 So we will talk about the Neural Colleges of Consciousness and I'm going to make some arguments why the NCC, the Neural Colleges of Consciousness, are not to be found in V1, in primary visual cortex. It's primarily based on experiments on data from monkey and the best data we'll hear in a future lecture and some data based on human psychophysics. And in the course of this, you'll learn something about human psychophysics today, namely about adaptation experiments. So the question is, and some people of course question the premise for this question, but you can ask the question to what extent is any one cortical area, to what extent is any one cortical, and to what extent is any one neuronal population critically necessary for consciousness? Now there are of course some people who say that sort of, that question is nonsensical because consciousness is a property for entire system it's a holistic global property of the entire system and so this question doesn't make any sense. Of course, as I'm paying to point out, the lesson from 20th century biology is that it's all about very specific mechanisms, very specific gadgets, and that particular using modern pharmacological molecular techniques you can perturb the system, you can take out particular component of the system and show that they are very specific deficits. And so to us it makes perfect sense, at least to Francis Crick and myself, it makes perfect sense to ask the question, is a particular cortical area, for example, all the neurons in V1, is that involved in consciousness? Clearly it is involved in visual consciousness in our case. Clearly it is involved in the sense that if I don't have V1 I do not see a lot. So if I'm a patient and I'm unfortunate enough to lose my primary visual cortex, primarily due to a stroke, could also be to a gunshot or to a virus, but mainly to a stroke, let's say I lose my right visual cortex, then I have what's called hemianopia. Okay, so I'm essentially blind in my left visual field. I don't see anything on the left side. That's what you would expect. It's like I take off the left side of my two retinas and of course I don't see anything. And the interesting point is people know that they are blind. They just don't see anything, it's like the back of your head, you just don't see anything there. For the most part people are aware of this deficit. I'm emphasizing that because there are some other diseases or pathologies which we'll study later called the most common being neglect, which is due to a loss of the posterior parietal area. Typically we think it's the right, part of the right posterior temporal lobe. And there you get what's called neglect where you also don't have access to this information, but you're not aware of that. People don't know that they're blind here. They bump into things, they fall, mainly they bump into things, they don't eat things on the left side of their plate. When they drive there they're likely to hit things on the left side. They're just not aware of the deficit. So here we have a situation, a somewhat paradoxical situation, a lesion early on that totally deprives you of all visual input is not a stability to the patient as a lesion in a higher area that we'll discuss more in depth in a future lecture. A lesion in a higher area that leads to neglect when you're unaware of this deficit. I think it's a very interesting point. That's called hemianopia when essentially you're total absolute blindness for things in one particular hemisphere, one or the other depending on which side the injury was. You can also get what's called something, quadrantanopia which is a much more specific deficit when you essentially have a deficit that's limited to one quadrant. So you fix it here and essentially can be limited to a upper quadrant or to a lower quadrant depending on where exactly the side of the lesion is. But the bottom line is without V1 you don't see. Now of course that doesn't mean that visual perceptions altogether eliminated. You can still have imagery and you can still have dreams. Although if you read some of the literature on people who become blind it's interesting some at least one subset of these patients after a while they lose the ability to image and to dream if they become totally blind. That might be sort of a neuronal degeneration phenomena where other people don't. I mean other people have read accounts of people who are blind who sort of still 20 years later have very intense visual memories. Oliver Sacks has written about this. Now sometimes you hear quite a bit in the literature on consciousness there's a phenomenon called blind sight which is of course a contradiction in term, right? Blind sight, you're blind and you can see something. So blind sight is a little, it's a phenomenon that was sort of discovered in the 70s based on first animal studies and then in humans by Larry Weisskrantz and his collaborators in England. And it's a phenomena that you have a patient who has lost his primary visual cortex who says he's blind but then you do some behavioral testing and you find there's actually, to the surprise of the patient there are actually visual abilities that still remain. So for a typical blind sight patient the doctor would, so this is now, I mean I'm the patient and this arm is the doctor's arm, okay? So the doctor would for example test the patient, you know, typically test you see, does this move? Can you tell me does it move to the left or to the right? And the patient says, no, I'm blind, that's why I'm here. And the doctor says, well just, you know, amuse me, just tell me, just guess. You know, does it move forward or backward? And the patient says backward, forward. Or you know, say guess, does it move upward or downward? And the patient says, well I don't know but I'm guessing upward, downward. There's a pattern, right? That these patients, which are pretty rare, that these patients do have remaining visual abilities in the absence of any professed conscious sensation, in the absence of any experience in that part of the visual field. It's not that totally blind, usually they're blind or sometimes the blindness might be restricted to smaller part of the visual field called a scatoma. And in this part of the visual field that's blind people don't have any visual awareness and they still have some visual abilities. So typically what they have access to is things like motion, if it moves fast, not if it moves slow, or if it's a low contrast, but if it's a high, rapidly moving, are those animals? Or is he at the MR machine? We're just below that, they're installing that. So typically these patients, they can detect single things when they move, when they're high contrast and rapidly moving, they might be able to detect orientation. There's some nice evidence from Petra Sturich, she's shown that some of these patients have access to color information. This is all unconscious in the sense that these people claim they don't know it, but if you put them in a situation where they have to choose, like a two alternative false choice, they more often than not choose the right one. More often than not, in this case, can mean 90% or 95%. There's even done some very nice animal experiments in a monkey, where you can also show that in a monkey, of course, it's much more difficult because the monkey can't tell you the experience, you can only signal its behavior, so there you have to do some interesting tricks to test, but she has done these tricks in a very beautiful paper with Alan Cowey, to show that the monkey, from a phenomenological point of view, has affected hemifield. So if you take all of V1 on the right side, this entire left hemifield will be blank, there will be nothing there, and from a phenomenological point of view, it's given the choice to push, so the experiment they did in one monkey, they took away V1, and then they had bright lights at various locations, and he had to, whenever there was a light, he had to push a button, and when there wasn't any light, so there was a tone, and then the light came on somewhere, and