The making, keeping and losing of memory | Ferrier Prize Lecture 2023
Transcript
- Ladies and gentlemen, good evening. My name is Annette Dolphin, and I'm very privileged
- to chair this event as a member of the Royal Society Council. A bit about the Ferrier Medal
- Lectureship. It was created in memory of the neurologist and psychologist David Ferrier FRS,
- and was first awarded in 1928, and it's to be given on a subject related to the advancement
- of natural knowledge on the structure and function of the nervous system, which of course is exactly
- what our lecturer tonight works on. So tonight's Ferrier lecturer is Professor Richard Morris of
- the University of Edinburgh. Richard investigates the neurobiology of learning and memory.
- That is how the brain mediates the acquisition and storage of knowledge and its later recall. Richard
- developed one of the most widely used tests, as we all know, for studying the causal mechanisms
- of memory in animals, the Morris Water Maze. Wouldn't we all love to have a piece of equipment
- named after us? He's used this to dissect the role of the brain structure, called the hippocampus,
- in spatial memory and navigation, and I'm sure we'll hear all about that from him.
- He wanted me to keep this short because he has a lot of a lot of lecture to get through but
- he did want me to say that, and of course, he's fulfilled many scientific and administrative
- roles and is currently chair of the Brain Prize Selection Committee. So now we have
- more time for the science which you've all come to hear about, and without further ado, it's
- my great pleasure to introduce our Ferrier Medal lecturer this evening, Professor Richard Morris,
- to give his lecture entitled The Making, Keeping, and Losing of Memory. Richard.
- Thank you, Annette, for your kind introduction
- and thank you all for coming. It's nice to see so many friends, scientific friends, other friends,
- family, and so on. I really appreciate you coming to share this evening with me. As my title
- implies, what I want to do is to talk about three themes in relation to the neuroscience of memory,
- making, keeping, and losing. By making, I mean what sort of brain regions and
- what activity patterns mediate the encoding of information into memory.
- By keeping, what neural mechanisms underlie the storage of information over long periods of time,
- and is there some selectivity to memory retention, which is an idea I've been advancing for some
- years now. Then come to losing, the benign forgetting that we all have from time to time,
- where on earth did I put my glasses, that kind of thing, through to more serious forms of
- forgetting. The latter being perhaps the more important things to come. Now, there are lots
- of people to thank. Every year we take a picture of the lab group, and here's one from a few years
- ago, and they represent the people who over the years have made lots of contributions to
- the research from my group. Now, please understand that if I fall into saying, I observed something,
- what I mean by that is we observed, but if I do say we observed, what I mean is they observed.
- Now, David Ferrier, as you've heard, was a renowned 19th century Scottish
- neurologist and experimentalist best known for introducing and refining the technique
- of electrically stimulating the brain. He followed up observations, particularly on the motor cortex,
- beginning some of the causal studies of brain function. If we wind forward 60 years from there,
- we come to the celebrated 20th century neurosurgeon in Canada, Wilder Penfield,
- who also used the technique of electrical stimulation and whose work helped refine
- the idea of a homunculus across the surface of the somatosensory and motor cortex, in which the
- area of the neurones is not in proportion to the size of the body part, but in proportion to the
- importance of the body part. So the hand area is huge, the back area is very small. Then if
- we wind forward another 60 years, we come to 21st century neuroscience, and I could mention a number
- of people, but the one who I think has made some of the biggest contributions is Karl Deisseroth,
- who's one of the architects of contemporary optogenetic methods of stimulating the brain
- using light, a technique that I shall return to, but let's turn to memory. For all of us, memory is
- a kind of conscious act of remembrance, something that happened in the past, happy events, birth of
- a child, Christmas, you name it, but also of sad events. Here is Angela Merkel on the occasion of
- the 30th anniversary of the fall of the Berlin Wall, lighting a candle. A moment that was,
- I think, particularly poignant because, of course, Angela Merkel was brought up in East Germany.
- Now, remembrance over time changes inside our brains and our feelings about things change,
- and it can get transformed from joy or sadness into expressions of national culture.
- An example of this is, of course, the ceramic poppies at the Tower of London some years ago.
- So that's remembrance in all its various different forms. But what is memory itself? Well,
- the textbook definition is it's a change in our response to things due to experience,
- and it reflects our capacity to change our understanding of the world and then our behaviour
- as things happen. Now, on this view, learning and memory are intertwined.
- Memory is more than just the folk concept of remembrance. It's a set of neural processes
- through which we encode information, store it, sometimes consolidate it,
- and then later express things through acts of skill, of recognition or of recall,
- and this is what the brain does. So let's start with the brain. Let's start with my brain.
- I think it's valuable for you to know that the person in front of you does
- have something inside his head here. I got this as a present when I was a subject in
- a scanning experiment at Massachusetts General Hospital, when I was working at MIT, and I was
- just relieved to see this image, but then came the moment which happens in many universities,
- and certainly in Edinburgh, in the medical school in Edinburgh, where new professors have to give
- an introductory lecture, an inaugural lecture. My parents came to this lecture and they were sitting
- in the lecture theatre, and since I had over many years given them many opportunities to
- doubt that there was a brain inside this head, I thought that I would show this picture,
- and I presented it very confidently as evidence that there was a brain inside my head, whereupon
- my father shouted out to the whole audience, 'No proof that it's working, my boy.' Not
- the last time my father upstaged me. So the brain has about 100 billion brain cells, they are 25
- microns in size, 5000 synaptic connections, which are really, really tiny, 20 to 40 nanometres,
- and of course, then 500+ trillion connections, and that's all in 1.4 grams. What we have across
- the brain is various different regions and neural networks, made up of cells whose neural activity
- represents experience in distinct ways and whose structure and connectivity changes with
- experience. Now, with modern imaging techniques, we can now see non-invasively the connectivity
- of the human brain and learn a lot about how this connectivity changes with ageing and in disease.
- So this brings us to the question of what might be distinct about the neural mechanisms of perception
- from those of memory, and there are lots of different facets to that but one is
- that perception and action is online. It has to happen as the events are being perceived
- and it acts in real time with those events as they occur. So if I show you this image,
- you will see lots of black and white, and then gradually over a few seconds,
- the Dalmatian dog will pop out in front of you. So that sort of perception happening in real time
- in your brain. We now know a lot about how this actually happens, beginning with the Nobel Prize
- winning work of David Hubel and Torsten Wiesel, who looked at the centre surround fields of cells
- in the lateral geniculate nucleus, which projected up to the visual cortex, and particular patterns
- of anatomy from this early stage through to the neocortical stage, gave them the opportunity to
- see line detectors. That was the beginning of our understanding of shape perception.
