed Mount Kilimanjaro before. When he was asked if he had ever been to this mountain, he said, "No, never, never seen it, never been there." But when they asked him where it was, he knew exactly where it was.

Um, so it's a, he had a dissociation. Having climbed it gave him a procedural memory, an unconscious procedural memory, but his semantic, uh, or or episodic memory, his episodic memory of actually having climbed it, entirely wiped out. No recollection whatsoever of having climbed it. So these injuries teach us, um, referring to the earlier question, that memory is broken down into different components, and these components, is these components, are distributed throughout the brain. With an injury, we might be limiting one component and not the other.

Now, here's a more recent patient that, uh, addresses some of the questions that a few of you just asked. This is the story of Clive Wearing. He's born in 1938, a very well-known musicologist, as a conductor, a keyboardist in London, and was one of the top musicologists and and conductors in in England in the 1980s. In 1985, one day he came home with a headache, and by the fourth day of this headache, he was no longer able to recognize his four-year-old daughter. And he went on from there to lose all of his anterograde and retrograde, uh, memory capability. He lost all ability to form new memories moving forward. He lost his retrograde memory going back 20 years to 1965. Before 1965, he could tell you things. After that 20-year gap, so he had herpes encephalitis. So the herpes virus, fortunately rarely, that just about everyone carries, occasionally tracks back through the nerves into the brain. It has an affinity for the temporal lobe, and it can wipe out the temporal lobe, including the hippocampus. Causes death 50% of the time. 50% of people survive. Many of the survivors end up with damage to the hippocampus and temporal lobe. He's one of them.

I'll show you a video, and I want you to note a few things: his interactions with his wife. This is his second wife. He forgot his first wife, um, and he forgot his, the names and a lot of features about his, his initial set of kids. I don't think he had children with his second wife. Um, he has a diary that's fascinating. Um, and we'll see a distinction here between episodic versus procedural memory. This is him sitting at the keyboard. Oliver Sacks wrote an article on him in the New Yorker in September 2007. That's where this picture's from. There's been a movie made about him. His wife wrote a book about him. There's been quite, quite a bit of this. If, are any of you familiar with Clive Wearing? Maybe three or four people. Just wanted to get an idea. Okay, let's watch. This video is four or five minutes here.

"You know how we got here?" "Don't remember sitting down." "No." "I reckon we've been here about 10 minutes." "At least." "I have no knowledge." "My started working now." "And what I've seen the whole time I've been seeing any all is that." "And do you feel absolutely normal?" "Not absolutely normal, no." "Completely and completely confused?" "Confused, yes." "You never eaten anything?" "Never tasted anything." "Never touched anything?" "Never smelled anything." "What?" "Right." "You assume you're like, but you are apparently." "Yes." "But I'd like to know what the hell's been going on."

Clyde was an outstanding magician. He would take his work very, very seriously. At the same time, he loved music so much that he just really threw himself into it totally. Clive Wearing was a musician of enormous integrity. He was the world's expert on, on Lassus, one of the great composers of, of the Renaissance, and he also worked a great deal in contemporary music and was chorus master of the London Symphony, which is Europe's foremost group. Music flowed out of him whether he's singing or playing or conducting. Clive Wearing, through a cruel twist of fortune, shows us how fundamental consciousness and memory are to our lives.

"So how are you feeling this morning?" "I'm conscious for the first time." "It's the first time I've seen anybody not been conscious before." "No, I've been here before." "This?" "I haven't seen you before." "No, you haven't." "No, I've not seen anything at all before." "Been completely blind the whole time?" "No taste at all." "This is the first taste I've had." "That's the first coffee you've had?" "Yes." "Cheers." "Cheers." "And how are you seeing things normally?" "The first time." "Colors?" "Yes." "I've never seen anything the whole time I've been ill." "No, black and white, nothing." "Do you remember me arriving?" "No, I don't remember any arrival." "I don't remember writing any of that." "Nothing to do with me consciously." "Me unconscious writing." "What are you doing here?" "Well, I didn't any previous, but I mean, that's the first one time I've seen any cars." "It yes." "Do you remember starting again?" "It all started with a headache."

Clive came home one day and said he had a very bad headache. The headache didn't lift, it didn't respond to analgesics. By the fourth day, he developed quite a high fever, and on the evening of the fourth day, for a little while, he forgot his daughter's name. By the fifth day, he was very delirious. Clive suffered from viral encephalitis, which has led to the damage of the left and the right temporal lobes, plus a good portion of the left frontal lobe. And the temporal lobes contain a structure called the hippocampus, which we know is implicated in memory function, and in Clive, it has almost certainly been completely destroyed on both sides of his brain. It's this that is primarily responsible for his severe memory impairment. In addition, the damage to his frontal lobes also causes a number of additional memory problems, which are manifest mostly in terms of him repeating himself a lot and generally showing, um, emotional, highly emotional behavior.

Clive's world now consists of a moment with no past to anchor it and no future to look ahead to. It is a blinkered moment. He sees what is right in front of him, but as soon as that information hits the brain, it fades. Nothing makes an impression. Nothing registers. Everything goes in perfectly well because he has all his faculties, his intellect is virtually intact, and he perceives his world as you or I do. But as soon as he's perceived it and looked away, it's gone for him. So it's a moment-to-moment consciousness, as it were, a, a time vacuum, and everything before that moment is completely void.

