Our Quest to Understand the Brain – with Matthew Cobb

right so i'm going to be talking about the idea of the brain and the first thing which is kind of difficult in getting your head around is this why do we think about the brain at all why do we think the brain is important and indeed for most of the time for most of us history most people have thought that thinking is done with the heart there's a fairly obvious reason for this because when you're excited it's your heart that pounds if you're frightened then it's your guts nothing much seems to be going on inside your head and we can see this we can see this significance of the heart just by some words that we use these phrases that we use in english same similar phrases are found if those who speak other languages will know them try replacing any of those hearts by brain and it just doesn't work it feels clunky so our language has like a palancest this layer of old ideas sitting in there because they make perfect sense so working out that it was the brain in the first place was really quite tricky and in general for this for most things in science we have to go back to the ancient greeks to find out what they thought about it and aristotle who's often the main man in this he thought that the heart was the center of thought because that makes sense because that's really what it feels like now some people began to disagree with him uh they were in the 5th century bce there were a series of dissections that were taken out by these people herophilus and aerostratus and they began to suggest that because the nervous system all the the organs the sense organs seem to be at the top that maybe it was the brain that was significant not the heart and hippocrates uh also argued that but the key thing is they didn't have any actually proof they couldn't show that this idea made more sense than the fairly common sense feeling that well my heart's pounding when i'm excited therefore it's my heart that's significant now an experiment was done which i'm not going to go into in any detail because it is pretty horrendous but suffice it to say that galen who did not look like that uh but galen who was a greek roman turkish uh surgeon and also a poet and a philosopher as well as creating what passed for medicine for most of uh the last 2000 years in in the west um he was able to show that if you stop the heart the pig was a pig was still conscious on the other hand if you pressed on its brain it went unconscious so he had some good experimental evidence that suggested that the brain was the significant organ but despite this you know people didn't agree and the followers of aristotle who remained incredibly dominant in western thought continued to argue no it's uh it's the heart that counts oops right and gradually the idea of the brain being significant did develop and this idea of ventricular localization so the ventricles are these big holes we have in our brain which are full of fluid and the idea began to circulate that these fluid filled bits of our brain are actually where things exist so there was memory at the back and then uh thinking in the center and imagination at the front there's no proof for this but this idea was repeated for over a thousand years people believe this and we're absolutely convinced that it was true and we can see this problem that people are tussling with right up until the 17th century in shakespeare when in the merchant of venice we have this lovely song tell me where is fancy bread or in the heart or in the head so shakespeare knew that people were still uncertain about what where this thinking business was going on because it feels like it's in your heart but the heart as was soon to be demonstrated is just a pump whereas the brain anatomical investigation showed us were is a really really complicated organ eventually by the late 17 late 18th century people were confident that it was indeed the brain and this is from joseph priestley a famous chemist so uh he was from yorkshire so you'll have to imagine this read out in a in a yorkshire accent a bit like alan bennett or something and he says in my opinion there's just the same reason to conclude that the brain thinks as that it is white and soft so it's self-evident for priestly there's no instance of any man retaining the faculty of thinking when his brain was destroyed and whenever that faculty is impeded or injured there is sufficient reason to believe that the brain is disordered in proportion and therefore we are necessarily led to lead to consider the latter as the seat of the former so what's happened over a period of hundreds of years is not as you might expect a decisive experiment there was no single moment where bang oh the apple falls from the tree there must be gravity there wasn't that for neat for that for gravity either but there certainly was no moment in showing that the brain is where we're thinking instead you've got this gradual certainty through a whole series of anatomical and experimental studies but as soon as you start realizing that it is the brain that's where we're thinking and everything then the question is comes how does it work and this is kind of the theme of the book and a lot of what i'm going to be talking about because throughout the last 300 odd years thinkers and scientists have used technology as a way of explaining what the brain does so we've got the earliest example of this is from descartes who in the 1630s observe these amazing hydraulic statues in paris parks and these hydraulic statues uh here we've got i don't know here's samson or somebody's going to bop this dragon on their head and simply by the movement of the water this would move down and hit the dragon in other cases a piper would emerge from the undergrowth player pipe and then go back into