when he saw the light, he had to push the button, and when there wasn't any light, he had to push a button called no light button, and when you put a light in its normal side of the hemifield, he pushed the button light. When you put the button in its blind hemifield, he pushed the button, not seen, no light there. So from a phenomenological point of view, apparently things in the blind hemifield looked to them like there was no stimulus there, yet then they did a different experiment where they retrained the monkey, so now the monkey had to touch, there was a touch screen, and the monkey had to touch the location of the light on the screen, so in that case, the monkey was able to do that. So you can show that both insurance and in monkeys, you can have this, again, I have to emphasize, there's a few number of patients only who have this, that there's some residual visual abilities in the absence of any awareness unless all these people are lying to you. Now it's fairly simple. People make a lot of hay made out of this discovery and observation. I think it's interesting for two reasons, mainly it's interesting because it tells you you can have behavior in the absence of any consciousness. We know that, I mean, we're going to discuss a large range of cases like that in normal people where there are all sorts of behaviors that we call zombie behaviors that you can do without being conscious, things like talking, etc. You have a nice form in some patients who can behave without having access to awareness. It's a rather simple form, so if you put up, I've never seen any evidence that these people in this blind hemifield can do complicated things like, for example, if there are two targets, to be able to track two targets, a point where, okay, that thing moves up or that target moves down. All these abilities are fairly simple. And so you can speculate about the pathways and people have done that and most of the time people think the pathway involved is probably a colloquial pathway. In other words, you have the output from the retina that projects down into this midbrain structure that I mentioned in the third lecture called the superior colliculus, which then in turn projects to the thalamus and then in turn projects to parts of cortex. And so that pathway might be in the absence of, you know, in the absence of the dominant pathway where 90% of the neurons go from the retina to LGN2, to primavisual cortex. This bypass is sort of all the evolutionary, all the bypass through the superior colliculus might be sufficient to mediate these simple behaviors. But they don't actually speak, so the existence of the blind side by itself doesn't speak to the fact whether or not you need V1. So clearly without V1, you don't see. Now the same argument, of course, can be made for the retina. Clearly without retina, you don't see. Again, you can have imagery and you can dream, but you don't see. So just because you're lacking a structure, the phenomenon doesn't occur, it's actually causally responsible for directly generating that. And in that sense, I made a point, I discussed in the different ways in which activity in the retina doesn't correspond to what you see. So I made this point that you don't see with your eyes, that the eyes are necessary to uptake the information, to take up the information, convert it from photons into electrical impulses that the brain can process, but you do not see with the eyes. The character of the electrical activity in the eye is too distinguished. In the blind spot, you have this, you know, this blur used to the rapid motion. You have this dramatic fall of cones. You have no blue cones at the center of the fovea. And all those things don't show up in perception. And likewise with V1, it's necessary, but that's not where conscious perception is generated. That's at least according to Francis Crick and myself. So let me talk about some evidence for that based on some psychophysics. So I'm stressing this because it's an interesting lesson in the general tool that psychophysics has used since probably 100, I don't know how long, 150 years or something, after effects. So let me try to see whether this works. So this is an after effect. Now there are different kinds of after effects. The simplest after effect is if there's flash photography, you know, you happen to look into the flash, while it's flashed, and for a few seconds you get these spots in front of your eyes. You might have something burned into your retina. That's essentially, it's after effect, it's short lasting. It's just due to photo bleaching. The fact that the very bright, the discharge sort of led to a dramatic loss of photo pigment in parts of the retina where that image has generated photoisomerization. And for those locations, for a few seconds the brain lacks, because it was very active at those locations, so for a few seconds you're going to get the complement image. That's why, you know, if you see something very bright, the complement are these dark spots. Here you have an after effect that's more specific that probably most people think happens in primary visual cortex or beyond. Namely the orientation specific after effect and spatial frequency specific after effect and next week we'll talk about a very striking one, motion specific after effect, the waterfall illusion. That's probably the most striking one. So let's see whether this works. So why don't you fix it this location, or you should have done it while I talk. So what you should do, you should just try to fix it this location for a minute. If you can without blinking, just fix it very steadily. So the idea is that what you're staring at are two gratings on top of these now, made these. And on top it's a grating with high frequency fringes, so you have more of these cycles per unit space and in the bottom one you have a low frequency grating. And now as you're looking at that, your neurons that respond either selectively to the grating from the upper part of the visual field or to the grating from the lower part of the visual field they're very active and they're going to adapt, either going to fatigue or they're going to adapt. I mean that's controversial what exactly happens in the brain, but they're going to adapt out. So now what you should do, you should look at this over here and what do you see? Are they the same or are they similar? Are they different? Does anybody see them different except Patrick who knows what he should see? I don't find it a very compelling after effect. I must say I tried this out yesterday. So the idea is that here you adapt to this low frequency, to this high frequency grating, your brain. Here you adapt to the low frequency grating because you're looking at it. So if you adapt to low frequency grating, your brain, so the activity pattern will shift such that you're more likely to see high frequency grating. In other words, this, you'll see a slight perceptual shift and you're more likely to see is a high frequency grating. Here the opposite occurs. This is a high frequency grating. So your neuron that represent the high frequency grating are very active. So that means they're going to be in their fatigue or they adapt so they'll be less active in the immediate future the next few seconds. And so therefore your brain is more likely to interpret this or another frequency as a lower frequency grating. And so although physically they're the same, this you should see as a lower frequency grating and this is a high frequency grating. So if you compare the two with the other two, they should be different. Well, let's try it again for orientation. Okay, so let's try this again. Let's get here for a minute without blinking. I mean, the longer you look, the less you move... I'm sorry, sorry, sorry. Look it over there for a minute. The longer you look, the less you move your eyes, the more powerful the after effect. Color and motion usually work much better. I should have... Yeah, I know which one I should have done. Okay, let me remind myself. Color works very, very powerful. Okay, so now... Patrick, let