And so, I just want to make one, you should be getting one key point. His view of the world is limited to that short-term memory capability. Remember, we said it's 15, 30 seconds. So take that 30-second window, move through life like this. That's what he's doing. If it happened before 30 seconds, gone. Um, so that's, that's what his wife's doing a great job describing. Feels as if he's awakening afresh the whole time. He always thinks he's been awake for about two minutes, and that's why he looks at his watch all the time to record it, to record the fact, "Ah, I've woken up. I'm, this is an important event. Therefore, I will write it down in my diary." So he writes, "11:54 AM. I am now completely awake for the first time." And he underlines "first time."

Patience begins because he's always playing patience, and the whole diary, every page is a succession of entries saying almost the same thing of first awakeness. And when he goes back and looks at his own entries, he doesn't acknowledge that they are genuine. He says he knows it's his handwriting, but as far as he's concerned, he was unconscious when he wrote them. So he, he quite often, he'll, he'll score out what he's written before. And so his life is an ever-repeating moment of first awakening.

The strongest thing in his life, I believe, and his diaries bear that out, is his love for me. And that's absolutely raw. And each time I walk into that room, it is as if it's the first time he's seen me for years. "Good. Hello." "You surprised to see me?" "First time I've seen anybody." "A draw?" "First person I've seen you." "You have not seen me before this morning?" "No, I've been here you or six times." "You walked in that day." "The first time I've seen anybody been conscious before." "No, I've been here before." "This?" "I haven't seen you before." "No, you haven't." "No, I've not seen anything at all before." "Been completely blind the whole time?" "Don't remember me arriving at all?" "No, no, I don't remember that you arriving me at all." "No, no, I don't at all." "Look, I bet you if I look to see what you've written, mention me on that page." "You've mentioned me on this page." "My first thought, I adore Deborah for eternity." "That people's entries in the diary are rubbish." "What does that mean?" "No, do you did you write that?" "I've no conscious." "Conscious me now for the first time." "But it's, is it your handwriting?" "Yes, it is." "I know nothing about it at all." "So how do you think it got there?" "I don't." "I presume the doctors don't know." "But you must know." "I haven't." "Who's listening to me?" "Please don't say." "When I say no, I mean exactly that." "I just haven't seen the book at all till now." "No, I'm all." "I've said no." "That means I haven't seen it." "I have no knowledge of it at all." "That's all." "There's no knowledge of that book at all." "It's entirely new to me." "Just use your intelligence." "That type of comment, which is intelligence, but you put, who would put that?" "If I don't know." "No, no." "Oh, I haven't said use intelligence for mistake." "I funny." "We use your intelligence."

Clive gets extraordinarily angry, and who wouldn't? Because here, you're not dealing with somebody who is demented. So just to make one point, this anger part of it might be frustration, but the, the herpes virus infection also damaged part of his frontal lobes, and when we see patients with frontal lobe damage, they often have lack of emotional control, and we call this emotional spillover. So there are multiple, uh, components here. Remember to be angry. Well, we'll get to that.

So there's declarative and non-declarative memory. His declarative memory is really dependent on that what was damaged, but as we'll talk about, his non-declarative memory, many of those things are intact, and that's part of non-declarative memory. In fact, he can still play the piano extremely well. That's non-declarative. That's procedural. That's that procedural non-declarative memory.

"Yes, well, the 100 times every day." "How does you remember you got right?" "Um, you know, I don't know." "We can only guess." "Maybe he's reinventing that." "I don't know." "These subjects are very difficult to study because it's might not be 100% of, of his, of his memory, of his anterograde memory. He might be forming scant new memories here and there. It's just very difficult to detect them."

"Yes, I'm not mistaken, they tried taking the book away and he never asked for it." "Yeah, that I didn't know that." "So, um, but yeah, so maybe it's available." "He, he thinks about it, and if it's not, he won't occur to him." "Who is oblivious?" "Who is gar?" "You are dealing with a perfectly lucid, highly intelligent man who has been robbed of knowledge of his own life, and he feels deeply humiliated to be put in that position, very, very frustrated that he can't grasp what's wrong with him because even as you're telling him, he's forgetting the previous sentence."

"Okay, so, so any other points that people want to make?" "So yes, is there a difference between the functions performed in the right hypothalamus versus the left?" "Is there a difference between the right and left, um, hippocampal functions?" "Not big differences. I mean, in the parietal lobes, vast differences. Language on the left, spatial on the right, etc. We'll see examples of differences in memory. Differences in memory do exist for the frontal lobes. I'll show you that. Not huge differences for the hippocampi, though."

"Yes, I'm related to this. I was going to ask one of the features of accidents in the news these days a great deal are concussions. In football, I wonder you never read of any failing memories out of potential." "Yeah, the question is, what about concussions? We hear a lot about those." "Concussion injury has a wide spectrum from mild, moderate, severe. With moderate, severe, there can be memory loss, and, and some of the elements of this can occur. This is, this is the most Clive Wearing is the most severe characterized case of memory loss in medicine, to date. So you're looking at extreme, but the extremes teach us a lot."

"Yes, I'm going to go across. Does the memory reset, or is it like a lag that keeps moving?" "He's basically as a moving window, as a moving window, right? So he can only look back about 30 seconds. So if you disappeared for 30 seconds, walk back in, starting all over. Who are you now?" "He does recognize his wife, and that might be part of non-declarative memory. I don't know if you noticed in the diary there, he says how much he loves his wife. That's his main constant. That's the one solid thing he has. He recognizes his wife, and he tells her that he loves her. I guess part of non-declarative or procedural memory in that case."

"Um, you know, which is interesting. He can sit at the piano and perform flawlessly. Now, people have asked, is that if that's pure procedural memory, is he like a machine? Is like an auto, you know, automated person at the piano? But apparently there's all of the full emotion of an elite musician at the piano, and the emotion's different every time. So again, the non-declarative, much of the non-declarative memory is intact. Again, the injury is teaching us that the brain separates out these different modes of memory."