the undergrowth all controlled by hydraulics and so descartes thought well if that's how these machines are working maybe that's how our brain works and there are fluids going down through our nerves such that when we're burned this message goes up the fluids go up and then it comes back down exactly like a statue so descartes is trying to use a metaphor an analogy of hydraulic power which he can see in front of his eyes to explain how the brain works now scientists straight away went away and chopped up nerves and if it was hydraulic power you'd expect kind of there to be gushing fluids coming out of nerves when you chop them in half nothing like that happened so people were fairly soon convinced that descartes was wrong it was nothing quite like the hydraulic power that he was able to observe but there was also another technology developing clockwork and this is an absolutely remarkable uh automaton created in the 1770s by a swiss watchmaker this is from a bbc four programme about machines and we're going to see this is a little you can go and see it it's a nice chatel in switzerland in um in a museum and yeah it writes just watch watch its eyes when we get a close-up so we can write letters oh never writes the same letter creepy eyes so this clearly can produce something like a you know a facsimile of human behavior but when you look inside the brain there's nothing like cogs so there was a problem with this metaphor you had something that looked very human-like but it was repetitive it always did the same thing and above all there was no way of finding what the what the physical substrate the equivalent of those cogs and wheels might be and the real breakthrough came actually in this room partly in this room and that's the discovery of electricity the discovery of electricity and the fact from the middle of the 18th century it was shown the electricity could move produce movement in frogs and then in 1804 there was the electrical stimulation of bodies that was shown to the public and humphrey davy in this very room carried out similar kinds of experiments and a young girl called mary godwin may have actually been sat on these chairs and been inspired by the power of electricity to move bodies to write her thriller horror shock novel frankenstein so this is to give you an idea quite how gruesome it could be this is aldini's description he didn't do this at the royal institution but he did it elsewhere all over europe i placed two cow heads in a straight line on a table in such a manner that the neck the sections of the neck were brought into communication merely by the animal fluids this is the blood and gunk and all the rest of it when this arranged i form an arc from the pile that's a battery because it's formed of a pile of uh disks but i for an arc from the pile to the right ear of one head and to the left ear of the other and saw with astonishment the two heads make horrid grimaces so that the spectators who were no suspicion of such a result were actually frightened people were terrified the fact that you could apparently even bring dead humans back to life by uh using electricity and the once you got the idea of electricity then as soon as that was put into a technological uh situation with the development of the telegraph system in the 1830s then people immediately drew a parallel between the telegraph system and the nervous system of the human body to the extent that somebody who was one of the the people i'd never heard of before i started researching this book one alfred smee who was quite remarkable inventor and all-round crank he said we really have electro-telegraphic communication in the nervous system that which is seen or felt or heard is telegraphed to the brain so he's actually literally saying this is completely the same thing as what we see with the telegraph system with a message going from uh one town to another and he went even further he actually tried to build a brain so this is much more ambitious than babbage's attempt to build a calculator what smee thinks he's going to do is actually build a brain and this is one of the parts of it and i've stared at this for hours i have no idea what he's talking about but he's convinced that this gives an analogous representation of a natural process of thought as perfectly as a human contrivance can well be expected to afford now all this is made out of brass and i think it kind of hinges and moves and he's trying to get some idea of conditionality he claims that it can represent a syllogism so he's trying to think about logic but despite his interest in electricity this is all about mechanics it's all brass and so on he comes up with a nice little explanation of how he thinks things are represented in the brain um and he claims that this is this is a lower animal so this could be a bird and this is a human so we've got muscles down here we've got the nerves going out and then the brain which is basically just lots of interactions and he claims that this showed how the idea of a nest may be implanted in the in the brain should say the brain of a bird of a comb in the wasp or b and so on so i mean again this is a nice idea but in fact has absolutely no bearing on the reality of how brains are organized or how behaviors organized and part of the problem with this if you just look at this very very interactive scheme is that there's no suggestion really that there's much localization going on that there are things being put in happening in one place that are significant and this is surprising because at this very time the whole of the united kingdom in fact the whole of the world became obsessed by what's called phrenology and everybody from karl marx to queen victoria was obsessed by it when moriarty meets sherlock holmes for the first time he