Melissa look at the grating. So now you should shift and look at here. And what do you see? The waving? That's me. Really? Well, it's a lot of you see this as waving. Incidentally, what do you see now? You don't see an after the image? I certainly saw an after image. Waxes and wings, really interesting. First it's gone, then it comes back, then it's gone, then it comes back, so it does it a couple of times. Okay, so what I saw was this... So on the left side, you see this pattern, and after you get on the right panel, you see this pattern. Did anybody see that? Okay, so that might have to do with eye movement. Of course, amplified. Eye movement, it's going to be amplified. Because for sure, the adaptation is very specific. I mean, specific to location and specific to orientation. If people do it properly. I don't know, maybe the contrast might not be strong enough here. It certainly works in a lab. Okay, so next time I'll bring you... Well, I'll show you two... I mean, one, we talked about in classes motion, that works very powerful. And I'll bring you the most powerful one is color, after fact, where if you do it right, you're convinced there's a color there, and in fact, there's no color there. It's just an after-effect. Okay, but here, once again, the argument is similar. The DSP is looking for a long time, let's say, of things that are oriented minus 5, and in the upper part of the visual field, things that are oriented at plus 5 degrees. And so the neurons that code for that orientation, they adapt out so they respond... After a while, they respond less vigorously. And therefore, if you're looking at a pattern like this, the brain will respond in a slightly shifted pattern, because the neurons that code for this orientation sort of fall now, because they're much weaker. And therefore, you tend to see this in the opposite direction. So it's called a tilt effect. Well, this one here, you see in the opposite direction of this. So whereas on the left, they look like this, on the right, they look like this. I have no idea where the wavy lines come from. I have no idea where the wavy lines come from. Now, this is an older, memorable and honorable tool in psychophysics that was used a lot, particularly before the advent of imaging. You know, if you want to study humans and you don't want to do it in patients and you don't have access to brain imaging, then there aren't too many other tools that you could use. This is like in the till the 1970s, well, really 1980s. There's only a limited number of tools you can use as psychologists. So this is one tool that people use to have specific after-effects. And they can be very specific. So I asked you to fix it here. Why? Because this orientation effect is specific for location. So in principle, this after-effect, you know, you're looking at these patterns on the lower part of the visual field, so you're not going to get this after-effect on the upper part of the visual field. In the upper part of the visual field, you're going to get an after-effect of this orientation. It can be specific to orientation, to color, to motion, and there's even a recent study where they show it's specific even to faces. They use an after-effect technique to create after-images for faces. And so that allows you, what it allows you to do, it allows you to infer something with the underlying neurons without knowing them. Because the argument is whenever there's a specific neuron population that, let's say, fires for these orientation, you can adapt it out, and that's how you can test from the outside for the existence of this population. If you can adapt it out, then you don't know whether there is a neuron population or there's not. Maybe there's one you don't have access to. But if you find a specific adaptation for something, then the argument is that tells you there's a neuron population or that code for that particular attribute that's adapted out. That's the argument which has been used for a long time. And of course, we know that. We know from, I briefly mentioned them in the lecture on primary visual cortex, and neurons in primary visual cortex and beyond. They don't really occur at the level of, they certainly don't occur in the retina and cells are only very weakly orientation-tuned in the lateral geniculate nucleus. But in V1 in higher areas, beyond V1, we know there are lots of neurons that respond very nicely to orientation. They might file for this stimulus, but they don't file to this or to this stimulus. So the specific experiment that I'm talking about was done by Schenghe and Patrick Kavanaugh at Harvard. They did the following. So they had one of two configurations. You fix it here and then in the periphery, like 20 degrees, fix it here, 20 degrees you have this pattern. And you look at this for a minute or so, okay? Firstly, you look at this pattern. And then what you do, I show the same thing. So this is one way how you can measure the, how you measure adaptation in the lab. You fix it here, somewhere in the periphery, you're looking at this for one minute. And then very briefly, you flash a grating, a very thin grating, at the same location where this one was, either at the same orientation or the orthogonal orientation. And you adjust the contrast so people can barely see it. And people always have to say, what's the grating, you know, I mean, you know, left diagonal or right diagonal. And then you adjust the contrast. How strong does the contrast has to be before people can rely on, say, it's a left contrast or it's a right contrast? That's one way how you can measure. So you measure the sensitivity of the system. Before, I mean, before your adaptation, there's no reason to expect and people don't find it. There's no difference in terms of what's the weakest contrast that you can see between this orientation and this orientation. The weakest contrast, whatever it is, it might be a percent out there, might be a few percent out in the periphery. So it's the same whether it's this orientation or this orientation. But afterwards, after adaptation, okay, so, okay, so here they plot the difference in, so this shows you the difference, the ratio between the contrast that you needed, what are some of the faintest grating before adaptation and after adaptation. And if it's the same, if you test it with the grating, it has the same orientation as the adapted one, then your grating now needs to be six times stronger. So it needs to be much more stronger before you can see it compared to an orientation of the orthogonal, a grating of the orthogonal orientation. Okay, so in other words, in order to see this grating after you adapt, it has to be much brighter than this grating compared to before. Why? Well, because the orientation is very specific. The adaptation is really only for orientations around, you know, around this left diagonal, and the adaptation does not extend to gratings that are oriented this way. Okay, so that's the standard way how people have done it since 100 years. Now here what they do, they use what's called lateral masking or crowding to hide the grating. So again, they have people now look at the same, look at five of these apertures, and they still adapt but this one is a critical one. It's the same as before and at this location, after a while they flash again after this has been on for a minute or two to burn it into retina and to get the after effect. Then again, they just briefly flash this and this and once again to determine what is the faintest grating that they can see of this or the other orientation. Now the difference is perceptually if you ask what's the difference between here and here, here you can see that there are five gratings but you don't really see this one any more in detail and you don't have access to its orientation anymore. The idea is that you're looking here and there are five gratings, if it's by itself you can tell with far above chance probability whether this or this orientation but if there are these other gratings above it and below it, if there's this one and these other ones