"Yes, are language functions just in a totally different part of the brain, or are they part of non-declarative?" "Or language, different part. Frontal lobe, and a lot of non-declarative, yes." "Years ago, when they did, I thought they did them in the front, temporal lobe, which is the same area that I thought he was." "Well, so you're asking an important question. Years ago, they did lobotomies. That's largely separating frontal lobes from the rest of the brain. Um, he had some element of frontal lobe damage, but not his entire frontal lobes. His primary damage were the medial temporal lobes, including the hippocampus. So he's not a full frontal, um, lobe type of patient. The full frontal lobe patients can't execute, can't plan, can't really do anything, and that's why frontal lobe lobotomy is so disabling."

"Yes, when you say he can still play piano, there are a couple of things." "He can still play from memory music that he learned 20 plus years ago." "Yes." "He can still read music." "From the music, yeah." "I'm not sure I'm thinking still do that." "Yeah, I, I think he can, but I don't want to say I'm not, I'm not sure." "I would wonder if he could read music and then later play the music that he recently read." "Yeah, my guess is if he read new music beyond 30 seconds, he wouldn't be able to play it. So maybe he can still has the procedural memory of how to read music, but I doubt if he could do anything with it 30 seconds later reading. But as he's going along, he can do, yeah, each measure at a time, certainly." "You measure, well under the 30 seconds." "Yes, yes."

"Do his dispositions change? So is one day he happy and the next day sad, or is it pretty much much a constant disposition?" "Well, apparently for the first five years or so, he was quite depressed. He had some insight that something horrible happened and he no longer could do what he wanted to do. And then, as beyond that, apparently his mood improved to somewhat of a normal mood, you know, given, given the situation."

"So I'll take maybe one or two more questions, but we have a lot more to go." "Yes, is he still alive?" "I think so. Patient, hm, recently died, but Clive Wearing's still alive as far as I know." "Okay, so we're going to go, um, into imaging and a lot of other."

"Many people have commented on memory. Tennessee Williams said, 'Life is all memory except for the one present moment that goes by you so quickly you hardly catch it going.' And this really, Clive Wearing experience, you know, reminded me of this quote. So we've seen a handful of patients, and there have been many more reported in literature. No two injuries are the same, but these patients with injuries have taught us quite a bit about what part of the brain might do what. It's an indirect way of learning. And in the modern era of imaging, we have a whole second chapter of learning about where in the brain memory works. And I'll show you a few examples of this, and then we'll get down to the cellular level of really how does a neuron capture memory."

"So we get back to this idea of distributed networks. We've got the hippocampus down here. We've got the, the region around the hippocampus called the parahippocampal region. We're still in the temporal lobe. And finally, we have these connections that are reciprocal, going back and forth between various neocortical areas throughout the cortex. And it's these networks throughout the system that are regulating memory. Sensory comes in, comes down to the hippocampus, gets integrated during short-term memory. If that occurs successfully, can go up to long-term memory. If we want to recall it, comes from long-term memory back down through here and into our, our consciousness."

"What does all this mean on scanning? Well, we're still discovering the networks that exist in the brain. I mean, just looking at the brain anatomically tells us a lot about structure, but we really know very little about the functional networks within the brain. And in fact, much of the exciting work is being done here at Stanford. Mike Gracious and Benon Menan in in neurology and psychiatry really collaborated doing advanced processing of functional MRI, uh, in a way that hasn't been done anywhere else in the world to really derive networks in the brain, neurons that are working together, and networks that have not been previously noticed. And for example, here's one. Here are the hippocampi, pair of hippocampi. They're connecting to the cingulate gyrus, another area thought to be important for memory, and that's connecting to the frontal lobe here. So this is just one of many networks that they've just recently identified in the past few years that are telling us a lot more about where memory is in the brain."

"We can go beyond just looking at the lesions, these messy lesions that patients have, and we can use functional MRI to see what parts of the brain are activated. We can put patients on the scanner and have them do tasks and see which parts of the brain are activated during MRI. Um, so this patient was given 40 nouns and asked to, uh, remember them. And during the encoding, when they were trying to commit them to long-term memory, the encoding, we can see the hippocampi lighting up on, on both sides. Then they were given a five to 10 minute distractor task, so we would see how much gets committed to long-term memory. And then after that, we're in the long-term memory phase. Now, they were given 32 of the old nouns, 16 new ones, in a randomized way, and they have to identify whether this noun belonged to the first set or not. In other words, they're retrieving, uh, from their long-term memory. And during this retrieval, we see the hippocampus lighting up. This side somewhat more, this time somewhat more so on the left, but it's not clear that this as is necessarily significant. These are coronal sections of the brain, cut like this. If we do a sagittal section, cut like this, this is where the hippocampus is right here, lighting up, uh, overlaid right here. So it's an example of how we're using this technology to see what parts of the brain are, are lighting up in response to, uh, forming memories or retrieving them."

"Um, this time, instead of concentrating on the hippocampus, we'll move to the, the frontal cortex. And again, we can look at that during encoding, during the attempt to remember words, we see this is left and right, a lot more activity in the left versus right frontal cortex. Remembering namable objects, more on the left, but bilaterally somewhat. Faces, interestingly, more on the right compared to the left. So an example of how we're learning what parts of the brain are important for encoding different things. So memory is not just all one big thing. Words are a little bit different than faces, um, etc. And that's why that finally explains why in the neurology clinic, we can have a patient whose memory is entirely intact except they can't remember a face if their life depended on it. It's called prosopagnosia, the inability to remember a face. Everything else is intact. It's probably because there's part of the brain that's responsible for faces."