makes a very dismissive remark about holmes's head so it was absolutely everywhere in popular culture throughout the 19th century and this idea was basically that there are bits of the brain that do different things now there's huge argument about how you could actually space these out on the brain above all nobody was actually looking at the brain at all instead they were feeling the lumps on the outside of the skull which is why this is just pure bunkum because you know your brain is inside your skull it's got all those bones you've got muscles on the outside you can't actually feel the shape of your brain uh through your skull and whatever lumps you may have on your skull don't tell you anything about your character but i didn't stop many people and novelists and queen victoria and karl marx from believing that this was actually true so it became in the end a bit like astrology today so it was a kind of not entirely serious folk belief but it carried on for right into the 1930s people were still convinced that you could tell something uh about people's heads about their brains from feeding their heads because there was localization of function now this actually took on a far more real idea with the work of this chap david ferrier who uh when he was only 27 he used very precise electrodes to show that different parts of the monkey brain control movement and perception so he could gently stimulate different parts of the cortex of the monkey and it would start to move its hand or it would prick its ears up as though it heard something so he was able to do two things firstly to show that there is localization of function in the brain there are bits of our brain that appear to do things particular things it's not just one incoherent mass but secondly he could draw a parallel between what is in a monkey's head and what's in a human's head and the numbers you can see are the same now he didn't do any experiments on a human but he could show that anatomically these are parallel areas and the only thing that seemed to be different between the monkey and us is what i'm doing now and that's speech and that's this area here called broca's area which was discovered in 1861 through a series of dissections of patients who'd had strokes and have been unable to speak and to brocker's great surprise he discovered that they all had lesions in this left frontal part of their brain and this was finally accepted that rather disturbingly it means our brains are asymmetrical so there's one part of our brain is doing something on this side which doesn't happen here and secondly although there was no clear anatomical difference between us and a monkey there was a clear functional difference we can't see that difference between a broca's area in a human and broker's area in a monkey but it's there at the same time scientists looking at the brain began so we got this idea of localization they began to realize that the brain is actually doing something quite complicated it's not just sitting there and absorbing stuff so helen holtz the fan most extraordinary physiologist of the 19th century he said that the brain makes inductive conclusions about the world and he made this on the basis of things like the blind spot so all of you right now are making an inductive conclusion about your visual field because this part of your visual field you can't see because that's where your optic nerve goes so there's no light detecting cells there but your brain basically partly because you're moving their eyes a lot but even if you keep your eyes steady your brain fills it in a bit like in photoshop it just copies it over we're not going to worry there's a little bit there i can't see it must be the same so you're making conclusions about the world you're not simply passively accepting what's out there you're processing it and even imagining without even thinking about it your brain is making these inductive conclusions and then lloyd morgan an early psychologist he came up with the idea that actually the key thing that our brain does is to control what we do and to control our otherwise active reflexes you can see this uh with an animal if you remove parts of its outer parts of its brain then you get uncontrollable reflex action and he realized that one of the things that humans are very good at doing is stopping that uncontrollable behavior so we've got quite a complicated view of the brain and it seems to be able to choose between different outcomes and possibilities but that's not what a telegraph system's like we go back to the metaphor telegraph you send a message from london and it goes up to manchester you've labeled it to manchester it's going to go up there but a new technology appeared that seemed to fit the bill the telephone exchange so all the young people who have no idea what a telephone exchange worth and how i know you just go in your pho complete mystery but this is how it used to work you'd phone up and you pick up the phone and the light would come on and a woman generally was a woman would then listen to you and she'd connect a cable to your light and then she would put another cable to the number you wanted so if it was in your local exchange you would literally be connected so you've got this flexibility if you wanted to talk to somebody else in a different region then it would be long distance and she would put it into a slot taking your message up to another uh exchange in manchester say and then you could explain the same thing again so you've got a an electrical system that is incredibly flexible and scientists started to use this idea and indeed i've still come across in you know prestigious articles in neuroscience people say well the brain is basically like a telephone exchange as i say i'm not sure that anybody under the age of 