above and below it you get what's called crowding, they all cowder at each other and the orientation is sort of gone, you don't have access to the grain because you get overlap at the level of the receptive field of the neon in cortex they interfere with each other and you don't have access to the orientation so all you know, yes there's a grating there but you cannot see and they test this they ask people to judge the orientation they don't have access to the orientation anymore. However, the level of adaptation is still the same so statistically this ratio cannot be distinguished and from this ratio. So the bottom line is and people have done this now in different forms they've done it for, I think called binocular rivalry and they've done it for very, very high frequency gratings very fine gratings that you cannot see that you don't need to see something in order to adapt to it that's the bottom line you don't need to see something you don't need to actually see the orientation in order for your brain to adapt to the orientation that's interesting because it tells you that the side, you can do what Bela Euler is called psychoanatomy you can use psychophysics you can use psychology you can use these tests in order to infer where in the brain might this process occur and what you can conclude based on these experiments is that the location where the conscious axis of orientation happens wherever is the neural correlate for the conscious perception of orientation has to be beyond the side in the brain where the neural correlate of adaptation is because your brain adapts to it even though it can't see it therefore since from other experiments because primary visual cortex is the first place in the brain where you have oriented cells and based on some other experiments in cat V1 is the first stage where you have orientation adaptation that really tells you that the location where orientation occurs has to occur in a higher place beyond primary visual cortex it's a nice use, I like this experiment because it demonstrates a general principle of psychology in the adaptation of psychology which usually is very limited in its use but here you can infer based on this psychophysical experiment you can infer something with the location in the brain where some process takes place you have said something I said it was psychology Patrick is a psychologist no no, I mean sorry I don't want to happen psychology I mean psychology is absolutely necessary to describe the performance of humans or animals but it's it's of course limited because you can just describe you have this very complex system the human visual system or the human sensory system and you're trying to infer something about it from the outside so you have this black box and the black box is amazingly complex in fact I got my PhD at an institute for biological cybernetics and tubing in Germany which was started in the 50s with a premise that once we understand it was started by Karl Tegai called Werner Reichert and the claim was once we understand once we use behavioral methods in the simple animals like fly simple animals in quote quotation marks like flies particular visual motion perception in the flies then we can even in the simple animals we can infer something about the underlying structures and that will generalize to humans and this program really although the standard model of motion perception is due to him Reichert-Hassenstein correlation model a spatial temple energy model and we believe something similar is also occurring in humans the underlying premise of this program that you can infer from the outside even in simple organisms such as flies something about the underlying structures just proven to be not working out just because of structures even in much simpler animals like C. elegans are so vastly complex and they adapt and they're nonlinear feedback mechanisms that you really have to open the box and you have to perturb the system as much as you can that's my little di-type on psychology and biology on the other hand, biology was out so a lot of people now talk about molecular biology how molecular biology will explain everything molecular biology by itself of course is also very limited if you would do the molecular biology of the visual cortex and you would catalog every single protein expression in the cortex that really wouldn't help you understand a lot about how you actually see in order to understand how you see you have the performance of people like in psychology or the description what they see the behavior of seeing and the visual the phenomenal aspects of seeing and tie that to the underlying neurons you really need all of these so it turns out there's an interesting story you don't dream with primary visual cortex now of course so here I'm making the theoretical assumption that visual consciousness for dreams is similar to visual consciousness for perception and this may or may not be correct it's very difficult to test because with one exception with one really nice exception I don't know in an animal how you would test the content of dreams the neuronal correlates of content of dreams it's obviously very very difficult because you don't have an external control and even humans so we study so I showed you we study these patients in the UCLA where the electrodes implanted and in principle we've had this debate I think Patrick and I had this debate how can you test so we have these electrodes they're 24 hours a day, 24-7 the patient's in the clinic for some days or something so in principle these electors are then there while the patient sleeps and they're being monitored because you want to know whether the patient has a little seizure without actually waking up and the patient certainly reports of dreams there's no question about that, they do report phenomenal dreams and so in principle you can now ask these questions for example these neurons I showed you in the first class that are very selective to things like individuals like I showed you a picture of Clinton or of dolphins etc. that we know are also active during imagery so we know the same neurons are active when you think about the dolphins or when you see a picture of the dolphins would that neuron also be active my gut feeling is yes but how do I know do I wake up every time the neuron spikes do I wake the patients up that's not going to work I'll never get permission to do that plus of course neurons spike all the time for various random events I don't really know how to induce a dolphin dream in a patient so it's not really a readily testable experiment there's been one nice experiment in rats where they had rats run a labyrinth for several hours a day and they recorded the activity with many many electrodes this is done at Matt Wilson's lab at MIT and then they recorded the rat where it was in slow way sleep and in REM sleep and they did see statistically speaking very similar patterns of neural activity as in a wake rat except they were sped up by some factor 5 or 10 or something like that so the first paper I know this just came out I think two years ago in neuron first paper I know where they've done this in an animal where you can compare the neural activity pattern and you can see statistically it seems to be very similar to the pattern that the rat that you saw on the rat when the rat was navigating in a labyrinth and so it makes sense I mean phenomenologically rats dream of when they run around in labyrinths all day that's what they dream of do rats dream of electric sheep ok now in humans so people have done this recently there's a number of papers that have appeared primarily using PET, positon emission tomography where you inject the radioactive isotope it's more sensitive in fmi it is invasive so you can only do a few comparison but it is more sensitive one of the problems is it's getting people to sleep in a scanner I think that's the main problem people haven't done this in mri because you're in this tight machine it's very restricted and the magnet environment for those you've been in is very very loud you have