"Yes, um, one question. So do we all, let's say, bird, have the same potential for memory in the same memory area? You do it often talk, people say, 'I'm terrible with names.' Is that number one, is that sort of a myth more than anything? Number two, is it true that we all would have the same potential for remembering faces as names, or or is that, is that a developed time?"

"Right, so the question is, do we all have similar capabilities across these different domains of memory, or might some of us truly be better at things than others? I think there is variability, and I'll show you a case in a few minutes that I think each of those domains can be, um, superior or inferior, and they can, that can happen differentially in different people. It's probably because different genes are expressed in different areas, and like many other genes, we have slight variations in those genes, and those genes encode proteins that are important for memory. So I think there is some variation across people."

"Yes, I have a problem, and I guess that would be the names about pronouns you use to talk, right? So you had a friend that couldn't remember nouns." "Um, that can happen. You can have people that can't remember names of vegetables. They're fine with fruits, plants, trees, just cannot remember the name of a vegetable. Could be a noun, could be anything. These are very segmented categories in different parts of the brain, which, which is just fascinating."

"Frank, what about, you know, most of, yeah, yeah. So Dr. P was asking about developmentally, why can't most of us remember, you know, episodic events from when we were two? And that we'll get to that when we look at the cellular molecular basis, but, just the myelination alone, you know, the insulation around the nerve fibers is so critical for their, uh, conducting ability, and we have so little myelination that these networks probably depend on adequate timely conduction, and the myelination is just not there. The insulation, or the myelination, really doesn't complete until around age two, two and a half. That's the main theory of why teenage judgment is so poor, because the myelination is last in the frontal lobes where our judgment lies. So, okay, let's keep moving. We'll run out of time here."

"This is just another imaging study. This is an associative memory paradigm where the subjects are asked to associate two things and to remember that association. And interestingly here, the, the hotter colors are, are parts of the brain that picked up their activity during this association. Here we've got the occipital cortex where the vision comes in. Carla Chat's probably talked about this, of course. When they're reading something, this lights up, and then other areas, brain, parietal, frontal areas, temporal lobe, hippocampus, light up for this association. The cool colors, interestingly, and this was the surprise here, are decreases in activity. So really, we're looking at a complex situation. For certain parts of the brain, get more active, certain parts get less active. So this complicated changing of pattern activation here."

"Here's an example of long-term memory retrieval, long-term meaning, I.E., a day, one day later. So the first two days, they're, they're encoding, they're trying to remember pictures or sounds. While they're memorizing pictures, occipital lobe, vision lighting up. While they're trying to remember sounds, inferior, superior temporal areas, inferior parietal areas lighting up. That's where sound goes. One day later, they try to recall it. Interestingly, a part of the same area that lights up for pictures is that we used to memorize it. And during recall, part of the areas that were used to remember the sound light up to recall it. This is fascinating. This REM, this suggests that memory is really, is a replay or an activation of the same areas that provided input to the hippocampus during the encoding. So during the encoding, I.E., the remembering, the hippocampus organized, working with the cortex, organized where the cortex would record this. So this network activity, and the cortex is the recording in the cortex, and we remember it a day later, we're activating some of that cortical function to remember it. It's almost as if we're just playing it back. Um, so we really are, uh, taking a previous time and playing it back at the present time. The hippocampus is in charge of organizing all that. If we damage the hippocampus, we can't access that playback. We can't perform that playback."

"This was an interesting study that involves sleep. These, uh, subjects were shown 320 pictures, and then they were asked to go ahead and take a nap. Some napped more than others. They didn't know why they were taking a nap, but most of them took advantage of it. After the nap, they were shown new 80 new pictures, and then scans were done as they tried to recall which pictures they had seen before or not. So they were shown 80 of the old pictures, 80 recent, and 80 new, and they were asked to tell me if, if the picture you're looking at was in the first set of 80. So clearly, you'd have to remember these first pictures from this first 320. So if we looked at these recognition performance, how well they did versus how much slow wave sleep duration they had, from 0 to 50 minutes, the more sleep they had during this nap, the better they did on this picture recall. So direct evidence supporting the recent hypothesis that during sleep, the hippocampus and the cortex are working back and forth to solidify these memory patterns. And if we don't get enough sleep, that, that solidification just doesn't work as well."

"Um, and during the, um, uh, on day three, fairly early on, when they try to remember, we can see the hippocampus is quite active. But what happens when they waited for three months? They looked at activity in the prefrontal cortex, out in a cortical area, and in the hippocampus. So during the 90 days after this experiment, the activity in the frontal cortex got higher and higher and higher, while at the same time, the activity in the hippocampus lower and lower and lower. It's as if during those three months, the memory is gradually migrating from the hippocampus up to the cortex, supporting the theory that the psychologists have had for many decades, but now in the, in the imaging area, we can see if the psychologist guesses were accurate. And so it's as if we're creating hippocampal space for new memory formation. Remember, we said this hippocampus is somewhat limited in capacity. So by moving things to the cortex, which apparently, as far as we know, is unlimited, we're making room, uh, for new memories. And the more we sleep, the better this happens."

"Now, what about lie detection? And there, you know, recently people have said, well, maybe if we can image different networks that light up during memory, we can distinguish an accurate memory versus a deceptive report of a memory. Um, so in this study, they showed the subjects, uh, one face, and then came back, a second examiner came back and showed them three faces, somewhat similar, it's not the same guy. And the, the second examiner asked them, 'Have you seen any of these faces before?' And the subjects were allowed to, the subjects never know what the purpose of an experiment is, generally, that's important, but they were given permission to lie whenever they felt like it. So they might have said, 'Never seen any of them.' But, but in their mind, they knew they had seen this guy, and they knew it was a memory study, so they probably were proud of the fact that they remembered it, they saw this guy. So that was in their mind, but they told the examiner, 'Haven't seen any of them.' And so they took scans for each report, and then they divided up into the scans, they went back later and separated out the lies versus the truths, and the difference in scanning is indicated by their color, and red. So, in other words, there were different brain responses to functional MRI of whether the person was telling the truth or a lie."