40 has any idea what that means but this is the idea you can see it's got flexibility in there within a certain limitations and in this very room in 1915 the massive racist sir keith siratha keith he was a really massive racist nespac it's true isn't adam he was a very bad man anyway um he gave the royal institution christmas lectures not about race um in the middle of the second world war and he drew this parallel here between the uh telephone exchange drawn here and the nervous system of a reflex arc so just like when you bash your knee and your leg comes out that this is very similar he says if you think about it a minute actually it's not at all similar because i when i have a reflex arc i can't it always does the same thing that's the whole point about it i can't control it it just goes up in here and down there but above all what's actually going along the nerves how can we understand this thing that's going is it like a telephone call or what how can we best understand it and around the same time because one of the things i learned about studying history is this very rarely just one person one brilliant person it's always the ideas are all there bubbling around at about the same time this chap meisner made what he called an it was an electric dog basically it's got three wheels and he shines a light at it and then the the photoelectric cell here drives the electric motor and it will move towards the light so you've got apparently purposive behavior in a very very simple device but there was even something even more amazing uh which this chap here alfred locker locker is well known for to ecologists some of you have made dimly remember hearing about lotka volterra equations which about how uh for example predator and prey cycle but locker did a lot of thinking about other things and he noticed this ladybird and i have one here because this is the wrong institution and so we have demonstrations it's not going to blow up i'm afraid um but it really i i got this especially because i thought i could share a video but i'm sure i can get hold of one so this is from 1925. locker saw this let's see whether it does what it does oh it's impressive no it really won't fall off okay we're going to leave it there that's the problem with demonstrations you've got to stop them it's not going to it won't go off it's going to carry on making noises um adam hold this don't let it make it don't let it make a noise there you go my glamorous system it's just put you just hold it i'm going to show you how it works that comes now so locker showed everybody how it works i'm happy to demonstrate it you can buy them they're still made in czechoslovakia it's exactly the same basically this is all it is you've got wheels here and the little antennae hold its head up so that such that if it goes over the edge of the the table the antenna drop down and then while still having the driving wheels pushing forward this freewheel which is not powered and is in fact set at a slight angle that now hits the ground and that turns it away so with a very very simple feedback loop you've got purposive behavior it looks like the damn thing doesn't want to fall off the edge but of course that's not the case you know it's clockwork and locker this is in 1925 this is before people started to mathematize the idea of information he said the toy is construing the information which is just bizarre way of thinking about it because this is just a clockwork thing but by taking something incredibly simple you can see how purposive behavior can emerge right what i'm going to do now is i'm going to stop so i'm a scientist and i know that everybody dozes off after about 25 minutes so i do so in all my lectures i stop for a minute and a half so i'm going to stop you can check your phone see what's happening with the virus are we all dead are there zombies outside and then we'll come back okay so a minute and a half just relax you'll see it works yeah you can let that make a noise put it on there [Applause] [Applause] anyway see that was better wasn't it see now we're all awake it's good okay well you can't go to sleep late right so here we are in the 1920s and we've got locker talking about a silly little toy construing information and at this point scientists are starting to think in this new way about how nervous systems might work and probably the person who did most for this is i think the most famous scientist that virtually none of you will have heard of an extraordinary man called edgar adrian not only did he win the nobel prize his two proteges won the nobel prize his son became an frs he was vice chancellor of the university of cambridge he went on to become chancellor after his retirement at the university of leicester an absolutely extraordinary man and he was a neurophysiologist he wanted to see what neurons did and what he came up with in the 1920s was the idea that nervous systems contain they transmit messages they contain information and there's a code and you can see the code here this is from different uh frog neurons and he's making them he's pulling on them with a heavier and heavier weight and you can see that this is the response of the neuron it's called spiking and you can see there are more and more of these with increasing weight now he didn't only study these isolated neurons he also studied himself at the end of the 1920s people discovered that if you put very sensitive electrodes on the brain and then used amplifiers which had been developed in the period immediately during and after the first world war you could record what was going on and he he did this with himself so this is him edgar adrian and he's got his eyes shut and you get what's called the alpha wave and then he opens them and it disappears and then he shuts them and it comes