- So then the reverse question, what might be distinct about the neural mechanisms of memory
- from those of perception? I put it to you that one key aspect is that the neural activity can
- mediate a kind of offline sense of mental time or space travel, back to a moment in time, be it last
- Christmas or what have you, that we can do while not believing that we've gone there but imagining
- that event from the past. Now, this is, I think, particularly striking in the case of music.
- I'd like to share with you one of the most memorable experiences of my life.
- I was attending an international conference in Sicily, in Taormina, late June one year,
- and that particular evening was my birthday. I said to a few people at the conference,
- let's go and watch the opera in the amphitheatre in Taormina, and of course, it being Sicily,
- the opera was Cavalleria Rusticana, and so there we were in this amazing amphitheatre,
- Mount Etna in front of us, the sun going down, seeing the twinkling of the boats,
- of the fishing boats in front of us, and for the first time I heard the Easter
- Hymn.
- [Music plays]
- So how is this done? Well, the truth is, we don't really know, but what is clear is that we have to
- not just think about neural activity, but all these pathways that I've been referring to,
- and many of these pathways are re-entrant, and that re-entrance may be critical for being able
- to go back to the past. Now, the founder of modern anatomical studies was Santiago
- Ramon y Cajal, who, working in Spain using the Golgi technique, was the first person to start
- looking at neural connectivity in a variety of different animal species, and he gave the Croonian
- Lecture to the Royal Society in 1894 about his ideas, which also included ideas about memory.
- Now, if you wind forward then about 100 years, you come to famous diagrams like this Feldman and
- Van Essen diagram, in which you can see from the bottom, the retinal ganglion cells and the LGN,
- and up to the visual areas, and then coursing through all manner of regions in the associative
- cortex, which somehow turn these perceptions into representations of knowledge, and so on,
- and finally at the top is a structure called the hippocampus. So let's turn now to making memories.
- So the first big distinction, which you'll all be well aware of, is between short term memory
- and long term memory, and although there's arguments about the exact neural basis of
- these different forms of memory, broadly speaking, short term memory reflects the continuous neural
- firing of a network of brain cells which keeps a memory active, and if you interrupt that firing,
- you lose the information. So if you write down a telephone number and then try to remember it,
- if somebody interrupts you, you'll lose it. So short term memory has great fidelity but
- is subject to interference. In contrast, long term memory involves changes in the strength
- of connections between brain cells, which store information through structural and biochemical
- changes that will remain passive for long periods of time but then can be reactivated later. So
- remembering back to Christmas, it's not you've had to remember every single instance since last
- Christmas. You can reactivate it with a cue. So let's now do a little experiment with everybody
- here. A short term memory experiment. What I'm going to do is I'm going to present some letters
- and numbers and symbols to you and I want you to imagine that you're in one of these short term
- memory experiments and trying to remember this information. So I'll present it to you briefly.
- And then take it away. What these nasty experimental psychologists then do to you
- is they make you count back in threes and things like that. So while you're trying to hold this
- information, you're told 54 and you have to then say 51, 48, 45, 42, and then it gets more
- difficult again and so on, and then at the end of that, after some period of time which might
- be 20 seconds or 30 seconds or a minute, you're asked to write down what that information is,
- and if you go beyond seven plus or minus two, it gets really quite difficult. So here are some
- of the characters. You probably saw two equal signs. Maybe you saw the capital A, capital M,
- don't know about F. The lowercase might have been a bit harder. You might have got the last letter
- which was M. So here it is again and some of you might have got it correct. Some of you might not.
- What I've actually presented to you is Newton's equation, force equals mass times acceleration,
- and Einstein's equation, energy equals mass times the square of the speed of light. Now,
- if I'd presented the information that way to you, you'd have found it much, much easier to remember
- because in your perception, you'd have accessed your semantic memory, your knowledge of science,
- and have got that information right away. So we come to a bit of a surprise, which is
- that long term memory is, in parts, an input to short term memory and not the other way around.
- Now, when we turn to long term memory, there are lots of different types of long term memory.
- One of them is procedural learning, actions, habits, skills, and that was the subject
- of the last Ferrier lecture by Daniel Wolpert. So I won't say more about that,
- and it's mediated by a structure called the striatum in the brain. A second is emotional
- memory which is mediated by the amygdala, and we have in the audience tonight, one of the world's
- experts on emotional memory. Another good reason for me to say as little as possible about it.
- So let me turn to declarative memory. Declarative memory is like Gilbert-Ryle's knowing how,
- which is procedural, knowing that. This is the knowing that bit
- and you can know things about events. You can know things about facts, semantic. You can know
- things about the spatial layout. All three are important components of declarative memory and
- they can't be fused into one another and they're mediated by the hippocampus and the cortex.
- So let's think about this. For skills, you repeat a skill many, many times and through reward
- processes, you gradually get better and better. Whether it's your tennis swing or golf swing
- or skiing or whatever it happens to be, and it requires massive repetition to get to that skill.
- In the case of episodic memory, it's completely the opposite. So here's Seve Ballesteros,
- who's just putted into the 18th green of the Old Course in St Andrews and won the 1984 Open,
- a moment which I'm sure he remembered for the rest of his life. The point is, there's no trial two.
- You don't have another putt. You've just got that one event. I was there at the time. I was
- a lecturer at St Andrews at the time, standing, and I'll never forget that push of his wrist. It
- was great. I didn't take the photograph. It was actually taken by a man called Peter Adamson,
- a professional photographer in the town who we all knew and who was
- the great grandson of Robert Adamson, famous for the Hill Adamson Calotypes.
- So episodic memory is about specific events, and these events have to be captured on the fly
- and then somehow remembered. I think the French photographer, Henri Cartier-Bresson,
- captured the notion that there's a decisive moment when something happens, and in this
- lovely picture, this famous picture of a un garcon carrying his van ordinaire back to his granny,
- he's capturing a particular moment of pride of this little boy, but also an aspect of
- Parisian culture from the 1950s. Now, semantic memory is more about our knowledge of facts,
- where we strip away the timing information and somehow create some kind of organisation,
- as you see here in this particular slide for the information. I won't take you through it, because
- I think it's fairly obvious that if we create effective mental models, we have a good knowledge
- structure that we can use in our professional life, and then there's spatial memory. Here's
- the city I live in, Edinburgh, which I know reasonably well, and this is also information
- that as you build it up over time, becomes something that the hippocampus helps us do.