all these noises so it would be very difficult to get people to sleep so here what they did is sleep deprived young volunteers you know undergraduates they sleep deprived them for two days probably just took normal people normal undergraduates in the normal state and put them in the in the PET so there you have you know you have a few you have to have the bolus of radioactive what is it? radioactive water I guess carbon and that's being injected in you and then you sleep in this in the PET and they claim that the sleep architecture for the various transitions from slowly sleep to REM sleep was statistically indistinguishable from their normal behavior when they were in sleep deprived because you could argue it's not normal sleep that they have in PET and yeah so what they saw there's a large science paper a couple of years ago in Braun that interestingly doing REM sleep particularly V1 is suppressed compared to slow way sleep that they couldn't distinguish the REM activity from quite awakened so if you just lay in the scanner in the PET scanner and your eyes were closed but you awake statistically they couldn't distinguish the activity patterns whether in REM and in quite awake with eyes closed and they were certainly suppressed compared to slow way sleep they did find very high regional cerebral blood flow again I mean PET is just like MRI essentially tracking hemodynamic activity you're not looking at neurons directly they found high regional cerebral blood flow in x striped cortex we'll talk about next Wednesday if you look from gyros these are areas of the brain where you have centers that are involved in color in color processing, in motion processing in face processing and middle temporal lobes these are structures that are involved in very high level visual representation and also in memory these are all very active this agrees roughly with the phenomenology of dreams almost everybody dreams most dreams, well let's put it that way all dreams I've seen they're always involved visual component auditory component more rarely do they involve some other sensory and almost never olfactory components but they're all very very very visual and so that would be you get all this activity in the brain so one of the interesting conclusion of this paper was that the dreaming brain is not identical the electrical signature if you recall from the brain is not identical to the wake brain when people discovered REM in rapid eye movement sleep in the 50s this was also called paradoxical sleep because from an EEG point of view the REM sleep looks very similar to a wake you have all this desynchronized activity in the EEG very similar to the awake state and so a lot of people believe for a long time the brain is essentially looks very very similar to the awake one, that's not quite true well some parts are very active other parts like V1 and some parts of frontal lobes again those parts of the frontal lobes that are inactive again agrees with the phenomenology of dreams one big difference between dream and awake behavior is you don't really have insight because we all dreams we have where we have these bizarres continuities I mean I might talk with a long dead relative and suddenly I'm flying and then I'm with my guard dog this all occurs in what is perceptually seems like very little time so although dreams I mean people always ask me whether dreams as a conscious state to me there's no question that they're conscious when I'm dreaming it's as real as life itself when you're in the dream itself it feels like life there's no way to tell the difference after the fact of course if you remember dreams you know that you didn't have insight these things were really bizarre clearly they couldn't have happened in normal life but not while you're in the dream you're quite conscious it's not like normal for example you don't have a transfer from short to long to memory that's why when you wake up you mean if you want to keep a dream diary you really have to write down while it's still sort of sticking around in short to memory before it diffuses away into oblivion interesting also there's been a large scale study by somebody called Son who made a study of dreaming or cessation of dreams which is called the Charcot syndrome sensation of dreams in various neurological patients so what happens if you lack this part or what happens if you have a stroke in that part and one thing he does remark about that people without V1 they still explain visual dreams I wonder about that it's just an aside whether the fact that V1 isn't active during dreaming whether that explains why any of you ever dreamt about reading I mean actually looked at the text and read it not just imagine you've read it in dreams do you have I mean have you actually read or just have you sort of did you imagine yourself reading I mean did you actually see a higher QD text I've never and I've talked to a few people I've never seen I mean the question it's an interesting question I don't know again how to test it what is your visual QD when you dream I can test my QD I can give you tests I can show you gradings that are ever ever finer and at some point you cannot distinguish the grading from a homogeneous line anymore or I can test you for other things called hyper QD so I can test you to ophthalmologist and give you a letter chart and I can ask you which letters do you see right 2020 are you better than 2020 etc so and the question is what is the acuity in dreaming because if it's true that V1 is suppressed in REM then you would predict that there are certain particular that all the neuronal representation that are only in V1 so that's probably going to be very high high fidelity information information about very fine details that's probably only accessible in primary visual cortex that information should not be accessible to the dreaming brain and so the question is how would you go about testing this I mean can you design an acuity can you test for acuity in dreams I don't know how people have done this thing called lucid dreams where people sort of claim you can give yourself you can sort of hypnotize yourself you can even do certain things so I don't know whether it's good enough to do acuity testing dreams it's difficult this is a they don't talk about they the patient doesn't report any significant difference now that's not to say that there wasn't fine differences but certainly I mean this is in a clinical setting so the stock of songs he went around and asked people whether they dream and do they still dream visual do they still have auditory component and he had a few patients with V1 lesion or lesion back in occipital temporal lobe that didn't seem to be a big interference that's a huge interference if you have lesions in the fusiform gyrus then often there are disturbances of visual dreams okay another another technique that people are beginning to use well it goes back I mentioned this it goes back to probably stone age the first time somebody hits somebody else over the head and if you hit them over the back of your head where the cuneus is here in the back then sometimes what you can see you can see flashes of light so we know since antiquity that appropriate mechanical stimulus delivered to the back of the head it's not a good thing but might give rise to visual stimulation now today you can do things more sophisticated you can use what's called micro stimulation so in micro stimulation this was first pioneered well actually it's quite old it's as old as neuroscience some of the first experiments done to localize a motor strip in the 1880s in dogs in Boston was actually done by electrically stimulating the exposed surface of a cortex so essentially what you do you open the cortex you know you take up the big part of it or today you just drill a tiny burhole and you insert a electrode through that and then you inject a little ac current little current pulse I mean hyperpolarizing depolarizing and you can do that they used to do it just on top of the surface when you need relatively large currents now you can also