back he could he became a dab hand at this he did this in demonstrations at conference he'd sit down there and close it and be firing away but he didn't only do it in people getting a water beetle and you get exactly the same thing water beetle in the dark it's got this rhythm all brains seem to have it they're just kind of pulsating away turn the light on the water beetle starts it thinking it's water beetly things and it goes away and then it comes back again so there's something happening there's electrical activity in brains at different levels there's this very simple code that you can see in neurons responding to sensory stimulation and there's much more complicated stuff which we still don't really understand of this kind of what's now called an eeg taking place in the brain of you but also of a water beetle final uh person i'm going to talk about and this change of ideas is another man you've probably not heard of his name was craig and he died just before ve day in cambridge he was knocked off his bicycle very young man in his early 30s and he came up with the idea that one of the key things that brains do is to parallel or model external events in other words your brain is getting those external features the environment and thinking about what might happen we're going back to helmholtz's idea of these inductions making predictions and craig has this very significant idea about what brains are doing in terms of representing the outside world and then doing something with them in ways that remain pretty mysterious final two uh weirdos i'm going to talk about are um especially this character here um this is walter pitts left school at the age of 14 but was one of the most the most brilliant mathematicians uh of his age and this chap here warren mcculloch and in the early 1940s they're trying to think about what how nervous systems work on a very very simple abstract way and they came up with an idea which was completely false or largely false but was very very influential and they suggested well if we've got nerves wired together we can get what's called boolean logic don't worry about the boolean business you can think of this to understand what's going on so here this neuron 2 is going to fire when this neuron fires so the signal will go from there this one here will fire if either this one or this one is firing so it's ore just like you've got in your computer programs this one here will fire if it's and they both fire and this one three will fire if this one fires and this one doesn't want fire so you've got a not these basic logical functions which have been developed and were uh philosophers of logic in particular extremely interested bertrand russell and so were fascinated by these they said you can see in the nervous system this is the imminent logic they called it of what was going on in the nervous system and this was to prove incredibly significant not in terms of primarily how we understand the brain but in something you're all completely familiar with the computer because when the computer was first being developed not by alan turing but by this man john von neumann the digital computer the one you've got in your pockets he said brains work in this way they use this logic system we should use that to build our computer so rather than the brain being a computer when it was first developed the computer was literally a brain they literally thought they were going to copy the structure of a brain and put it into a machine now as it turns out nervous systems are much more complicated than that but the significance of this idea was well we can see it all around us and just to give you uh some idea people trying to put this into devices and here we've got something which uh you can go and see in the science museum actually this is toby it's got various names it's not really a tortoise and this chap here grey walter uh was a psychologist from bristol and he built a little robot you may see some similarities with celino the robot dog from 1912 but in 1950s britain this was pretty hot stuff dr of bristol why the torch well here's the reason it's turbid a mechanical tortoise with an electronic brain which functions like a human mind it doesn't it doesn't function like [Music] it's brilliant [Applause] magic eye is a photoelectric cell constantly revolving until it picks up the strongest source of light to which it is that attracted in this case an ordinary electric torch guides the mechanical tortoise in any direction its inventor chooses it can also negotiate obstacles when it hits an object the pressure on the shell causes a short circuit of the photoelectric cell mechanism and the tortoise moves at random until it is free of the obstacle very dramatic thunderbirds in music um and actually i think probably speak the rest of the talk like this because that's the way we should be speaking um you can see this at the right at the uh science science museum and in fact toby would even go back into his hutch when he was feeling a bit low and his electricity was dropping he'd go back into his hutch and recharge so you've got this very what looks like purposive behavior but despite the flim flam from the path newsreel commentator walter didn't actually think that toby was conscious it wasn't working like a brain but rather you could get like with a little clockwork device you can get purposive behavior from very very simple structures so this takes us up to the 1950s and my argument is that basically we're still working in the world that these all these thinkers have been describing uh created that's where we still think what we think the brain does so what does a brain do what is our current idea of a brain something like this it contains symbolic representations of the outside world that it manipulates through