- Now, where is the hippocampus? Well it's sitting in the medial temporal lobe just inside here and
- it's critical for the formation of these three types of memory. Now, if I was to come in and cut
- through the hippocampus, we would in fresh tissue, not see very much, but nowadays we have amazing
- new fluorescence indicators. This particular picture is from my colleague Francesco Gobbo,
- of the lines of cells that you have in the hippocampus, which are cleverly into a careful
- line, and a very interesting architecture, which I shall come on to in a moment. Now, just as I
- was finishing my PhD, I met two people here in London, John O'Keefe and Lynn Nadel. John had just
- done the amazing experiment of lowering a small electrode down into the brain of a rat to these
- very cells, and started recording the activity that he saw there as the animals ran around,
- and he made an amazing discovery, which is that some cells will only fire if the animal is in a
- particular position, other cells where the animal is in a different position and so on, and he
- mapped these across the surface of the hippocampus and identified what we come to call place cells.
- Later, further discoveries were made about these spatially selective cells of the hippocampus and
- adjacent regions. The famous grid cells of my former postdocs, Edvard and May-Britt Moser
- and with Edvard and May-Britt, John shared the 2014 Nobel Prize for their amazing discoveries.
- So what you've got is wires that are being put down,
- and you try to represent the height of activity for a particular point in space,
- which you have here in this circular arrangement, and that's what our place cells are. So
- having just met them, I then got a job at the University of St Andrews and went up
- and started doing my lecturing as a young lecturer, and then I had a stroke of luck
- because I discovered that I'd accepted a job in a department which had absolutely no lab facilities.
- So at the end of the year, having done my lecturing, I went to the Professor Malcolm Jeeves,
- and I said, this is great, I am enjoying being here, I've done my best with the lectures,
- but I'd like to do some research, and he said, leave it with me, I'll find you a
- place. So about a week later, he called me in and they said they had found me a place
- at the back of the Gatty Marine Laboratory, which you can see is a rather forbidding place. The
- reason it was lucky was because to go into my lab, I had to walk past lots of tanks of sea creatures,
- time and time again, and so I thought, I wonder if rats could swim. Maybe I could
- use this to solve the problem that John had given me, which is to try to devise a behavioural task
- in which animals would move to a position which is not defined by visual cues,
- not defined by auditory cues, not by olfactory cues, but a point in space.
- So what I did was I developed what you've heard is called the Watermaze. I would
- like to add I've never called it the Morris Watermaze in any of my papers, but anyway,
- I know that there are people who have. So it's just a pool of water with a platform
- hidden underneath which the rats can climb onto. So let's just show you a little movie. This is a
- well-trained rat because he's on television, so he should do what he's supposed to do.
- So he swims to the platform, which is great. So that's what he's doing. Notice he heads off
- in the right direction from the starting point. Then my assistant pushes the platform down. It's
- on a spindle held by an electromagnet. Now we do something that the rat's never had. This is the
- first time he's been filmed. It's what we call a probe test, where we're trying to investigate
- how much spatial knowledge this animal has, and you can see that we're tracking him in
- real time and I hope you can be quite quickly persuaded that he has some knowledge of where
- this safe position in the water is, despite there being no local cues that define it.
- Now, the good thing about this platform is that after a while, we can release the electromagnets,
- we can bring it up and rescue the animal as he climbs onto the pool.
- Now I set this up with a technician, buying lots of stuff from a local yacht chandler,
- resin and polystyrene and stuff, and I didn't at that time have a research grant. We wanted
- to make the water cloudy with milk because we didn't want the animal to see the platform,
- and we were going through quite a lot of milk, and then we decided we'd switch to powdered milk.
- So that seemed to work. At that point, being the sort of naïve young lecturer I was,
- I thought, right, I'm going to write to a company.
- There was a company at that time which made most of the powdered milk. I think it was called
- St Ivel. I wrote to St Ivel and said, I'm doing this incredibly important research in St Andrews,
- and we're using a lot of powdered milk, but if you could supply me with a huge amount
- because I don't have very much grant money, I'd really appreciate it. I said, look, when I publish
- my results, I'll be sure to thank the company. So a few days - well, actually a couple of weeks
- later, a huge pile of powdered milk arrives with a letter from the managing director. He said,
- we're very excited about your research and you should have received a package of powdered milk,
- good luck, but please don't thank the company in your paper. Now,
- place cells tell an animal where he is. So one set of cells will fire for here, different cells here,
- different cells for here, and so on. What they don't do is tell you where he's going.
- So if you can imagine a set of cells that correspond to place A in this diagram,
- there'll be different cells that correspond to place B. Place cells are defined as cells that
- fire, if and only if, the animal's in that position. So there's a logical conundrum
- here. How does the animal at place A access the information that he wants to go to place
- B and not to place C without first going there? Because by their definition, that can't happen.
- So we're not sure really the solution to this problem yet. People have worked on it quite a bit,
- but my stance on it depends what you train the animal to do, and I'll try to give you an idea
- of how we've recently tackled this. So this is a new piece of apparatus called the Event Arena,
- which is basically like the Watermaze, except the animal has sand wells that it has to dig in
- to get food, and what's better about this is we can now do physiology, which we can't really do
- in the Watermaze. So the animal finds a sand well, digs away, finds a nice big pellet, and like any
- self-respecting rat, he takes it home to eat. So a new technology has emerged called calcium imaging,
- and what you do here is you put a virus into the area of brain that you're interested in,
- such as the dorsal hippocampus. This virus has a fancy name, GCaMP6f,
- if the cell fires and releases calcium, you can detect that calcium with a camera. So what we
- have is these miniature cameras that weigh less than two grams and they have a CMOS unit in it,
- just like you have in your mobile phone. So you can get very good quality images. Then
- a gradient refractive index lens being pushed down into the hippocampus to detect these cells.
- So you see that some of the cells here are firing, the lines in the different colours,
- and there are lots of regions of interest where there are possible cells. So here in
- panel A, it shows the virus, B is our camera, C is the lens coming down towards the hippocampus,
- and the raw data is a bit of a muck, but if you take the delta F over F, you can see the
- individual cells and it's wonderful to watch this on the screen. You see them firing away.
- Francesco Gobbo did this experiment, training animals in the Event Arena, getting as many as 190
- cells in hippocampus in each animal, and what you can see here is that they… I've lost my pointer,
- but it doesn't matter, is that the animals gradually get better and better at the task over
- a period of about 20 sessions or so. So that by about session 18 to 21, we're able to record from
- these animals, and in terms of the paths shown in panel M, they're a bit all over the place at
- session two. They're getting better by session nine and by session 19, they're doing great.