go inside like a hair comb you can go inside and stimulate them with fine patterns of current of course it's still very crude compared to neurons and the currents are still very large and probably excite neurons anywhere within 50 to 100 micrometers that's sort of the estimate now people have done this in a clinical context I think it was first done in Breslau in Germany in the 1920s and then it was particularly pioneered by something called Penfield the very famous neurosurgeon at the Montreal Neurological Institute and he pioneered the exploration of the brain of of course everything he did was in disease patients by default I mean you're not going to get a human to volunteer to have a skull opened and his brain stimulated and he found he has sort of catalog of a few thousand patients mainly but not exclusively epileptic patients and a catalog of which site give rise to stimulation and we'll mention it further on in the lecture he does report he has a few patients where you have where primal visual cortex was exposed and there you can get very simple visual phenomena called phosphines so that's usually flashes of light but sometimes but this is sort of most of these are sort of more exesite cortex in eye areas outside of V1 sometimes they also report in V1 where you can see things like flickering light star star wheels thing, simple geometries when you stimulate of course it's quite famous, these are all the Penfield experiments, when you stimulate in a higher level part of the brain, particularly the middle temporal lobe or the more anterior part of the fusiform gyrus sometimes people report these little vignettes that they remember oh yeah with those grandma and the church I remember looking at the cross or something like that when they have a very vivid very vivid percept they tend to be the same thing over and over again you know you can sort of step through the catalog it's not like you have sort of film canister after film canister of the person's life history in the various parts of the brain typically when you get these things like this memory of grandmother in the church it tends to be the same thing and because it's done in patients in epileptic patients it's actually parcel that what the surgeon did he caused a little micro micro seizure it's one possibility people are doing this now there was a patient at NIH where this was done in a volunteer a few years ago she was no she was born normal sighted you should never do this with people who are born blind because it doesn't work but you take a normal sighted person who lost sight due to some destruction of the retina or macular degeneration or something and then this was I can't remember it's like a 50 year old lady they implanted like 15 electrodes into her primary visual cortex and they left them in there for a few weeks and then they de-planted them, they took them out and there was a recent wide article on a similar case done by a doctor Dr. DuBelle in New York where people, so a number of teams are trying to do that now where they're trying to bypass the effect of retina retinae, particularly in age-related macular degeneration and which is unfortunately quite common and the idea is to directly stimulate primary visual cortex through various techniques usually you have, well in the different techniques so people do see discs when you stimulate them you see a rough topography, in other words if you have three wires and you're stimulated versus this and that you tend to see sort of a systematic shift the one light will be here, the next one will be here and the next one will be over there most of them they're not colored most of them they're flickering lights they don't you can go into the microstructure of this if you stimulate for several seconds usually you only get a light at the onset now this is still very unnatural stimulation, it's a wonderful thing to be imagined that if you ever have the unfortunate to become blind then we can put this prosthetic into your brain and we can bypass your defective retinae but for now, for the foreseeable future it's going to be very, very primitive technology at best it's going to be like 10x10, 20x20 not because so much because we don't have the micromachining abilities because we don't understand the interference there are all sorts of non-linear effects between the electrodes, we don't understand the code still used by the brain of course the big problem is also compatibility, right, every time you stick an electrode into the brain you want to leave it in there many months or many years in humans, you're going to get all sorts of degeneration, you're going to have some gliosis you're going to get all sorts of problems Richard Anderson at Caltech of course is one of the leaders in your prosthetic devices not for primary visual cortex in his case he's not a parietal but a lot of these problems are similar okay, so you can say well isn't this now, I can directly stimulate V1 and so therefore that disproves this idea that the NCC isn't in V1, it's just like the retina if I stimulate your retina, you know if people have done this, in human volunteers if you directly stimulate the retina you're going to see flashes of light so once again, you don't know what probably happens if you stimulate V1 and then you get all sorts of secondary this cascade of neural activity that somewhere triggers an NCC that you can finally see so that really doesn't answer the question the best source of information we have and I mainly talk about a couple of weeks from now is really at the single neuron level and that's very limited of course in humans so you have to do it in monkeys and so there are a number of experiments where people have done this in a monkey where they record it during various forms of visual perception in the monkey and then try to see to what extent do neurons in primary visual cortex follow so one interesting observation is called origin of eye you should try this, you should just take any oh no, you can actually do it I mean how many of you do you know about dominance, dominant eye how many of you know you have a dominant eye any? few which eye is it, the right, the left? right what, how many are right eye and how many left I wonder what they call it with handedness are you guys also left handed by any chance? no? but you are okay for those of you who don't know what I mean so you should test it, just take a distant point let's see a point way over there in the corner and then you know put out your finger and look with both eyes and see what's in the background and close one eye and close the other eye if you're like me, like most I'm right dominant if I look with both eyes or close my left eye my finger doesn't shift but if I look with my close my right eye my finger shifts substantially compared to with binocular viewing no, no, no, no most people have dominant eye either right or left some people have, don't have dominant so there it shifts a little bit when you close one, yeah so you don't have a dominant eye it's perfectly fine yeah, because you know when I see with both eyes or with my sorry, it's the other one then of course it's dominant so most of my input comes from my right eye because if I close my left eye I see things over here but my it doesn't really shift, if I close my right eye it shifts substantially so for example you are not aware of this fact I pointed this out so this is just to get your glimpse of the observation I mentioned already once before that you don't have access to origin of eye information that if you do this experiment carefully so the way you do it one way to do it I put on these two tubes you put the light into your left or your right eye and I ask you without blinking, without moving your head whether that light came into your left eye or your right eye then if done properly there isn't any evidence that you have access to this information that you can tell with more than chance probability yes that light came into my right eye that light came into my left eye and one way to demonstrate that sort of intuitively that many people