computations and that enables it to predict what will happen and to produce appropriate behaviors because the brain's not just sitting there it's enabling us to do stuff that's the whole point of it and amongst the processes that take place are feedback inhibition and it's also calculating something about probability so this is the kind of general agreement i think that most neuroscientists would have they'd disagree with various bits of it but this is the world that was created by all those researchers and emerged in the post-war world and we're still using that so the question then comes is how exactly are these processes represented in the brain if we all agree this is basically the kind of stuff that's doing how on earth does it do it and that's when it all gets a bit tricky so the first point is if it's something like a computer well as i've already explained your phones the computers you use are all digital yeah that's what von neumann had his system based on that's the essence of those little uh boolean logic and or not but neurons aren't digital we can see that in adrian's data look each of the then you're only the fires or it doesn't so that's digital you've got a spike or you don't but the way it's responding to increased intensity of weight is to shove an increased intensity of firing it's not digital at all people realize this very very quickly i mean von neumann figured it out very very soon he said no no no this isn't actually going to work in terms of understanding the human brain or any brain because nervous systems are not digital they're doing something else they have a digital component but they also have an analog aspect and they're much more interesting frankly than a transistor what about how neurons meet when they meet when the signal goes from one to another it's a little representation of what a synapse might look like well again they're horrendously complicated it's not like a a link in a computer we have dozens of different neurotransmitters that enable the neurons to respond we have activation and inhibition so this can go lots and lots or it can turn itself off they're about in a human synapse there are about 5 500 different protein molecules sitting there we have no idea what most of them are doing and there are hormonal effects so we have hormones coursing around our body that change we're going to amplify or reduce or alter these activities of our synapses and change them perhaps on a very very long term basis so it's not simply some kind of digital wiring diagram like in a computer something very very different give you some example so this is what i work on this is the maggot and maggots they wriggle and to know that they've wriggled they have these little things called stretch cells which basically says i've been pulled this is what the little muscle the little neuron looks like um and there are loads of these uh along this this isn't even in its brain this isn't its body wall as it's wriggling along each of these cells has 53 inputs 18 outputs is connected to 74 other cells and most if not all of those synapses have more than one neurotransmitter so that's not doing anything complicated that's not trying to work out you know the meaning of life the universe and everything this is simply saying oh i've been stretched you might want to move a bit that's all it's saying and yet you have this incredible level of complexity which we don't properly understand here's uh a tweet by sophie scott professor at ucl uh which with her permission i've reproduced in the book she's an audiologist she's interested in how hearing works in what's called tonotopi how the signal is preserved down the wire and basically this is your ear here and she saw this drawing in a book spent today reading about the subcortical subcortical auditory processing like this bastard one euron away from the cochlear all hell breaks loose eight different cell types five different parallel processing streams all preserving tonnotopi this organization that she was interesting and that's just the start this is just the very very beginning of our auditory processing how do we ever hear anything and you can do the same thing with i work on the sense of smell and it's similarly horrendously complicated as soon as you get beyond those first neurons sending a a response to a sensory stimulus and scientists have actually tried to use the analogy of a computer to see whether we can actually understand it so these two people are jonas and cording they took this mos-6502 processor which the older amongst you will know was the processor that used to play donkey kong amongst other things and they used all the tools of neuroscience they established its connectome its connections they took bits out they measured what the various bits were doing and they tried to understand using the tools of neuroscience how this thing worked they failed completely now there is a there is a logic to this this is the macaque visual system and who knows what the logic is going on there so our tools at the moment are insufficient even to understand something that has actually been designed not something that has evolved and has got all sorts of strange bits going on in it finally what about that fmri you've all seen those images of the brain lighting up brains do not light up this is simply how it's shown this is the the big breakthrough this appeared on the this is the front cover of science sciences one of the two weekly magazines this was a really big deal so they put the image the first time they got fmri images and it got on the front cover of science now there's a rather harsh fact about fmri we'll leave aside as to quite how significant it is just think about the resolution if what we want to know is how a process is represented in the brain those