- Now comes the critical part of the experiment. So we now know the animals can perform this task.
- They can go to different places each day. What we're now going to do is record in the start box.
- We're not really interested in what happens in the Arena. We're interested in what's happening in the
- start box. What you see in panel A is the activity gradually creeps up. Something's happening.
- If you then train a decoder to see what cells are firing, you'll see that in the start box, the
- cells that are firing are the ones that correspond to the destination the animal should go to.
- So it's as if the animal's imagining the future in the start box, I've got to go this way,
- or I've got to go that way, and we can see this in the calcium imaging right in front of our eyes.
- So if we turn to panel D, this is brought alive a little bit further in that you can see that we're
- only recording in the start box, but we're locating where the place fields are of those
- cells if we just let the animal run around, and in this particular case, at -8.4 seconds,
- the animal maps out in his mind a path to the nearest sand well. Same at -7.7, the same at
- -7. Then he has second thoughts. Maybe that's not the right one. Maybe I should go to this other one
- over here, and you can see here he is still in the start box, you see the cells that correspond
- to this other region start to fire, and then he changes his mind back again and does things right.
- So we're able to predict with high accuracy where the animal is going to go on the basis
- of the brain activity in the start box, and we've also used what's called population coding
- using manifolds through a collaboration with Simon Schultz at Imperial College,
- Rufus Mitchell Higgs is the expert here, and he is able to measure the centroid of the two
- dimensional cell pattern and what he finds is beautiful circular paths, so that if you have
- the thing very dark blue at the beginning, when the animal's in the start box and dark red at the
- end when he gets to the destination, what you've got is circles. So the centroid starts at one
- point and comes back to that same point time and time again, and you can see several circles here.
- So what we think is happening is if you've got an animal who's been trained to do a
- particular task where he has to anticipate the future of a point in space, we can map that
- out in this early period, which I think is quite exciting. So let's now turn to keeping memories.
- Now, we all forget, and of course, forgetting has been the subject of study for over 100 years.
- It was a classic study by the German experimental psychologist Hermann Ebbinghaus,
- who used himself as a subject, and he found that in the particular task he'd set himself, which was
- sort of nonsense syllables, quite a hard task. He only had 33 per cent retention at 24 hours.
- So forgetting happens naturally during the day. The argument I want to put to you in
- this next section is, thank God it does, and I'll tell you why.
- Even though we have forgetting, are memories lost randomly, or is there some selectivity?
- Well, our intuitions here are probably broadly correct. We remember what we attend to, of
- course, it must be downstream of some attentional filter. We remember what we're interested in.
- We remember better when we're surprised by something and so on.
- So from the stream of events during the day, some things are remembered well, but not everything.
- Now, an important human brain imaging study was done in 1998, led by John Gabrieli at the
- Massachusetts Institute of Technology, and he invented something called event related fMRI.
- So he was going to measure the bold signal as people learned information in a scanner,
- and information was presented roughly every ten seconds or so,
- and he measured that bold signal, and then after he'd done that for about, I think about 45 minutes
- or so, he brought the people back the next day and tested them on how well they remembered the items.
- He found some were remembered, some were not. So he then went back and resorted the bold signals
- as a function of whether the information was remembered or not remembered, and what you can see
- from the red signals is the bold signal's bigger for the information which is later remembered.
- So it suggests some kind of neural activity is happening at the time of encoding,
- which is predictive of what we might be able to remember better the next day.
- So encoding matters, but I want to argue that that's not the only thing and may not even be the
- major thing, because then comes another great big discovery about the way the hippocampus operates,
- which is critical for the next part of the story, and that was the discovery
- by Tario Lomo, later aided by Tim Bliss, in a big experiment they published together in the
- Journal of Physiology in 1973, the discovery of long term potentiation. They had a control
- pathway which has the open symbols, which you can see remains pretty static over six hours,
- and then where the arrows are, they gave brief bursts of high frequency activity, which might
- be a bit like the neural activity that would have happened in these event related bold experiments.
- What they found was that on the other pathway where that was done, the signals got bigger.
- The synapses, as it were, were getting stronger and they called this long term potentiation.
- It's been intensively studied ever since. Now, I don't know whether you noticed,
- but at the position at about quarter to five hours on the diagram, there's no data
- for about 45 minutes. Tim tells me that this is when both he and Terry fell asleep.
- So activity dependent, neural activity dependent synaptic plasticity occurs in the hippocampus
- and we have this physiological phenomenon of long term potentiation,
- which is a process through which the strength of the connections can be enhanced but enhanced in
- an activity dependent way. It uses an automatic Hebbian rule for synaptic change, not an error
- driven rule like in the striatum. It's so-called unsupervised learning. Soon after this discovery
- came another big discovery, this time done by Graham Collingridge together with colleagues
- when he was working in Canada, which is there's a particular type of glutamate receptor, now called
- the N-methyl-d-aspartate receptor, which if you block it, blocks LTP. It doesn't block neural
- transmission, but it blocks the ability of those synapses to change in strength. It's a fascinating
- thing because we now understand the molecular basis of the NMDA receptor in much greater
- detail and realise that it's a kind of conjunction detector, and while I can't take you through the
- full details of this, the idea is that ordinarily, if you release transmitter onto the NMDA receptor,
- nothing happens, but if the postsynaptic side, the receiving side is depolarised,
- then what happens is that a particular ion, magnesium, is ejected from this channel, and then
- other charged ions can flow through. So it's a conjunction detector. It detects when there's
- activity on the presynaptic side and activity on the postsynaptic side. It then signals this
- through downstream signal transduction mechanisms and as you can see from the diagram on the right,
- the structural biologists are getting into the act and understanding exactly how this operates.
- So these NMDA receptors exist in large numbers, probably about 20 or 30,
- on each of these tiny little synaptic connections, and here in this further image from Francesco,
- you can see these little spine-like protuberances coming out of the dendrites and each of those will
- have these synaptic connections. We're just seeing the postsynaptic side here.
- So the basic idea is that there's a basal state where you've got so many receptors, and then you
- can potentiate it, you can have more receptors for reacting to glutamate, or you can depress
- it and have less of them. So here, we've got a mechanism which is certainly a memory mechanism,
- but on its own just as a synaptic mechanism, it won't do very well. We now have to,
- just like Howard taught those years ago, embed this into some kind of anatomical circuit.
- So let's do that because this next bit is kind of magic actually.