"Now, they only could find this difference if they averaged 20 or 30 subjects together. It doesn't work that well if it's just one subject, and that's why this is by no means an accurate technology to detect lies. But that won't stop commercial application. So you can pay for this, just as you can go out and get a private, uh, traditional lie detector test. You can hire somebody, um, to do a functional MRI and tell you whether you're quote, or whether the person is lying or not. Um, it's thought that maybe in some number of years, this will be perfected. It'll be interesting when that happens, and the courts and, and the legal worlds are already, you know, assessing this technology."

"I want to introduce the idea of schemas, and that's these are pre-existing knowledge structures. And these were been speculated to exist in humans, but no one had any idea if they existed in in animals. Now, explain the animal version, we'll come back to the human version. So what's a, what is a pre-existing knowledge structure? In this experiment here, rats were, um, uh, introduced to a food flavor. It might have been bacon. Apparently, rats like bacon a lot. And then these are little sand, little holes in this chamber filled with sand, and you could put a little piece of bacon under the sand, and that would prevent it from being smelled from a distance, apparently. And then the rat is introduced into one of three, one of four doors to go into this chamber, and it'll go around exploring, sniffing everything, including each of these sand pits, and it'll randomly happen on the one with bacon, and that's a good event, uh, for the rat. And then you could take the rat out, wait, let some time pass, leave the bacon in the same place, and the rat will again go around sniffing pretty much everything. Now, after enough trials, five or six trials, it'll finally get somewhat more efficient, just going directly here, or perhaps only going here, then here, you know, less mistakes, getting more and more efficient. It's learning, but it's taking five or six trials to learn the location. It's committing the location to memory, likely, and all of this can be quantitative. Now, that wasn't too surprising. The interesting part of the study, the surprise that's why I was reported in science just recently, is you could, once that rat got pretty good at going directly from here to here, so learned it, you could take that rat several days later, or a long time later, give it an entirely new food group, watermelon or something, put it in an entirely different position, and it would only take two trials to learn it, much more efficient next time around. And this speculation was, it had, by that first round, it had created for its first time, it never done this, had it never had this experience before, a pre-existing matrix to think about this situation. And once it had that pre-existing matrix, its ability to acquire a new long-term memory was much more efficient."

"What's the human equivalent of that? Well, I, I barely know how to golf, and I want, I went golfing with one of my friends who's, who's an outstanding golfer, and I suffered through 18 holes. My goal is just to survive and see if I could make it occasionally in under 10 shots. And typically, somebody might make three, three shots to get in a hole. And so afterwards, it was a blur to me. I remembered two or three of the holes, a few of my shots, and I was amazed that the friend of mine remembered every single shot, where it was, where it went, exactly how many he had for all 18 holes. He doesn't have a great memory, but he has a pre-existing matrix for for that activity. He had something, he had a, a matrix to put those new memories into. I had nowhere to put those new memories other than frustration. For me, my matrix is medicine. I recently saw a patient, and they brought a 500-page stack of old medical records, and, you know, it took me about 15 minutes to go through it, and after doing that, I could tell you what medications are on all their medical problems, etc. If somebody gave me a 500-page legal brief, I could probably spend a week on it and tell you absolutely nothing about it because I don't have any pre-existing matrix for that. So these pre-existing matrices support the rapid acquisition, the rapid creation of long-term memory. In fact, some people think that's so powerful that might be one of the few times that an input could go straight to long-term memory, skip the whole short-term memory thing, if you have a very robust pre-existing, uh, matrix."

"So let's get into the, uh, popular culture. Memory is so important that it pervades everything. Yes, develop, develop very positive things. Yeah. Is there a body of work or an organized way to develop a pre-exist, an existing matrix, matrices? Not that I'm familiar with, but maybe just going to school for endless years, years, is is a way of doing that. But I think it's a good question. There might be more efficient or focused ways of doing that for whatever task you're interested in, and there probably are informal ways that that we do that. Um, so it's a popular culture is a very useful vehicle for thinking and learning about memory. So it's not just entertainment. And you can actually look here at the top 20 grossing by grossing opening weekend, uh, movies that have a theme of memory. And the, the Bourne series plays big in that. In fact, there's Jason Bourne right there. And these generally involve retrograde amnesia without anterograde amnesia, and as we've learned, that's extremely atypical, but it fits the plot. It's a good plot device. Now, occasionally, somebody will get it right. In Memento, I think did pretty well. Let's take a look at that."