calculations and all the rest of it i mentioned then we need to know quite how fine the resolution of an image like this is or the ones you see in the papers all the time and one vox cell so think of it like a pixel the smallest unit of an fmri image contains about 5.5 million neurons maybe 55 billion synapses 22 kilometers of dendrites that's the input part and 220 kilometers of output i checked all those figures i didn't believe them they're absolutely right now that was from 2008 our systems got a little bit better but they are in the same order of magnitude so when you see one of these images you just need to remember that in terms of actually understanding really what's going on we are just it is hopeless now most of what we understand about the way that the brain works and what it's doing has come through the study of vision humans are particularly visual animals but there's plenty of reasons for trying to understand vision because many of the animals that we study in the laboratory have got very good visual systems and this has produced some quite bizarre and remarkable results that challenge some of the basic ideas we have about how things work and certainly it made me think very much and changed my ideas so one of the earliest findings was they won the nobel prize for this was from two people called huble and weasel and they discovered that as they uh put an electrode into a cat's brain that as they showed different lines of different angles different cells within the cat's brain would get very very excited so it seemed as though there are detectors for lines at different angles in the cat's brain in a kind of columnar organization there's similar things in primate brains as well and they suggested well okay maybe if we've got little dots we've got three dot receptors so these are bits of the visual field where i can see a dot and you wire them up like this and they're in an angle then you this cell here should respond to an angle okay you can understand that maybe that's true but then i've got to put all these images together all these little dots and lines and all the rest of it into my head and i've got to be able to recognize things and how can i do that is that all in one cell and this became a a kind of joke when i was a student and was used to kind of ridicule this whole idea uh that you'd have to have a cell for your grandmother you'd have to have a cell for your grandmother standing up sitting down playing the ukulele backwards it would always be your grandmother you must have a cell for everything in my case it would have to be seeing my grandmother when she was getting married in june 1914 with her three sisters all the gentlemen survived the war i'm glad to say uh i'd have to have a cell to recognize her and i there isn't enough space in the known universe for that so we dismissed all that as just rubbish and then about 10 years ago there were some researchers studying people with profound epilepsy and these uh subjects more patients very kindly said okay you're going to do things to my brain yes before you do the operation you can poke around in there with electrodes and show me things and to their great astonishment they found that in one patient there was a cell just one cell that they recorded from that responded to jennifer aniston if they showed this picture it wasn't interested wasn't interested in brad pitt at all had to be just jennifer aniston and they could have us standing up sitting down all the things that we laughed about and it would always say oh yes that's jennifer aniston it wasn't all rubbish uh there was another chap who uh got very excited at pictures of the sydney opera house so there seems to be an astonishing degree of specificity in the brain now before you get all excited okay we have got this is a cell for jennifer aniston what you've got to remember is they're just recording from one cell but that cell is connected to millions in a network and there's going to be overlapping patterns of activity such that yes that might get very excited by jennifer aniston but maybe a picture of her grandma might the subject's grandma might excite another half of those cells and so on and so what you're getting in fact is not just one cell but whole networks of activity producing our perception the fact that we can recognize people and recently at the end of last year some researchers did a very clever experiment they did a similar kind of thing but this is a was on a monkey and they they put the electrode into the monkey's brain and uh then they sh gave it just a a grey screen and the computer then changed the image randomly and if the cell that they were recording from got more excited then it kept that and then changed something again so effectively they were trying to tune in to what that cell was really really really interested in and it was this so heather knows what monkeys are seeing but there are cells in their brains that are i mean it's supposed to assist in it's in the state that are composed of these strange mixtures that i get really excited when they see this and much less excited if you take one of those things away so overlapping with millions of these different cells somehow you get perception emerging now you might think well okay maybe we can use computers to get at that and people have done this so here's an example you may have seen this this is jeffrey hinton who now works for google and they gave one of their ais 10 million pictures from youtube to look at and they didn't say what it had to do they just basically gave this data to the computer that just went through it and it was looking for patterns and out of these 10 million pictures which of course weren't actually pictures they were data streams so it was a one-dimensional data stream it was looking at