- Let's suppose you had a simple reflex network where you've got an input coming to an output, and
- you then induce LTP at a particular point. What will happen is you'll get a bigger output for a
- given input. Many people have supposed that that's the basis of learning but that's a very limited
- form of learning. If you instead have a slightly more complex circuit in which you've got one cue
- coming in from the left, with activity at the red synapses, but not the blue ones,
- and then another cue coming in from the right Q2, what I'll show you is that
- you can build an association between those such that once you've got that association,
- if you get Q1 come along, then Q2 will be the output from the neurone. You build associations.
- I can't take you through this in great detail but imagine four wires running horizontally
- and vertically, and we'll present information to it as a binary number, which is sensible because
- neurones fire with action potentials. So 0011 comes in on one side and on the other side, 1010.
- Well, if the rule is that you've got to release transmitter on one side and be depolarised on
- the other, the Hebbian rule, then you'll only get synaptic change where a one meets a one,
- you won't have it at zero meeting zero, zero meeting one, one meeting zero. Only at one
- meeting one. So you end up in the first diagram with just these four points. So the memory is
- distributed. Then you come along with 1010 which is this one, with another signal, 0110,
- and you add in some more synaptic connections. Now this is just a Mickey Mouse version. It'll
- get saturated very quickly, but importantly, if you now come in with 0011 here to the final
- state of the matrix, you'll get 2120, which, if you divide by the number of active elements,
- gives you 1010, which is your association. So these distributed associative matrix are really
- clever. They can do all sorts of amazing things with different types of anatomical circuitry,
- content addressability, pattern completion and so on, and this is the simplest form
- of memory system. Much more complex ones called deep learning with different sorts of learning
- rules, and now, at the heart of what the people up the road at King's Cross in DeepMind are doing,
- really giving us great insights into how computational techniques may help
- us with aspects of artificial intelligence. Now, why present this to you? Because if we
- look at the circuitry of the hippocampus, it's just like what I've just shown you.
- You've got the dentate gyrus area, CA3, very like what I've just shown you, area CA1. So we have
- lots of reasons for thinking that the hippocampal anatomical network is a distributed associative
- memory network. Now there are different flavours of synaptic plasticity in the hippocampus.
- The vanilla LTP, LTP1 decays to baseline on a timescale of about three to six hours. But there's
- another one which depends on RNA translation and the synthesis of plasticity proteins, I call it
- the real McCoy. It lasts for much longer periods, at least 24 hours. Then there's an LTP3 as well.
- So that if we were to make a memory where we've got memory trace length on the Y axis
- and time on the X axis, then we can store the information, but it'll either decay away with
- trace decay or it may get consolidated, and what determines that is dendritic mRNA translation,
- the synthesis of plasticity related proteins, and so on, and they stabilise the synaptic change.
- So some years ago, I published a paper together with Oov Frye, who was then working in Magdeburg
- in Germany, which we called synaptic tagging and long term potentiation. We later in my own lab
- went on and did similar experiments, and here we're looking down the microscope at a slice of
- hippocampus not stained like Francesco's using, but with three stimulating electrodes and one
- recording electrode, and we're now going to try and do what I call a synaptic tagging experiment.
- Now, if you record extracellularly, then excitatory potentials go downwards because you're
- outside the cell instead of inside the cell. So what we would see is waveforms going downwards,
- and we're trying to measure the slope of those waveforms or the height of those waveforms as we
- induce LTP. So here's a classic LTP experiment. We've got a control pathway which is the yellow
- symbols, and then you stimulate with a strong tetanus, just like Bliss and
- Lomo did all those years ago. And you can see that you get about 50 per cent potentiation,
- which importantly is stable from two to about ten hours. So that's with a very strong stimulation,
- but suppose you have something just a little bit weaker, a weak tetanus. Here just four trains of
- five pulses you still get an LTP, and in fact if you stop the experiment after an hour, you'd say,
- well, I've got 20 per cent LTP, which many people would regard as fine, but what Redondo who worked
- with me on this experiment did was enable these experiments to last for ten hours and what he
- always saw is it decayed back to baseline for the weak tetanus. So he was able to show that
- there was a statistically significant decline, but now suppose you put the two things together.
- What he then routinely observed is that the weak pathway does not decay. So what's
- going on here? We did further experiments with drugs, but the gist of the idea is as follows,
- let's imagine a neurone, we've got a dendrite sticking out from at the bottom,
- and it's receiving a neuromodulatory input and strong stimulation as well.
- So at the time of strong stimulation, the synapse gets bigger, and that's due
- to actin polymerisation changing the structure of the synapse, and that change is temporary.
- But that strong stimulation can also act on RNA that is distributed through the dendrite
- and cause translation in the white coloured RNA molecules, symbols there, and that gives
- you the plasticity proteins which bind to the tag and stabilise the new synaptic change.
- Now suppose we only had weak tetanisation, we'd still set a tag, we'd still get this early LTP,
- which I've shown you decays, but there'd be no neuromodulation and there would be
- an insufficient stimulation to trigger RNA translation,
- but now let's put the two things together, and they don't have to be exactly the same time. They
- could be half an hour apart, or even we found up to an hour and a half apart. Now what happens is
- the strong stimulation again on the left, and the neuromodulation causes the RNA translation, the
- synthesis of plasticity proteins, but now the weak stimulation guy hijacks the other guy's proteins.
- So the weak stimulation stabilises also, and you get lasting memory on the weak pathway,
- and we did a lot of experiments to try to test that idea. Of course, eventually, we wanted to
- go back to behaviour. So we went back to our Event Arena in which the sand well moves from one day to
- the next. In this particular day, it's in row two, column two, and then we have a probe test
- either 30 minutes or 24 hours later. What we get, as you might expect from this weak tetanisation,
- is really good memory at 30 minutes, but the memory is all gone at 24 hours.
- So now what we need to do is to bring in some strong stimulation,
- and that strong stimulation might be novelty, surprise, something like that,
- which can upregulate the synthesis of plasticity proteins. We did that just simply by putting the
- animals into a novel box in the Arena, which they could walk around for five minutes, and we knew
- from histochemistry that this upregulated various immediate early genes. So then what we did was
- we did our experiment again and now put this novelty in 30 minutes later, and lo and behold,
- as you can see, 24 hours later, the animals can now remember precisely where that sand well is
- located. So this is the behavioural analogy of the physiological synaptic tagging experiment.
- Now in those experiments, we went on and showed that this effect is sensitive to anisomycin,
- which is a protein synthesis inhibitor, and also to dopamine D1, D5 receptor antagonists as well.