"So this is a gentleman, his wife was murdered, and during, while he was trying to protect her, he suffered a severe head injury, and now he's trying to figure, and he has a form of amnesia, and he's trying to figure out who killed his wife." "Whoops, let me see if I can get this to work here." "I have this condition." "The condition?" "It's my memory." "Amnesia?" "No, no, no, no, no. It's different from that." "Okay, let me stop them right there. They got that wrong." "He has a severe anterograde inability to form new memories. We all learn that's anterograde amnesia." "In Hollywood speak, because he can remember his, I'm sorry, he's having trouble forming new memories." "He's, but in Hollywood speak, amnesia usually means just loss of old memories." "That's why I'm saying he doesn't have amnesia. It's different than amnesia." "But this is amnesia. Since my injury, I can't make new memories. Everything fades. If we talk for too long, I'll forget how we started. And next time I see you, I'm not going to remember this conversation." "What's the last thing that you do remember?" "My wife." "That's sweet." "Dying, I guess." "I've already told you about my condition." "Oh, well, only every time I see you." "You don't remember where you've been or what you've just done?" "No, I can't make new memories. It's like waking. It's like you just woke up." "When you find this guy, what are you going to do?" "I'm going to kill him." "Maybe I can help you find him." "Are you sure you want this?" "My wife deserves vengeance." "Do not trust her. She's going to use you to protect herself." "I think someone's trying to get me to kill her own guy." "You can question everything. You can never know anything for sure." "Teddy, don't believe his lies." "You wander around playing detective." "Well, maybe you should start investigating yourself." "Who did this to you?" "You did." "I want my life back." "Why are you asking me?" "I can't remember what I've done." "I have no short-term memory since my injury. I can't make new memories. Everything fades. If we talk for too long, I'll forget how we started. And next time I see you, I'm not going to remember this conversation." "I don't even know if I've mentioned before." "I've told you this before, haven't I?"

So he had traumatic brain injury, and he can't make new memories. That's pretty good. So it's not the typical Hollywood where they just have no trouble making new memories, but forget everything in the past. So those tattoos, they, he, those are, um, he's using tattoos as notes to, so he doesn't lose his notes. So he's taking pictures, putting notes on his body. That's his memory, just like Clive Wearing's writing all these notes.

"Okay, uh, what about real versus fake, uh, memory loss?" "We are running out of time. Is a hard stop at 8:40?" "Is that, um, 8:20, I meant to say." "Okay, so we'll see. We might have to prioritize a few things." "No, um, once you get past age 40, there's a few things." "So, so how do we know if somebody's faking or not? And if you've had the pleasure of seeing patients brought from the jail, which I have, when I worked up at San Francisco General, this, this becomes an important issue."

"So this is Anna Anderson. She claimed that she was Anastasia Romanov. So Nicholas, Tsar Nicholas II, and his family were all killed in the early 1900s, and there was a rumor that the youngest child, daughter Anastasia, escaped and might still be alive somewhere. So Anna Anderson in Florida said, 'Well, that's, that's me.' The only caveat was that she forgot the entire period of her life that she was with the royal family. So, I.E., selective retrograde amnesia, but no anterograde amnesia. She was doing fine in, in the current life. Again, highly atypical pattern. So people have collected these kinds of patients, and these are the atypical patterns that are thought to go along with fake, uh, memory loss. No anterograde amnesia. We've talked about that. A memory score poorer than chance. You can actually fake people out and do this testing so that you can actually be poorer than chance would predict. Um, inconsistency, face versus events. If you really do have this retrograde amnesia, generally affects both of these to some extent. Could be differential, but you'll have people that will, um, forget events, typically, especially during the, the crime they committed, but be able to name faces, um, easily from that same time period. Inability to recall one's name. We can wipe out almost everything, and most people will retain the name. Lack of imaging evidence for brain injury. With the modern imaging, even mild memory loss, we generally can see something, um, on the MRI scan. And there's a frequent premorbid psych history. For a more typical focal retrograde amnesia that's real, we, we generally, brain injury by imaging, there's a graded retrograde amnesia. The amnesia is more severe right up before the injury, getting less severe as we go back in time. There's generally some kind of some degree of mild anterograde amnesia accompanying the retrograde amnesia, and the, the deficits are more consistent. That's how we look, but you never know 100%. Still a difficult area."

"So recollection is the only paradise from which we cannot be turned out. So again, interesting thought about memory. Let's get down to the cellular mechanisms. The gentleman asked earlier about that. Let's get down to the neuron. I think you're all educated about neurons at this point. We've got the cell body, we've got the dendrites. Messages are coming into the dendrites, typically going out the axon. The axon has these myelinated insulation. The action potential, uh, the electrical action potential travels along the axon, gets to a synapse. So the presynaptic terminals here, there's the synapse, the space in between, the postsynaptic terminal, this generally a spine, and the spine is sitting on a dendrite, I, one of those things. So this neuron is transmitting to the next neuron, and that neuron will transmit to the neuron after that, and on these dendrites. So here's another image of dendrites. If we blow this up, we see these spines, and these spines are delicate, critical processes because the synapse, the connection between the neurons, the comm, electrical communication between the neurons, is fully dependent on that spine. It's not good to lose spines, but they're very delicate. We lose them with Alzheimer's, age, um, etc."

"Now, this is sort of a complicated diagram, but this is the hippocampus. We're looking on the cellular basis. We've got memories that have come in through the senses into the cortex. They come down to the entorhinal cortex. That's the cortex next to the hippocampus. They come in through the perforant pathway to the hippocampus, and these fibers."

Are synapsing on granal cells. Here's a granal cell right here. Here are its dendrites. Fibers are synapsing on it, and the granular cells are sending fibers along through the mossy fiber pathway to paramal cells. These triangle neurons here, synapsing on those. And then those paramal cells are sending their axons to other paramal cells. And finally, these paramal cells are sending their axons to the output track, output back to the hi um cortex. And so we see into the down from the cortex, this loop through the hippocampus, through this chain of neurons, and then back out up to the cortex. So this is this hippocampal cortical coordination.

When we create memories, we're modifying these connections. We're making a given connection stronger or weaker. That's the bottom line change that occurs to form a new memory. If some of you in here are computer scientists or engineers, I'll I'll welcome your comments on this. I'm not.