patterns of zeros and ones it came out with this you probably can see it this is a cat as it sees there's there's there's it sees there's its face you can just about see its eyes and this if you showed it a picture of a cat that had never seen about 20 percent of the time this could recognize what was going on said it was a cat now you might say okay well this is going how we're going to understand the brain we can use these systems these complicated ai's to get at what's going on well probably not because this is what hinton said and there's the separate problem which is we don't know entirely however these things work right no we have really no idea how they work so this is from the man it's not me he says we haven't the foggy is how it works and you can see the problems with ai with this so uh an ai researcher called janelle janelle shane a fantastic book she used gpt2 now you may have heard of this this is a program which has basically devoured all the fan fiction on the internet and can then write things for you so if you start writing and then stop a sentence it will complete it in the new yorker they wrote a whole article just giving this program prompts at the beginning of paragraph and then it would write a lot of stuff then you give it another prompt and so on and it can just churn stuff out so she very cleverly said okay hey make me some love hearts with your great knowledge and this is what it came up with so i i tweeted swizzles of new mills saying you should use these they'd be much better so the problem is the ai has no idea what it's doing it's just fishing it out and spewing it out they're good so to give you some idea of the problem there is of trying to understand nervous systems and brains i'm going to close very briefly with the work of eve marder who's another scientist who probably most of you won't have heard of she is absolutely brilliant for the last 20 30 years she has been working in brandeis university in bo in boston and america trying to understand the lobster's stomach right now the lobster's stomach has got 30 neurons it's a little bag behind its head and it's got kind of teeth in there and it grinds food up and these neurons these 13 neurons is not its brain it's just its stomach they produce a rhythm they can produce two or three different pulsating rhythms and despite all her smartness despite all the work of her and her researchers they cannot understand how these rhythms are produced by these 30 simple neurons there is something about the way they're wired together that still after all this work escapes them and that's the challenge we have to understand how brains work and how different they are from a computer because even very very simple structures for the moment uh escape us that doesn't mean to say that we're we've got no hope my view is we should be studying small brains the final bit of the book is the future what how can we try and understand what's going on so for oops for my for my money uh we can study platinumis this is the lava of a of a worm that lives in the sea this is one model system we about to have the connectome of this uh drosophila and the drosophila maggot both of which i've worked on drosophila maggot brain has about 10 000 neurons we should have a connectome of this of a maggot just one just one maggot because maggots differ from one to the other they've sliced up one maggot and they will know exactly how it's wired up within the next five years or so this has taken dozens of people years and years and years uh the drosophila fly has about 50 000 neurons in its brain and the zebrafish lava if you really think it's got to be a vertebrate then the baby zebrafish has about a hundred thousand just to give you an idea a mouse has 70 million and you have who knows 80 billion plus there's loads of other stuff it's not all just neurons in there there's all sorts of stuff in there so for my money i think we can by focusing on these simple systems where we can manipulate them we can turn cells on and off we can activate them optogenetically by a flick of a literally a flick of a switch this is some way we've got a possibility we have of understanding brains then modeling them in computers and seeing how particular outcomes may be affected by altering them if we can do that and correctly predict what will happen something that has so far escaped eve mada then we have really been able to say we've got the idea of the brain so the future then what can we draw from all this well you've seen how science culture and technology are all intertwined there's a mixture science isn't outside of culture and it relies heavily on technology for its metaphors to try and understand things said that and these mod metaphors of course that all models are limited they've got tremendous explanatory power but they can also limit you and when you say this to scientists you say well actually you know what you can imagine is actually determined by partly by the technology about you they get terribly excited they they think that's fantastic so that means that new technology is going to alter how we can imagine what the brain does and it will change our view and to which my response is well yes that is going to happen when we have that technology so what will that be well luckily i have a device that enables me to see into the future so we're going to see the future oh it doesn't seem to work you'll have to make it up for yourselves thank you very much terrific

Channel: The Royal Institution

Views: 44,252

Rating: 4.8683386 out of 5

Keywords: Ri, Royal Institution, matthew cobb, brain, neuroscience, the idea of the brain, ai, brain computer, machine learning, anatomy, cognitive neuroscience, brain imaging, life

Id: NifNfkliCos

Length: 50min 37sec (3037 seconds)

Published: Thu Aug 20 2020