- But we were able to find the source of the dopamine, and this was a bit of a puzzle because
- we knew that the substantia nigra VTA projects to the striatum, and many people suppose it also
- projected to the hippocampus, but it turns out the projection of dopamine to the hippocampus from the
- STN is very, very thin, and so we decided to turn to some optogenetic experiments, and we switched
- to using particular transgenic mice, which had a virus expressed downstream of tyrosine
- hydroxylase. Using this technique, we were able to establish that instead of doing novelty,
- we could just fire up the locus ceruleus, and that would deliver dopamine to the hippocampus,
- and you'd get exactly the same result. I won't take you through the details of it.
- So the model I put to you for an unsupervised learning is forgetting happens all the time
- and it's great because would we want to remember every single event of our day? But you can mark
- things off, and one way you can do it is with novelty. Another is with interest as it connects
- with your prior knowledge in some way. So what I'm trying to show is that if you have some
- period where these plasticity related proteins are available, then the trace strength will remain.
- Now there's a phenomenon which we're all aware of called flashbulb memories,
- that certain events are very well remembered by the particular generation or geographical
- locale for which they're important. My mum never stopped telling me about her joy on V.E. Day.
- I'm just old enough to remember the assassination of President Kennedy. I
- was in Sacramento for Princess Diana's death, and of course, there's 9/11.
- Today is actually the actual anniversary of the Columbine High School Massacre,
- which may not have affected us so much in this country, but you can be damn sure it did in the
- West of the United States. Now, a distinctive feature of these memories is that we also have
- a memory of the many trivial events of these days, and there have been experiments on this to show
- that actually, those memories are somewhat inaccurate, but they're not that inaccurate.
- We do remember where we were. We do remember who we were with. We do remember whether we
- were having a cup of coffee or what have you. So I'd like to put it to you that
- this penumbra of influence that happens from something surprising is part of the basis by
- which we remember these trivial events of the day. So let's now turn to losing.
- As I've said already, we all forget, and I've tried to put to you that this is actually
- happening much more than we realise. We think that boosting memory is the
- most important thing. I actually think boosting forgetting might be quite a good thing to do too,
- but I think it's worth thinking about this. Dan Schacter wrote a lovely book
- called The Seven Sins of Memory. The sin of transience, the sin of absent mindedness,
- the sin of blocking, sin of misattribution, sin of suggestibility, bias, persistence.
- Misattribution, telling a joke back to the person who told it to you the day before,
- that kind of thing. It's not great. So these reflect weakening breakdown of attention,
- implanted by a leading question, very important studies have been done on the legal implications
- of that, and the repeated recall of disturbing information in the sin of persistence. Schachter's
- argument was that these failures of memory, while they're exacerbated by age in many
- people as they get older, complain about them, he says they're not really failures at all.
- They reflect the proper operation of a finely tuned system, not vices,
- but virtues. They're by-products of otherwise adaptive features of memory. Sin of persistence,
- post-traumatic stress disorder. Very disabling. I don't in any way
- underestimate the importance of that and of treating that effectively, but persistent memory
- of successful escape from any life threatening situation is also adaptive. You can imagine that
- that's important for animals. So there's always a yin and yang to this. Now another aspect of
- forgetting is whether the excessive use of digital devices is undermining effective memory,
- something we chat about over dinner, but what's the evidence? Is cognitive function, including
- working or long term memory being affected by too much social media? The question on the minds of
- many parents, particularly of teenage kids. Now, so far, the picture from experimental studies, and
- I've read a few of them, there's a growing body, but it's there's very mixed actually. Some report
- clear impairments and there's worry about enhanced impulsivity and reduced attention span, but others
- even report memory enhancement. So the truth is we just don't know yet but it's a very active
- area of research in experimental psychology. Now, forgetting can become much more serious,
- and of course, this brings me to a topic that many of us think about of old age, dementia. Now,
- Alzheimer's disease is the most common cause of dementia. It affects millions worldwide. There's
- no cure or effective treatment, and in many respects, it's a disease that affects the carers
- as much as the victim, as emphatically described in John Kelly's account of his wife, Iris Murdoch.
- It goes through stages. There's a forgetful stage,
- then there's confusion, then there's severe dementia, and then a terminal final stage.
- The pathological hallmarks of this are called plaques and tangles. The tangles are shown in
- black here with an arrow, and the plaques, this sort of orange, in this slide from Dennis Selkoe,
- who's one of the world's experts on Alzheimer's disease, published about 20 years ago. Now, this
- is a very important area of research, and I think as well to say that here in this country, there
- is great strides. Not only was there the original discovery by John Hardy and his team at St Mary's
- of the genetic basis, identifying the specific genes that were the basis of familial forms of
- Alzheimer's disease, but also the research has expanded in many universities across the country,
- and I think a very welcome innovation from David Cameron was the creation of the Dementia Research
- Institute here in the United Kingdom, which I know Mark Walport played a key part in creating. So I
- think that we can be comforted to know that there is exciting research going on in this country
- trying to investigate this, but there are some puzzles as to what is the chicken and what is the
- egg, because it's all very well seeing certain forms of pathology but end stage pathology may
- well be compensatory rather than causative, and it's very hard to work out what's what.
- Does the neuronal injury cause the altered tau metabolism or is it the other way round?
- And so on. Now, I became quite interested in this following a paper published in Nature
- in 1995 by a company in South San Francisco called Elan Pharmaceuticals. They showed the
- first successful mouse that developed plaques, didn't develop tangles, but it developed plaques
- and it was published in Nature in 1995. So the very same day, I sent them a fax,
- because that's how it was, and I said, are your mice forgetful?
- They got the spirit of my approach because they replied also in one line saying, would you like
- to find out? Which was great, which was sort of what I wanted. So began a collaboration
- with Elan. I never took any money from them because really what I wanted was just the mice.
- So just briefly, let me take you into a personal foray of animal models of AD with the so-called
- PDAPP mouse, which overexpresses a human minigene with John Hardy's mutation. What we found was that
- these mice are indeed forgetful, and we were fortunate to publish a paper a few years later,
- and here's the data. So if you're a non-transgenic animal, you forget a little bit. As you go from
- young to middle age, you can still do about nine problems. But if you are a PDAPP animal
- developing these plaques and tangles, then by the time you're old, you can learn about
- half the problems, and that was true both in a cross-sectional study and a longitudinal study.
- So then a key member of the team, Dale Schenk, came up with the idea that it might be possible
- to immunise with antibodies against excessive beta amyloid. So the idea is that this a-beta
- leaches out from the cells and forms into oligomers, which then disrupt synaptic function
- in various ways, and there have been some very nice electrophysiological experiments on that,
- and eventually form into these huge great amyloid plaques that you see at the bottom here.