As we move on a little further, so if we magnify this here, we've got this again. We've got this neurite, this axon coming in here, synapsing on this granal cell. Um, and the the discovery. There are three ways to modify the synapses. The first is by this concept of long-term potentiation, or LTP, that was discovered in 1973. And that was the idea that if we stimulate with an electrode this incoming axon, making it more active, and then record at the receiving neural on when we stimulate, we get a we look it on oscilloscope and we stimulate. This black line is the normal peak we would get the first time we did it. But if we stimulate it a lot, 100 pulses a second, we get a much higher peak, a much bigger response from that receiving neuron. And if we go away for a few hours or days or even weeks, and we come back and just stimulate at once, we'll still get this extra size, extra large peak. So, in other words, when this connection was stimulated and was used a lot, it became more efficient. The secondary neuron responded with a bigger response.

There's an opp opposite phenomena called long-term depression, where the peak actually gets smaller. This is the key electrical change that results, uh, from more firing. If this axon is more active, we're actually building the strength of that synapse, and that's called long-term, uh, potentiation.

What do what's the chemical or protein basis of long-term potentiation? Again, we've got this pre-synaptic axon coming coming in. This is the spine. This is the receiving end. So the pre-synaptic axon, the electrical impulse, the action potential travels down here, hits the terminal, releases the neurotransmitters that are in these vesicles. They're released into the synaptic cleft. These neurotransmitters, glutamates, one shown in purple, bind to these receptors. One's called the NMDA receptor. When it binds to that receptor, that receptor opens a channel. Calcium comes blasting through. When calcium comes in, it activates a whole chain of signaling proteins that activate this, uh, factor called CREB, that causes gene activation. And the gene activation causes this neuron to make more of these receptors so that they can, uh, handle even more of these transmitters and fire in an even more powerful way. So that's the molecular version, very briefly, of how this firing from the pre-synaptic terminal enhances the ability of this post-synaptic, uh, axon neuron, uh, to fire.

We can get more advanced and we can look at the associa of LTP. And this deals with a question that was asked earlier. We can have one neuron here, it's dendrite, and we have three different neurons here synapsing on different parts of this dendrite. We can have a weak input, and the weak input from this neuron might not be enough to activate this neuron, so the message would stop right here. We could have a strong synapse where if this draw fires, that's enough to activate this. And we can have another weak synapse here. So this wrong synapse might be coming from the amygdala, the emotional part of the brain. This neuron might be coming from part of the brain that's responsible for remembering somebody's name. A weak synapse. Now, to remember this person's name, ultimately, we want this neuron to fire. Now, I'm showing you an example of a handful of neurons here. There might be a hundred thousand neurons and synapses involved in remembering a name. If we try to remember this name with just this connection, this synapse is not strong enough. This neuron won't fire. We'll fail to remember the name. We fail to move it to long-term memory.

Changing these connections is the process of forming long-term memories. However, if we're trying to learn the name of somebody that gives us an emotional element to us, this person triggers some kind of emotion in us, the amygdala neuron will fire. If that's a strong synapse that's capable of activating this neuron, and there's a principle that if this neuron's already activated when this weak synapse fires, this weak synapse will convert to a strong synapse. That's part of associative long-term potential. This will remain stable as a long as a strong synapse that's part of this, the stable retention of a memory. So next time we want to remember that person's name, we don't need the emotional input. We don't need the amygdala neuron firing. This synapse here is strong enough to trigger this neuron. And so we've succeeded in forming that long-term memory of remembering that person's name. So I know that might be a little confusing, but that's what's going on at the synapse, uh, level, and that's long-term potentiation.

So long-term potentiation reshapes the population of which cells are being activated by given input, and it's the changing of that network that's the that's the new memory, that's the long-term memory. We have long-term memory means permanently changing the pattern of these networks by changing the synaptic strength. So we can illustrate that here. We're getting visual, auditory, in tctl input from our from the cerebral cortex. It's coming down to the medial temporal cortex and hippocampus. The black lines are strong, our connections with strong synapses. The the blue lines are weak. We can put in some new event from the world that causes a change in the pattern of strong synapses. It's transient, but if we can rehearse it, convert it to a stable change in synapses, even though even after the sensory inputs gone, uh, it stabilized into a stabilized change in the network, resulting in the successful conversion of that membrane from short, uh, to long term.

Now, if we take go go back and look at these proteins, some of these proteins that are responsible for this, what if we design a mouse that has extra receptors so that this synapse will fire even more readily, uh, for a given input? And this was done in the 1990s, and the mice ended up being smarter than regular mice. Uh, this was a very prominent, uh, paper reported in Nature, and they're called Doogie mice, named after Doogie Howser, MD, in the in the 1990s. There that show about some young kids, a physician, and they call these Doogie Howser mice because these mice outperformed any other mice that had ever been tested, and they were just expressing extra amounts of this NMDA receptor. Now, what does that mean, um, in human, uh, terms? And since then, over 30 proteins have been found that if you express extra amounts of that protein, all regulate all related to synapse function, the the mice are smarter. Um, and so we get to the whole human, uh, element of of hyperthymesia, the ability to remember extreme amounts of information better than other humans can.

And the first case was described by the Russian psychologist A.R. Luria. He has this book published in 1968 where he describes, uh, aist, a man named S, that could re remember, uh, amazing amounts, but it turned out he was really using a mnemonic trick more so than probably exceptional memory. Um, and if you gave him a list, an incredibly long list of items, somebody asked about visual technique. He had memorized every nook and cranny of Gorky Road so well that if you gave him a list of a hundred things, in his mind, he would place each object at a certain part along Gorky Road. And then the next day, if you asked him to recite the 100 objects, in his mind, he would take a walk, walk down Gorky Road, say, yes, fruit, banana, etc. So, very powerful, uh, visual technique. Um, but more recently, we have this, uh, case published in N 2006, which I'll show a video of, the woman who can't forget. Maybe she, they call her AJ. The subject of most medical studies are usually anonymous. She's a woman with the most amazing memory known to science. By understanding how her her brain does this, we may help write a new chapter of memory research. It's got to be just like any reference. I mean, you go to the internet, you look in an encyclopedia, or you ask my sister. AJ remembers details of what she did every single day since she was 14, instantly recalling dates of news events in her lifetime. She has the ability in a split second to tell you dates, times, what she was doing, what we were doing, what was happening in history. It's hard to describe for her. Being able to forget isn't easy. She became a fascination around the world, the human calendar, doctors, journalists wanting to meet her, her voice heard on the radio just once. If I'm able to cure a disease, it's a gift. But to remember like the end of end of every relationship, or, you know, anything, it's, it's, it's hard. But it has formed who I am because I remember everything. Until now, she's been anonymous, dealing with the interest in her brain, her life, almost too much for her to bear. She chose not to reveal herself to the world, uh, for several years. You can imagine it yourself, uh, if if suddenly the whole world knows about you, there's a lot of people out there. And we are joined now by AJ, who is really Jill Price. We're saying it on television, a 42-year-old school admin administrator from Los Angeles. And as I said, it is her first ever live appearance on television. And we had a chance to sit down earlier and do a taped interview. Good morning, you again. Can I throw some more questions? Sure. I love that you don't mind that we keep testing this way. All right. I know television is one of the things that that you remember the what you saw on television on the days in your life. So when when was JR shot on Dallas? When he was shot, or when did you find out who was he was shot? Who shot him? Uh, we were shot on March 21st, 1980, and we found out November 21st, 1980. What about, uh, the end of All in the Family? Well, that was in 19, March of March of 1978. What about Nancy Carrigan attacked? Something that interested you. January 6th, 1994. Oh, Lord. You're right. The LA riots, they started on April 29th, 1992. Unibomber arrested. Ted Kaczynski arrested, um, April 3rd or 4th, 1996. April 4th, yeah, 1996. Does this mean you were good in school? I was not good in school. See, this is inconceivable. You, you didn't have the ability to take a poem and then suddenly recall it. My memory is autobiographical. So I could tell you my life, but to memorize a poem or a monologue was very excruciating for me.

Which raises this old question that we have talked about a lot, which is, we all think it would be bliss to be able to remember every day of our lives. Is it more of a comfort or a burden? It's both. Tell me about the burden. Tell me about what it is to be able to remember every single painful incident in your life. It's horrendous. Small things. Uh-huh. Everything. Yeah. Regrets, choices, and I could literally go back to the exact moment I I was in that fork in the road where I could have made this choice, but I made this choice, and then that leads to this, which leads to that, and I just can't, I can't forget that.

Okay, so too much memory. She didn't, she wasn't a great student in school. And this really points out to different kinds of memory. As she said, autobiographical, so a part of her declarative episodic memory, she's really good at. But these are things that related to her. If you ask her, it doesn't fall flow necessarily into semantic or or world knowledge. So again, just a subset of memory is exceptionally good in her. She's the first case in the literature that's been described in medical literature formally like this. Two or three others claim to have this, and they're they're being investigated, but this is a relatively new phenomenon. My own speculation is that just just like all of our genes, all of our proteins, you know, we all have slight variations, you know, of the 30 or 40 proteins relevant to that synapse, might there be rare variations that people have, uh, that some synapses are incredibly strong in in one out of, you know, millions of people? We'll see as time goes on.

I'm going to try to wind down here because I know I'm two minutes over, um, but I want to do show you a second way a synapse can change because to me, it's incredibly inspirational. Um, many people think you, you had a lecture on the heart. The heart has elegant movement. The brain's kind of boring. It doesn't move. It just sits there. Um, in fact, here are some heart cells grown in tissue culture right here. They beat. That's really cool. And for all of us who have grown cells in tissue culture, including myself, we've all, we've put the heart cells in there, and you can see them beat. It's really fascinating. You wouldn't expect a brain cell, uh, to have any movement like this. Um, however, these are dendrites, and those little things coming off them are spines. And if you ask me, as a neurologist, to to share with you the most powerful vision I have of the brain in my 30 years of neurology, I'm going to share it with you right now, and this is it. This is brain tissue. It moves. Why are those spines moving? Because they're looking to establish new connections. With this imaging technique, we're just looking at the receiving end, the spines. Uh, the axons coming in that want to meet with them are not being visualized here, but they're there. Okay? So if a spine is randomly searching around, if it happens to find an incoming axon who's active, that connection can stabilize. If that incoming axon is not active, it's not attractive. We won't have a stable relationship formed. The brain is incredibly dynamic. It's a very much use it or lose it, um, organ. We can look at this as a closeup look at this thing. Look at this. This is looped here. Uh, we see these connections forming, changing, reaching out, looking for a connection. And if there's an active put coming here, this connection will be stabilized. So we have 100 billion neurons, probably over 100 trillion spines doing this right now. If we can stabilize a long-term memory, this is what's going on in her mind. So my friends and cardiology say the friend brain is boring because it doesn't move. This is, this is amazing. This is amazing. Here. Let's say. Okay. Thank you. [Applause] Wow. Use it or lose it, you say. All right. Well, I think this has been a fascinating, uh, display, obviously, and a wonderful presentation. Frank Longo, thank you so much for that incredible, uh, all right. I'm sure you're going to get lots of questions, uh, and you're all going to descend on him, uh, right now. And please do. For the rest of you who don't remember anything you've heard tonight, go home and come back next week. All right. Thank you again. For more, please visit us at stanford.edu.