don't know they have a dominant eye and so most of the much more information comes into the right eye than the left eye but you're not aware of that now that's interesting because the only place in the brain where you really have lots of cells that are monocular selective is prime and visual cortex so the information of course is kept separate from the left and the right eye and it's only really merged at the first location where you have new onset of access to both eyes is prime and visual cortex by the time you get outside prime visual cortex into v2 almost all cells are binocular to various extent which means almost all cells are input from both eyes to a variable extent sometimes more from one sometimes more from the other, sometimes equally in other words if one of these cells these binocular cells fires you don't know did it fire because there was an input from the left eye or the right eye it's only in v1 that you have lots of neurons that are clearly only fired because they see they get input from the right or the left eye so the idea that you don't have access to origin of eyes is perfectly compatible with the idea that you don't have access to v1 the conscious view doesn't have access to v1 because the brain doesn't use origin of eye information it's critically for certain types of stereo computation people have shown that but that information is not made accessible to the conscious view more interesting are experiments done a range of paper by Peter Tia who is going to be here next week from Germany to show I remarked already about the fact in the eye that the eye that if you look at a neuron in the eye it experiences every time you move the eye smoothly or to move discontinuously it has these big blurs remember the movie I showed you of how the world looks from a point of view of retina ganglion cell so you know roughly every 200 milliseconds you get these big blurs now or same thing if I move things incontinuously now perception clearly it makes a huge difference whether I move my eyes smoothly across this image or whether the image moves across the retina in the opposite way physically just from a point of view of the visual information if you take a camera and you take a scene I can either move the camera in one way or I can have the camera stationary move the image in the opposite way and from a point of view of input it's exactly the same the only way you can tell those two conditions apart if you have access to non-visual information for example if you have access to the command signal that your brain gives to the eye or if you have access to feedback from the muscle if my muscle tells me I move my eyes then of course I have that information accessible but if I do the experiment carefully and the PDT has done this in monkeys where I for example have monkeys make a series of eye moves of either smooth or these rapid saccades over a visual scene and then I record those eye movements and I play them back in such a way that visually they are indistinguishable I can now ask the question does a V1 cell tell the difference between motion in the outside world or eye motion movement of the eye and V1 does not in a higher area I mentioned already we'll talk about more next once they call MT a motion area those movements can certainly tell the difference they respond much more vigorously to stimulus motion and actually to eye motion clearly you want to distinguish if something is moving out there or if something is moving my eyes of course the other thing the remarkable thing is that if I constantly move my eyes but the world doesn't seem to move outside I mean if I move my eyes here the world doesn't jump about as it should physically speaking so again V1 in that sense is like the retina that it cannot tell it cannot tell the difference between motion of the eye or movement of the image yet perceptually I can so that's rather nice and the data is very very clear on this I don't know the difference at all MT many cells not all but many cells can so in other words V1 doesn't have access to this information it's a key part of our visual experience that if I move my eyes the world is steady yet V1 cells don't know about that now the nicest evidence against V1 but I'll talk about it in a separate class because it's some very compelling visual phenomena called rivaling flash suppression the nicest data comes from a monkey again from the lab of Nikos Logotitas and his collaborator David Leopold who was here at the beginning of the week that you can have that you can train I mean it's one of these experiments similar to the motion use blindness I showed you where you have those yellow spots and you have a cloud and although you have the same physical input sometimes you just see one thing and sometimes you see the other while the physical input never changes so what they could show very very nicely that you can have an image out there that certainly it's not seeing it's perceptually suppressed just like in motion use blindness you can have the yellow spots but sometimes you don't see them yet you have probably literally 100,000 cells in V1 firing millions of spikes yet none of that furious activities actually gives rise to to phenomenal content none of that activity makes it into phenomenal content so the analogy they didn't do it with motion use blindness but the yellow spots give rise to lots of neural activity in V1 yet this activity doesn't show up when it's suppressed the monkey suppresses it at a time you don't see the yellow spots yet they still give rise to neural activity in V1 now that's somewhat controversial because in humans if you do the same experiment the one I just mentioned by natural rivalry you do brain imaging you get a different result so in humans and we'll talk about more in other words when the person constantly sees the stimulus there's much more hemodynamic activity in V1 than there is when the person doesn't see it but there's some serious methodological issues there because what you're seeing is hemodynamic it's not neural activity and people are realizing that now and there can be in a certain condition a big disconnect between there have been several cases now and this is something one cannot emphasize often enough in several cases in the animal literature both in rat and in monkey when you can have strong hemodynamic signature what people measure is called bold blood oxygenation level dependent contrast changes you can see this fmi signal yet it does not actually correlate with a change in firing rate that you can have these large hemodynamic activities that represent use of metabolic energy the brain something is going on there's neurotransmitter release there's probably action potentials going on but the big relay cells the one that project outside they're not active yet you still get this big hemodynamic activity so it's going to force us to reevaluate a lot of the imaging data I think now it has been argued that while a lot of these results are mentioned are all true these data are mentioned the monkey data is not really controversial it's more the interpretation in terms of human that's controversial because the imaging data that if all of this is true there's still there's some claims that there's that what you need you need feedback so primary visual cortex just like any other area it's heavily connected not only in the forward direction so V1 sends out axons to all sorts of other brains but what's typical of the brain it also receives heavy feedback so if you count synapses probably 10% of the synapses in primary visual cortex 5% 10% it comes from the input in this case the LGN and the retina ultimately and roughly 10% of synapses comes from outside from a high area that feedback that provides feedback signal and that's typical of most areas if you look in the brain there's lots, I mean most synapses are with local neurons so neurons like sometimes older people they tend to talk to themselves mainly and you have a minority of synapses that are made by the input and you have another synapse that is made from feedback and the