- Now Dale Schenk then did develop an antibody called 3D6, which if he gave it to PDAPP animals,
- whereas the control PDAPP animals would show these plaques, the vaccinated animals were
- completely cleared. You can imagine this was an incredibly exciting moment for the company.
- So we teamed up with them to do a mouse study, and the idea was to immunise in these animals
- which we knew would start to develop plaques about eight months of age, either immunise
- them when they're about four months of age and then carry on or wait until they're 11 months
- of age and then immunise them. Our study was only a partial success in rescuing the memory deficit.
- You can see from these grey bars that there were some animals which were performing well
- at about nine problems correct, but most of them very much poor and highly, significantly poorer.
- Now, this paper hardly cited at all, buried in the literature, nobody knows about it,
- but I think it makes an important point, which was that only the animals which were immunised
- prior to the onset of plaques showed any rescue. If you waited until they were 11 months of age,
- nothing happened. You couldn't do anything.
- Now these mouse models are very poor. They're sort of basically furry test tubes,
- and there are all sorts of problems with them. So I don't want to make a big claim about it.
- But we were publishing in 2007 that if you want to immunise against Alzheimer's disease, you have to
- catch the people before they get it, and that's a bit of a problem. So we've got a challenge
- for the early detection of Alzheimer's disease and this is a big focus in the DRI institutes.
- So let me now just show you what's called PET imaging, positron emission
- tomography imaging of amyloid plaques using something called the Pittsburgh Compound,
- a paper published in the New England Journal of Medicine in 2012. So what you will see is a brain
- 25 years prior to onset, they used familial cases and they knew the onset time of the parents.
- Watch this. You can see the time clock. We're now 20 years before the onset. 15 years before.
- Ten years before.
- Five. And now we're approaching the point at which people are diagnosed.
- So we've got a sort of huge challenge here, which is can we develop biomarkers,
- things like the Pittsburgh Compound and others, which can give us an insight into what's
- happening. Well, fortunately there have been some breakthroughs with antibodies. Here's lecanemab,
- which was just published last year, which does appear to slow the course of the disease,
- but they had to use people in very early stages. So let me just end with my concluding thoughts.
- What I've tried to do is to take you through how we neuroscientists look about the question of how
- the brain forms memories. I've said a little bit about the brain regions and activity patterns.
- One of the key points here is that there's a great deal of interest in brain imaging, people doing
- experiments on humans and neuropsychologists, on the whole idea of the hippocampus being involved
- in imagining things. Well, I put it to you that our little rat experiment is a bit like that.
- He's imagining where he's got to go to get the food, and we can pick this up in the activity of
- the cells in the start box. It's the same sort of idea. The second to do with the keeping of memory.
- My basic argument is it's pretty important that most things are forgotten, and that we just select
- a bit of it and that selection process we've investigated so far only in relation to novelty,
- but I'm sure the same kind of ideas would apply in relation to connection with established interests.
- Finally, with respect to losing, whether it's benign forgetting or more serious, this is one of
- the challenges of our age, to try to develop ways of spotting the onset of these conditions earlier.
- So to end then, I'd like to explain why I'm so excited about memory. It's something that
- helps to define our own individuality, the stories we tell, the memories that we have,
- and I think shared memory within families and amongst friends acts as a kind of glue
- that holds us all lovingly together. So with respect to your memory, cherish it. Thank you.
- Well, thank you so much, Richard. I'm sure we've got time for a few questions.
- There are microphones, hopefully, roving around. Who would like to ask the first
- question? John?
- No, there's a microphone so the people at the back can hear.
- There's people online who are listening to this, so we need to.
- Is there any downside to immunising everybody with lecanemab?
- Is it working?
- Is there a downside to giving everybody lecanemab?
- I don't know the answer to that yet. The first study, the New England Journal paper,
- suggests there is some slight increase of adverse events, and I guess Biogen is
- looking into what that's about before we can. So there are a few steps ahead before we know.
- Another question at the back.
- hand.
- of studying the brain as modules rather than networks? Which one is better, studying as
- individual components like the hippocampus or as like neural networks that are all together?
- So the question is what are the advantages of different types of neural networks or?
- No, of looking at the brain as individual modules or as neural networks?
- Well, it depends a little bit on your level of analysis. For people doing,
- say, human brain imaging, they've got very interested in networks and very interested in
- effective connectivity between networks. It's not just a field of sort of blobologists anymore. So
- I think we're all interested in networks at some level,
- but what you'd see in human brain imaging is the activity of maybe 10,000 or 50,000 cells,
- not of tiny groups and that's where the animal studies come in, because you can actually
- start to study things at the level of single cells, which you can't really do in humans.
- The example of a distributed associative memory I gave you was a sort of slightly Mickey Mouse
- example, but it explains the basic principle that you don't just make synapses stronger, you embed
- them into particular networks and then out of that comes some very interesting properties.
- If you look at other networks, say the striatum, there, you've got a kind of a network where
- there's re-entrant connectivity between the hippocampus and striatal circuits themselves,
- and the role of dopamine as a reward predicting signal coming in and the intricate circuitry is
- perfect for doing precisely that. So it's set up beautifully for doing the learning of skills
- and very different from the hippocampus. I think one has to look at these in different ways. The
- cerebellum is also different from the point of view of timing but we need to study these things
- in much more detail to get a real understanding of what's going on. In particular, I think we need to
- understand more the role of inhibitory neurones. I haven't mentioned them in this talk. I thought
- of doing so, but I thought, well, it will just get too complicated, but they actually can sculpt
- various aspects of dendritic computation, which is another very exciting area of research.
- Where do you see sleep as fitting into the processing of memories?
- Fascinating question. Very important, Tony. Sleep research was in the doldrums until about 20 years
- ago, because it was all based on sleep deprivation experiments, and the problem with sleep
- deprivation experiments is you induce all sorts of other things other than just deprivation of sleep,
- like stress and so on, but some innovations were made about trying to look at what was happening
- in the various different stages of sleep, the early slow wave sleep, which seems to be very
- important for the consolidation of declarative information, and the later REM stages of sleep,
- which seems to be important for procedural memory. There have been quite a number of studies,
- a lot of them led by Jan Borne in Germany, which is pointing to that. So that's at one level,
- in terms of the psychology. Now if you get down to the neurophysiology,
- then what people are very interested in is so-called sharp waves and cortical spindles,
- which may be part of the process of transferring information from the hippocampus to the cortex.