claim is that this feedback is really critical for conscious perception it might be true right now it's really very difficult to test in the absence of a tool like molecular biology or pharmacology that would allow us to selectively deliberately and delicately and transiently most important has to be done transiently and reversibly to interfere with the system in other words for what we would like as a tool in a human really very even more difficult to probably in a monkey or in a mouse first that would for instance block all the feedback synapses so the feedback synapses structurally probably have slightly different structural proteins or functionally active proteins and the feed-forward synapses I mean they both use the same neurotransmitter glutamide but we know there are lots and lots of subtypes and people have found molecular, for example signatures that distinguish a connection coming from the LGN into cortex to cortical-cortical connections so it's quite plausible that there will also be differences molecular difference between feed-forward cortical passes and feedback pathways and what we need, we need a tool that we can deliver like a viral infection you want to use a virus like you use a little molecular syringe a disposable molecular syringe that essentially you put something you put some some molecule into this virus the virus infects all the neurons in the particular part of the brain and it silenced them transiently it silenced but only it could just go to the synapse and somehow turns off those synapses people are thinking about those things actively but only those synapses that feed back not the feed-forward synapse so now you can take your same animal it grew up normally you show them some stimulus, it does some behavior now you take this animal, you infect it with a virus either by injecting it which works very quickly or by infecting it nationally which might take a few days and now you retest the animal and you ask the question, well are there for example specific abilities that the monkey is now unable to do people are doing that those sorts of things right now we don't have yet these molecular tools to enable us to dissect the synapses but we're beginning to develop the tools to inactivate specific genetically identifiable neuronal populations so I think there's going to be a revolution coming there coming from molecular biology that will have these very very specific tools it's just a summary so just a summary of what I said I mean before I come to that you can ask, let's just evaluate on an epistemological level this claim that we want is not necessary why should that be interesting after all as you'll see there are probably 50, 60 different areas in the brain at least 50 different areas in the visual part of the brain the visual part of cortex so if one is involved in consciousness why is that particular interesting I think it's mainly, well it's interesting for two reasons if it's true it's interesting A because it tells us that not any cortical activity gives rise to conscious perception so I think that everybody agrees on I mean some people thought for example for a long time well things that are in the brainstem or outside cortex proper those are things that are not represented in consciousness or cortical activity it's very high levels, very specific it's all bound to be represented consciously but all those experiments I showed you certainly with the monkey are you strongly against us you can have nuance find a way very vigorously define a way to stimulate that the monkey does not see so clearly cortical activity by itself this shows cortical activity by itself does not give rise to conscious sensation it's more interesting this claim because the way to test it because in principle these sort of tools psychophysics and mapping psychophysics on neurophysiology and particular neurophysiology allow you to evaluate these claims if anyone cortical area or if anyone specific neuronal population actively involved in conscious perception or not and so here I just listed some of these so here we talked about explicit representation it has to be that one of these things underlying conscious perception anywhere in the brain has to be an explicit representation for that particular feature you do have explicit representation you explicitly represent things like location and you explicitly represent things like orientation yet that's not necessary that's just sorry that's not sufficient it's necessary for a new colleague of conscious but not sufficient you have to have what Zecki calls an essential node in other words experimentally if you think that this population is critical analyze consciousness then if you if you lose it you should lose the particular attribute and that's true in a sense that if you lose v1 you don't see anything again if you think you have identified the NCC then there should be this causal relationship that there should be that if you stimulate this particular part of the brain then you should give rise to specific percept that's also true in v1 however the things that are not true is things like there is at the neuronal level very little correlation between the perception of the animal and neuronal activity or put differently in many cases you can show there's a clear dissociation between what the animal sees and what the neuronal activity is in v1 oh I didn't mention of course we blink just like we move our eyes constantly we also blink our eyes not quite as often as we blink maybe a couple of times a minute while we move our eyes a couple of times a second but we certainly blink and there's been one paper where people studied the response to blink and find that the neurons in v1 follow when you blink the visual input is briefly shut off and v1 neurons follow that and clearly unless I pointed it just out to you you constantly blink but you're not aware of that it's not that suddenly the world goes black as it would be if I for the same let's see a blink is like 60-80 milliseconds if I briefly turn off the light for 60-80 milliseconds you would all notice that but you blink where your own visual system shuts down you don't notice that the reason Krik and I postulated this in 1995 or so was that we postulated this before any of this evidence became available and the reason we postulated that was that we'll talk about it later it relates to the function of consciousness consciousness as a property of an evolved system has to have a function otherwise we wouldn't have evolved it and the function has to be if you look at all the tasks that involve consciousness involve things like planning but there's no evidence that you can do a sort of sophisticated planning planning for contingencies what if this might happen or this might happen if I do this or that route what are the various scenarios that seems to be one of the key functions of consciousness whenever you have a specific deficit in consciousness you have a deficit in that particular aspect of planning and so therefore you would expect that the the neurons that call it for consciousness have to directly access the planning stages and the planning stages are in the frontal part of the brain very crudely we'll talk about more later specifically there in the frontal part of the brain and if you look in the monkey again there's no evidence that neurons in primary visual cortex project into that part of the brain they project into other areas of the visual system but there's no evidence that neurons in the back here in primary visual cortex project directly in the front so therefore out of this theoretical perspective we surmise that you're probably not conscious of primary visual cortex okay that's all I had to say the next time it's the next the next Friday I won't be there I'm sorry I'll be at Stanford so there's no class but I think because some of the other classes we have in our space we can make that up so I'll talk about anatomy of visual cortex in general I'll talk about cortical hierarchy and I'll talk about some of the properties of neurons outside primary visual cortex