- We're not entirely sure, but it looks like it is. Since these patterns of activity have such a very
- specific waveform, people have developed very clever electronic tricks called ripple killers.
- So when a ripple starts, you can kill the ripple instantly, and if you do that in animal
- experiments then you can show that they don't remember so well. So that suggests that there
- is a causal role for these sort of particular ripple events. Now over and beyond that, I think
- there are various other aspects. One I would like to mention, which may not be directly related to
- memory, but I think is very interesting, is the ideas of Maiken Nedergaard and the so-called
- glymphatic system. The idea is that there is some sort of paravenous flow, which passes out through
- the brain and sort of cleans the brain at night. It's a kind of night-time housekeeping is her
- argument. There's still some uncertainty about those experiments, but her study in Science in
- 2013 is, I think, one of the great papers that's ever been published about sleep. So I think there
- are lots of different dimensions to sleep, and I think we need to pay much more attention to it.
- Thank you. Do you believe we can improve our memory? I mean, why some of us are
- more forgetful compared to others.
- kinds of answers. I'm sure we can. One approach is to try to develop drugs which improve aspects
- of glutamate transmission or LTP, that kind of thing, and various compounds have been developed.
- A lot of interest at one time in D-cycloserine, which seems to make NMDA receptors work a bit
- better. There's also interest in various types of nicotinic drugs, and experiments
- have been done which suggest that this can be helpful. However, when we come to humans, I'm
- slightly more sceptical of whether that is the right way to go about doing things,
- because I think our ability to remember well depends so much on how well organised our mind is.
- So what I would say to a kid trying to learn something, and what I say to the students in
- the university which I teach, is get your mental models, your frameworks organised,
- and then once you've done that, you'll find it's so easy to add extra information. You need to have
- some kind of framework. So the metaphor I give to you is the difference between what's happening
- in the attics of so many of our houses where we just throw things up and it's a complete chaos,
- from one where there's a superb organisation and if you've got superb organisation, you can go and
- get the information quickly. So I think that that mental effort in organising information probably
- would be more effective than any drug, but I don't know. I think we should pursue both approaches.
- That was wonderful. Richard. What's going on in the mind of people who have a sort of ridiculous
- level of memory, who can remember pi to umpteen thousand decimal places, or to
- the autistic savant like Stephen Wiltshire, who can sit him down in front of St Pancras station
- and he will draw it in incredible detail?
- know. In the case of somebody like Stephen, I mean, the drawings are just spectacular,
- aren't they? They really are extraordinary. I think that the people who become memory experts
- do use this sort of memory theatre approach, and they'll build frameworks, as I was saying to the
- lady who asked me earlier, and then they'll place information into this and they'll build a story,
- and then once you've got a story, it's a bit like a piece of music.
- Each note takes you to the next one. I think that's the sort of trick that they're using,
- but how you get to the levels of proficiency they have, I just don't know. It's staggering, yes.
- There was another question. Another person put their hand up there.
- Some people will be getting thirsty.
- Good evening. What role do you think hormones, especially female hormones or the lack of them,
- during pregnancy, menopause, perimenopause have in making, keeping, and losing memory?
- To be honest, I don't know the answer to that. I think there has been quite a lot of interest
- in how various different types of hormones, including adrenal hormones, can influence
- the way in which memory systems operate. For example, we do know that stress, in addition
- to making us sort of more inefficient in cognitive processing, actually has structural effects on the
- brain. Some brilliant work done by Bruce McEwen at Rockefeller University showed it actually changed
- the structure of hippocampal neurones if you had stress. So that's something that's mediated
- by adrenal hormones, and there could well be other influences of hormones across the oestrous cycle,
- in other ways. What happens at the time of something like the birth of a baby,
- it's beyond my competence to talk about, I'm afraid.
- I think we now ought to draw this to a close. I'm sure there are many more questions, and you
- can address your questions to Richard during the reception, but for now, I'd like to thank
- Richard enormously for this hugely entertaining and informative lecture, which I think has brought
- new aspects of memory to all of us. It's obviously a topic that that we're all fascinated by and is
- ever important in our minds. So it remains only for me to give Richard this scroll and
- the Ferrier medal.
- Thank you so much, Annette, thank you very much indeed.
- Now there's a reception outside at the back, is it?
Ferrier Prize Lecture 2023 given by Professor Richard Morris
The concept of memory is used in many branches of science. In neuroscience, it refers to experience-dependent changes in the nervous system that collectively constitute memory traces of varying accuracy, and from which we can later recall earlier events, places, facts or learned skills. Remembering the birth of a child, the layout of the city where we live, that canaries are yellow, or a well practised tennis stroke. Memory is important and it helps each of us to travel in time and so define our own individuality.
Analysis in both humans and animals typically distinguishes the separate processes of encoding, storage, consolidation, and recall. Exciting major advances have been made in recent years that reflect deepening understanding of these processes. The making of a memory trace about an episode is now believed to involve specific patterns of brain activity that release the major neurotransmitter of the brain – glutamate. This binds to N-methyl-D-aspartate receptors at synapses in the hippocampus that act as coincidence detectors to trigger memory encoding and to tag specific synapses. This discovery rested on the shoulders of brilliant earlier physiological discoveries about synaptic plasticity.
Keeping a memory - storage - involves a different set of so-called AMPA glutamate receptors that are shuttled into the synaptic junctions between neurons to help enhance their strength. Embedded within appropriate neural circuitry, the result will be a set of distributed memory traces mediating altered connectivity across large numbers of neurons and their synaptic connections. Cellular consolidation, like the fixing process of traditional photographic images, may then kick in to enable a subset of these traces to be kept for sufficiently long to be eligible for the overnight brain-wide component of consolidation that occurs during slow-wave sleep. In the absence of consolidation, forgetting can take place, but forgetting is deceptive as it is sometimes true loss but other times merely a failure to access memory traces that may be still there. Memory loss is feared, but all too often forgetting is benign – a valuable feature of a system that guards itself against saturation. It can, however, become severe and notably in neurodevelopmental disorders and neurogenerative conditions such as Alzheimer’s Disease, a condition about which there is now growing hope that we can ameliorate facets of the disease process or at least the symptoms of memory loss associated with it.
About the Royal Society
91TV is a Fellowship of many of the world's most eminent scientists and is the oldest scientific academy in continuous existence.
/
Subscribe to our YouTube channel for exciting science videos and live events.
Find us on:
Bluesky:
Facebook:
Instagram:
LinkedIn:
TikTok: