 Hello and welcome everyone. This is Acton Flab guest stream number 16.2. It is June 11th, 2022. We're here with Mark Solms. This is the second part in our discussion. We had an awesome part one. Today we are going to have a recap on some of the points that were addressed in the first part in 16.1 and then we're going to head into some unexplored territory. So Mark, thanks so much for joining for these sessions and please take it away. Thanks Daniel, glad to be here again. I was just testing out sharing my screen a minute ago and it seems as if I'm incapable of showing you my face and sharing my screen at the same time. So take a good look at me. You're not going to see me for a while. Here comes my screen. Please God, there we are. Great. And here comes my pointer. There we are and here's my title. The emphasis is on physiological and philosophical considerations rather than computational ones. So I'll start with a brief recap. I made a few claims and I didn't get to the end of this list. So I will just remind you what those claims were and then I will pause when I get to the point that we stopped at last time, which was here. That affect is an extended form of homeostasis. I started with the claim which really is the main claim of my presentation. It is that affect feeling is the foundational form of consciousness and that it is intrinsically conscious. I quoted Freud, who more than anyone else introduced the notion into mental science that mental processes are not intrinsically conscious that much of our cognitive processing goes on unconsciously. Even he made this point, the point that I'm making which in his words was that it is surely of the essence of an emotion that we should be aware of it. That is that it should become known to consciousness. Thus the possibility of the attribute of unconsciousness would be completely excluded as emotions, feelings, and affects are concerned. And I emphasize the word feelings because there's a whole lot of reasons why some people claim that there are such things as unconscious emotions and unconscious affects. Just to be absolutely clear what I'm talking about here, I'm talking about feeling and you can't have an unconscious feeling. In other words, you can't have a feeling that you don't feel. So whatever might be meant by unconscious emotions and unconscious affects, they can't apply to feelings. They can't be unconscious feelings. But that might sound like a semantic point, just a matter of words that you can't have a feeling that you don't feel that's an oxymoron. So I then went on to set out various reasons, empirical ones, why the claim that feelings are intrinsically conscious or also has a solid empirical basis. I drew attention to this discovery that was made in 1949 already, that the brain mechanisms for consciousness itself, in other words, what wakes us up in the morning and puts us to sleep at night, the sort of switching on and off of the lights of consciousness that this is the business of the reticular activating system. The reticular activating system is the brain system for arousing consciousness, for activating cortical processes and rendering them conscious. I also mentioned the periaqueductal gray, which plays a crucial part in all of this together with the reticular activating system, but I'll come back to that later. For now, let's just focus on the reticular activating system. To give you a sense of what I mean by how these structures, their intrinsic function is the switching on of the lights of consciousness, I pointed out that you can make a lesion or you can suffer a lesion, a smallest two cubic millimeters in size in the peri-brakeal complex of the reticular activating system. In other words, a lesion the size of a match head and that is enough in human beings reliably to obliterate consciousness entirely. So I was saying, let's look at these structures if we're wanting to understand something of the intrinsic consciousness generating mechanisms of the brain. And I made the point that if the cortex, which is where we normally shine our light in terms of looking for the neural correlates of consciousness, if the cortex were the seat of consciousness, then there's an easily testable prediction, which is that if you have a case in which there is no cortex, then the patient should not be conscious, not so. Therefore, I showed you one such case. This is one representative example of its type that here is a patient born with no cortex. The condition is called hydranincephaly. Here you see in an MRI scan of her brain that where cortex should be, there is just cerebrospinal fluid. And you see that her brainstem is perfectly intact. So on the view that this brainstem area is where the consciousness is generated, that it doesn't require cortex. This is a critical case. On the brainstem view, she should be conscious. On the cortical view, she should not. And here she is and she's conscious, as you can see. She wakes up in the morning. She goes to sleep at night. In this sense, the lights are on. But much more interestingly, she's not merely blankly awake. She, her wakefulness has a content and a quality. And the quality that I'm talking about is affective quality. You see how she responds to her baby brother being placed on her lap. She responds with some form of pleasure. And so there's a content to this mental state and a quality to this mental state that she's displaying. Why that's important is because when we first learned in 1949 that the reticular activating system is prerequisite for all consciousness, that there's no such thing as cortical consciousness without brainstem arousal of the cortex. We had the view that the cortex provides the qualities and the contents of consciousness and the reticular activating system merely provides the quantitative dimension, the sort of level of consciousness. So it's as if this was a power supply like a television set needs to be plugged in at the wall. The reticular activating system is the prerequisite sort of booting up of the system. But the television set itself, the cortex where the contents and qualities of consciousness or processed could still be claimed to be the seat of consciousness and the reticular activating system merely as a power source. This is why it was important to point out that this child is not only conscious in some blank sense of wakefulness without content and quality, but rather that she displays affects. She's emotionally responsive to her baby brother being placed on her lap. Just as this child who also has no cortex is emotionally responsive, here you can see. She's responding with pleasure to a stimulus. And this is quite generally the case for these children. Here's Bjorn Merkur's summary of his observations in many, many, many such children. They express pleasure, they smile, they laugh, they show aversion and fussing by arching their backs and crying. Their faces are animated by these emotional states. They build up play sequences. They smile, they giggle, they laugh. They show great excitement, et cetera, et cetera, et cetera. So I've highlighted all of those words to show that these two cases that I just showed you with no cortex, they are conscious and they are responsive. And in particular, they are emotionally responsive. And on this, I base the claim that these children do have a quality. They do have a consciousness and this consciousness does have a quality and it does have content. Many people are perplexed as to how this could happen since they have no cortex. How can they respond to things like their baby brother's being placed in their laps? And so I just inserted this slide. I don't think I showed it last time. Just to point out that our sensory end organs, here's the example is the eye, same applies to the skin and to hearing and taste as well. The optic nerve projects to the lateral geniculate and from there to the visual cortex, but not only to the visual cortex, it also projects subcortically to the superior caliculi of the midbrain, of the brain stem, which is immediately adjacent to the periaqueductal gray, which I said earlier I was going to mention again. So these children receive information in the brain stem, which is not conscious. It's not cognitively conscious perception. Conscious perception is generated in the cortex, but unconscious sensory information goes into the brain stem where it is responded to consciously by the emotional structures that are the main focus of what I'm talking to you about. Now, of course, many of you and like many of my colleagues will reasonably say, well, how do you know that there's something it's like to be those children? These might just be reflexes or instincts. These might be the equivalent, these kids of philosophical zombies. In other words, they look as if they're conscious, they behave as if they're conscious, but we can't know that they're actually conscious. And so last time I tried to address this objection by drawing attention to what happens in cases who lose great swathes of their cortex and because they haven't lost all of their cortex, they are able to describe to us what it's like to be them. This is one way of getting around the objection that because these kids have no cortex and therefore can't declare their conscious states, we can't be sure that they have conscious states. So I focused first of all on a case who has a massive lesion of the prefrontal cortex. The reason I did that is because this is a favorite part of the cortex. For those who claim the cortex is the seat of consciousness, like the global workspace theorists, they say it all comes together in the prefrontal lobes. All of this information that's processed in the posterior cortices is re-represented or accessed in the global workspace. And this is where the sentient being, the subject of the mind comes about. And I pointed out last time that if that were the case, then again, we have a falsifiable prediction. A patient who has no prefrontal cortex, like this patient of mine, patient W, he has no prefrontal cortex, but he has a sliver of language cortex. He, if his sentient being was contingent upon the integrity of prefrontal cortex, he should not have sentient being. And so I showed you what it's like. I talked to him about what it's like to be him. He claims to be consciously aware of his thoughts. I asked him to imagine something for me, to imagine two dogs and a chicken, to see them in his mind's eye. And then I asked him to count the legs. I thought this would be a reasonable test of whether he's actually got conscious mental imagery. And please note the person I'm talking to refers to himself as I. So he seems to think that he exists as a sentient being. And I'll ask him how many legs they are in total if you have two dogs and one chicken. And to my disappointment, he said eight. And when I questioned his answer, he pointed out that in his mind's eye, the dogs are dead, the chicken. So I thought that was maybe not a great joke, but it certainly suggested that there was somebody at home. And I made the point last time that these patients, generally patients with massive frontal lesions, that this tendency to make pure rile jokes is considered to be a rather common part of the frontal lobe syndrome or the frontal lobe personality. And this is part of a bigger story that I mentioned last time, which is that these patients are generally quite emotionally disinhibited. And why that's important is because remember what we are considering here is the question as to whether the feelings that you saw in those patients, those kids with no cortex, whether those feelings could be generated in the brainstem. And the claim of corticocentric theorists like Joe LeDoux, for example, is that the feelings literally come about when they are registered or re-represented or labeled or even named some people claim that it's only once you are able to re-represent these subcortical survival circuits in declarative consciousness in prefrontal lobes. That this is literally what brings the feeling about. And I think it's quite interesting that these patients who have no prefrontal cortex don't have a dearth of feeling, which is what you would expect if the machinery that brings feelings about is absent, then they shouldn't be able to have feelings. But in fact, what is generally accepted is these patients have an excess of feelings. There's disinhibited emotionality in these patients. I made much the same argument about the other major cortical area that is associated with the sentient self, and that is the insular cortex. Of course, this is associated above all with the work of Bud Craig, but has been very widely accepted, that the feeling self comes about when the state of one's interests and the receptive body is re-represented in insular cortex. And so again, we have a falsifiable prediction. If you take a case who has no insular cortex, like this patient of Demasius, then there should be no sentient self present. And I showed you this interview where Demasius speaks to him about his sense of self, and the patient is perfectly adamant that he exists as a self. He speaks about himself as I, I, you know? And then this interview ends with Demasius saying, you're aware that I'm aware, and the patient B says, I'm aware that you're aware that I'm aware, you know? And this patient, just like my patient W, is not deficient in emotionality. In fact, all the basic emotions are present, including both bodily effects and emotional effects and sensory effects. And not only that, he's also a little too emotional. And this is a little disinhibited in his emotionality. And this is what we quite generally see with patients with insular lesion. So again, it's very hard to sustain the argument that the self actually comes into being in the cortex. Now, of course, those two cases, those examples of their kinds, patients with massive frontal cortical lesions, patients with massive insular lesions, of course, they are just examples, but the point I'm making now is that there's lots of cortex left in those cases. So the argument is, well, you know, maybe the sentient subject is generated in the remainder of their cortex. And that's a bit of a circular argument because remember, the children who have no cortex, we're told, well, how do we know that they're conscious? They can't report their conscious states. I then gave these two outstanding examples looking at the areas of cortex that are most bound up with sentient subjectivity, according to cortical theories of consciousness. And these patients seem to have intact sentient subjectivity, but now I'm told, well, you know, the remainder of their cortex might be what's generating the consciousness. So we can't use only lesion methods. There's no way out of that on pass. And so I then showed you last time evidence of a different kind. I showed you what happens if you stimulate reticular activating system nuclei. There again, there's an easily tested prediction. The prediction is that if these brainstem nuclei, the reticular activating system in periaqueductal gray, as I'm claiming, if these are the structures that generate consciousness and feeling, then stimulation of these structures should stimulate conscious feelings. And I showed you that here's a case who had a steep brain stimulator placed into the substantia nigra, this part of her reticular activating system. And within five seconds, it generated, it produced a profound depression. The patient was actually suicidal. She didn't want to live anymore. This is a patient with no psychiatric history. The electrode was placed in a brainstem for the treatment of Parkinson's disease and stimulated the wrong nucleus. And that's how this came about. 90 seconds after the stimulator was switched off, the depressed feeling disappeared. And the patient generously agreed to allow for further stimulation in the reticular activating system and out of it. And it was only when that particular nucleus was stimulated that she fell into the depression. So this is the kind of evidence. And remember, again, I'm just giving you examples. You can stimulate intense affective states by stimulating reticular activating nuclei and periaqueductal gray. You get the greatest intensity and the greatest variety of affects from stimulating there. And you get nothing of the kind from stimulating cortex. So this is a different line of evidence suggesting that feelings are actually generated in the upper brainstem. And then I showed you another line of evidence that was positron emission tomography of people in intense affective states. Here you see research participants in states of sadness, here of anger, here of happiness, here of fear. And in all instances, the activation is in the brainstem and the circuits arising from it, the subcortical circuits arising from it. That's what we see. The cortex is, by contrast, largely deactivated. So this is a further line of evidence that the, you know, an entirely different line of evidence. Remember, we've got lesion evidence, then we've got deep brain stimulation evidence here. We've got positron emission imaging evidence that the part of the brain that's generating the feelings is the part of the brain that switches on the lights. And this is why I'm claiming that the basic, the foundational form of consciousness, this prerequisite form is affect. Affect is the elemental form of consciousness. A further line of evidence is pharmacological manipulations. If you tinker with the neuromodulators that are sourced in these reticular activating nuclei, like, for example, noradrenaline or serotonin or dopamine, all of which are sourced in the reticular activating system. Serotonin is, of course, regularly pharmacologically manipulated for the treatment of depression. Dopamine for the treatment of psychosis, noradrenaline for the treatment of anxiety. Noradrenaline is sourced in locus eruleus complex, serotonin in the RAFE nuclei. Dopamine, at least the one that's important for psychosis in the ventral tegmental area. All of these are parts of the reticular activating system. If all that this system did was switch on blank wakefulness, it might be of interest to anesthetists. But in fact, it is the main target of the drugs of psychiatrists who are treating emotional disorders by manipulating the chemistries that are the source nuclei for which are in the reticular activating system. So on the basis of all of that, I argued, remember, that affect is the foundational form of consciousness and it's intrinsically conscious. I said that feelings, we don't own, not only on semantic grounds, I'm saying that feeling is the basic form of consciousness. The reason I'm saying that is because the basic consciousness generating tissues of the brain, the reticular activating system, which is prerequisite for the activation of consciousness in cortex, that these structures generate feeling. That feeling, therefore, is prerequisite, is foundational for all forms of consciousness. I then made, this is a sort of a sidebar. The reason I went into this third point is that namely that the claim that affect is not synonymous with inter-receptive inference. In other words, that it's not just an inter-receptive form of perception as opposed to the extra-receptive forms which have been the major focus of consciousness studies over the last few decades. Vision being the main focus. I'm saying that affect is not, and why I say this is because this is increasingly being argued. In fact, it was argued by Bud Craig himself that affect is just inter-receptive perception. The perception of the state of the own body is the equivalent of the perception of the outside world in the more typical form or extra-receptive form of perception that's been equated with consciousness. I made the point that there's good reason to believe that affect isn't just a sixth modality of perception and the evidence that I presented for that was of various kinds. For example, that there are affects which are clearly not inter-receptive. For example, getting a fright or being startled or feeling pain when a pin is stuck into your finger. These are all extra-receptive forms of stimulation and yet they arouse affective responses. So affect clearly is not uniquely inter-receptive. And then I also made the point that there are many inter-receptive perceptual states which are not affective. You can feel your tummy grumbling or you can feel your heart beating or you can feel your lungs expanding. These are not intrinsically affective phenomena. So inter-receptive perception can happen without affect and affect can happen without inter-receptive perception. So I was trying to draw a line under the idea that affect is just another modality of perception. I think that by detaching affect from perception casts some new light on the hard problem. And so that's where I went next. And I reminded you that David Chalmers who coined the hard problem said that when we look at functions, perceptual functions like vision, which was as I said, the model example derived from the or grounded upon the assumption that consciousness is a cortical phenomenon. Cortical vision became the model example of consciousness. This was following Crick's initiative in the mid 1990s. Crick's idea was that if we can identify the neural correlate of consciousness in the case of cortical vision, then we can generalize from that by discerning the mechanism whereby visual information processing gets turned into conscious vision in the cortex. By isolating that mechanism, we will be able to understand the nature and function of consciousness. And Chalmers said that that's not true. If you isolate, identify the mechanism of visual information processing like this map here does, it doesn't tell you anything about why there's something it is like to see. And he used the well-known knowledge argument of Frank Jackson, the story about Mary, the visual neuroscientist who knows everything about all of this. And I slightly simplify the story by saying, well, let's imagine that Mary is blind. Even though she knows everything about the functions of cortical vision, she knows everything about the mechanism whereby visual information is processed in the cortex. Because she's blind, she knows nothing about what it is like to see. And if she were to be gifted suddenly with a normal sight, then she would learn something completely new about vision. She would learn what it is like to see red and blue, what blueness and redness, et cetera, are like. None of which is accounted for by this information processing flow diagram. None of what she knew about the functional mechanisms of cortical vision would have prepared her for what it is like to see. In other words, there's something else about visual information processing other than the sort of things that mechanistic functionalist dissections like this provide us with. It doesn't, this mechanistic account doesn't predict that there should be anything that it is like to see. And this is the grounds upon which people like Chalmers say that a mechanistic, an account of the functional mechanism of cognitive and perceptual processes, it doesn't tell us anything about why there's something it is like to perform these processes. And this is the essence of the hard problem. I said that I thought that this might be because they were looking in the wrong place, that visual perception and all forms of cortical perception. And indeed, not only a perception but learning and cognition more generally that this can readily go on unconsciously that these are not intrinsically conscious processes that it is perfectly possible to see and to recognize faces and to read with comprehension and even to discriminate colors. These are all uniquely cortical processes and they can all go on in the dark as it were. In other words, you do not have to be conscious of what you're perceiving in order to perceive it. And so this costs a lot of different light on Chalmers's point that all of this information processing can go on in the dark. So consciousness isn't accounted for by our normal functionalist mechanistic way of doing cognitive neuroscience. So I was saying, well, that's because they're looking at functions like perception and functions like learning, which are not, these are not intrinsically conscious processes. I drew your attention to this other statement of Chalmers's from his famous paper in which he says in summary, there is no cognitive function such that we can say in advance that explanation of the function will automatically explain experience. And my point was, well, that's because they're talking about cognitive functions. Could Chalmers make the same statement if we were talking about affective functions? I'm saying, no, he could not. I'm saying that we can say in advance that explanation of the function of feeling will automatically explain experience because the function of feeling is to feel. It's intrinsically conscious, is this function? Unlike vision and perception in general and cognition as a whole, none of that is intrinsically conscious, but feeling, affective feeling is intrinsically conscious. You could not understand the mechanism of affective feeling if it didn't account for why it feels like something because that's the whole point of feeling. So this question of Chalmers's why is the performance of these functions accompanied by experience? Why doesn't all this information processing go on in the dark, free of any inner feel? I'm saying that that question is perfectly reasonable when asked of these cognitive functions which are not intrinsically conscious, that kind of information processing can go on in the dark, free of any inner feel, but that is not true of feeling, of affective feeling. And so my claim, and this is where we got to last time, this is now, I'm now starting with new arguments. My claim is that if we can identify the functional mechanism of affective feeling, then we will be able to explain why this sort of information processing doesn't go on in the dark. So I hope that that's clear. That's kind of like my main point. My main point is we've been looking in the wrong place. We've been looking to cortical vision and cortical perceptual and cognitive processes in general in order to isolate the neural correlate of consciousness. And Chalmers has said, well, it doesn't work. You can isolate the mechanism whereby these sorts of information processing goes on. And it doesn't tell you anything about why there is something it is like to see, et cetera. And so my argument is, yes, that's true of those processes, but it's not true of affect. Affect, remember what I've said to you in my summary now today, that this is the most concentrated consciousness generating tissue that there is. The reticular activating system and periaquiductal gray is where the lights are switched on and that you can switch those lights off with as small a lesion as two cubic millimeters in extent. So this is the place where we should obviously have been looking in the first place in order to identify the neural correlate of consciousness and to understand its mechanism. And much more important than that is the fact that the kind of consciousness it generates namely feeling, it is an intrinsically conscious mechanism that feeling, it wouldn't exist if it wasn't felt. And so unlike vision and learning and cognition in general, which does not have to be felt, affect does have to be felt. That's the whole point of affect. So I'm saying that if we could understand the mechanism of affect, in other words, the mechanism whereby feeling comes about, then we might make some progress with this hard problem of consciousness. So remember, that's how far we got last time. And now I go on to, and I'm now going to slow down a little bit because I'm now going to argue, try to identify what the functional mechanism of affect is. But I thought perhaps before I do that, I should pause for a moment in case there are any questions arising from my summary of my argument so far. Carry on. Thank you, Mark. Okay. Thanks. So my, you see, I'm saying affect is an extended form of homeostasis. And by the way, this is not my argument. This is an argument that was first, to my knowledge, first introduced by Yark Panksep in the 1990s, subsequently popularized by Antonio de Marzio. So I'm just summarizing my version of what is in fact now a good 25-year-old argument. Homeostasis is probably the most basic biological mechanism. It could be said that homeostasis is what enables living organisms to resist the second law of thermodynamics. In other words, it is what enables them to be living organisms. They don't just dissipate. They remain in an organized form. And the basis of this is homeostasis that rather than just explore all possible states, we living things have to remain within highly specific states. And these are called the settling point or the set point, the viable ranges of the organism. And this applies across multiple different parameters. Let me use the example of core body temperature. For those of us who think in degrees Celsius, we have to remain between 36 and a half and 37 and a half degrees Celsius. That's where we need to be. If we deviate too far from that very narrow range, then we are at risk of dying. And that doesn't apply only to temperature. It applies to water, to oxygen, to salt, to sugar, to blood pressure, to all sorts of things about our bodies. They have to remain within very narrow ranges that are viable with the preservation of our living state. If we move outside of those ranges, in other words, if we explore all possible states, then we die. So we have to work against that dissipative trend. As I said, that is this entropic trend. We have to resist entropy and remain within our viable states. Very narrow specific ranges of viability across these multiple dimensions of our physiology. And this is why I say that homeostasis is what keeps us alive. Now, a deviation from that settling point then becomes a demand for work. The organism has to do something to return itself to its viable bounds. And that's the basic mechanism of homeostasis. That's how all of those autonomic functions I was talking about earlier, like core body temperature and blood gas balance and so on, that's how they work. If you're moving outside of your range, you have to do something to return yourself into the range. That is what homeostasis is. It's the mechanism that returns us back to our viable range. Now, what I'm saying is that affect is an extended form of homeostasis. And how it extends homeostasis is that when we move out of our viable range, we feel it. We feel an unpleasant quality. That means I am moving out of my viable range. And by contrast, if you're moving back towards your viable range, you have pleasurable feeling. This seems to be the basic function of affect that enables the organism to know how it's doing in terms of its organismic viability. So the organism feels when it's moving outside of its viable range as unpleasant, which means this is bad. And it feels moving back towards its viable range, back towards its ideal settling point as pleasure, which means this is good. This predicts my survival. This predicts my demise. So feelings are rooted in a value system. In other words, that there is something good and there's something bad. And what is good is to survive and as it happens to reproduce. And what is bad is to not do so. This is, of course, the basic value system that underwrites all of life. This is the value system that underwrites natural selection. So feelings are rooted in that value system and what they do is they enable the organism subjectively, the organism itself to know whether it's moving out of its viable bounds. In other words, whether it's doing something bad within that value system or something good within that value system, that's what feelings do. They enable the organism to know how they are doing within that value system. Now, why does this get added to homeostasis? Because not all homeostasis is felt. Most homeostasis is entirely autonomic. I was mentioning earlier, blood pressure, for example, when you move out of your viable range, your heart rate changes and your blood vessels dilate. In order to return you to your viable blood pressure, you don't need to know, in fact, you do not know anything about it. In fact, it's clinically notorious how blood pressure regulation works because you can be way out of your viable bounds and know nothing about it. And I'm just using blood pressure as an example. There are many, many, many ways in which your autonomic nervous system maintains you within your viable bounds without you knowing anything about it. So why do we need to feel it? Well, what feeling adds is when you're in a situation of uncertainty and a situation where you do not have a readily pre-prepared reflex which returns you to your viable bounds. So you don't need to feel how you're doing if you don't have to make any choices. If you have automatic predictions which return you by reflex, like in the case of blood pressure regulation, to your viable bounds, then you, you, the sentient being have no part to play in the process. What, where feeling comes into its own is where the organism finds itself in a state of uncertainty. For example, in a novel situation for which it's innate pre-wired, as it were, reflexes have no preparedness. You find yourself in a state of surprise and now what you're going to do. If there's no reflexive solution available, then feeling enables the organism to make choices. Choices have to be rooted in a value system. There has to be a good choice and a bad choice. Otherwise it's random. So whether you're doing the right thing or the wrong thing is announced to you by how it feels. So this enables voluntary action. I really must emphasize that this enables you to choose. It enables you to decide for yourself what to do. Things are getting worse. So I'm going to change my mind. Things are getting better. I will stick with this policy. This is working. So these are not, these are not hardwired. These are not innate predictions. These are choices made on the fly by the organism here and now. And they enable the organism to survive in states of uncertainty. In other words, in unpredicted situations, in other words, in novel situations. And God knows there are many of those in life. So just to put flesh to those bones, let me give you an illustrative example. It's normally respiratory control is autonomic. You don't need to make any decisions about breathing. It just happens automatically. So you automatically remain within your viable bounds in terms of the ratio of oxygen to carbon dioxide. But that's only when the normal autonomic reflexes manage the situation because you are in predictable circumstances. Now imagine that you are in a carbon dioxide filled room. Now suddenly you move out of your viable bounds and breathing normally, normal reflexive regulation of your respiratory system doesn't work because ordinary breathing in this carbon dioxide filled room is going to rapidly lead to your demise. So now you've got to do something. You've got to get, you've never been in a burning building before. Let alone this particular one. So you have no reflex or instinct, no innate prediction as to what to do. So you feel your way through the situation. And please note at this point, at the point when you find yourself in this unexpected situation, this is where your need for oxygen becomes conscious. So this is a very important point. And otherwise autonomic function now becomes a conscious function. You feel what we call air hunger or suffocation alarm. And you now you move about in this building, deciding which way to go. Remember, you have no prior knowledge of what to do. And it's only on the basis of how it feels. So if, for example, you go upstairs and the oxygen supply diminishes, you feel worse air hunger. If you go downstairs and there there's a greater provision of oxygen, then you feel better, you feel relief from the suffocation alarm. And so the feelings tell you whether what you're doing is working or not. And so your choices are based on feeling. And so you are able to feel your way through this problem and survive. This is not a small advantage. The ability to survive in unpredictable environments is an enormous adaptive advantage. And so that we believe is what the function of feeling is. That's why in this narrow example, why respiratory control suddenly becomes conscious that it dramatically intrudes on consciousness. Suddenly, your need for oxygen and the purpose of this is to enable you to calibrate your choices, to change your mind about your policy, your current policy on the basis of whether it's working or not, which is exactly what feeling announces for you. In the absence of feeling, you would behave randomly and one in a million will do the right thing and that one will survive and reproduce and the rest of you have had it. So this enables us to make choices within our own lifetime. We don't have to let natural selection do it. Within our own lifetimes, we can adapt to unpredicted situations. Of course, once you've done that, you then also can learn from the experience within your own lifetime. And so the next time you find yourself in a burning building, you might, on the basis of learning from experience, you might have somewhat less uncertainty about what to do. So again, let me just make sure that I'm getting across my main point because this is the mechanism of feeling. This is the function of feeling. This is what feeling does. This is why the organism must feel it. Now, remember what Chalmers had said that, you know, why doesn't all this information processing go on in the dark and so on? That was built upon an earlier argument by Tom Nagel who said an organism has conscious mental states if and only if there is something that it's like to be that organism, something it's like for the organism. Then he went on to say, if we acknowledge that a physical theory of mind must account for the subjective character of experience, we must admit that no presently available conception gives us a clue about how this could be done. So this is what I'm trying to do in this talk. I'm trying to give us a clue about how this could be done. This is Nagel's way of formulating the hard problem. Why and how is there something it's like to be an organism? Something it's like for the organism. My point is that that question only makes sense in relation to things like visual information processing, which it does not have to be something it is like to see. There does not have to be something it is like to discriminate red from blue. You can do that unconsciously. The cortex can do that automatically. That kind of information processing can go on in the dark. But would you, would Thomas even have asked, I mean, Nagel even have asked this question if we were talking about feeling. I've just explained to you what the function of feeling is. And it makes it kind of absurd to ask why and how is there something it is like to be an organism? Something it's like for the organism. I hope that you can see what I mean. Why and how there is something it is like to be an organism, something it is like for the organism has everything to do with feeling. Feeling is what it is like to be an organism. How you're doing as an organism, how much longer are you going to remain in existence as an organism? That's what feeling is like and this is why and how it exists. So if we can get to what the physical mechanism, going back to what Nagel was saying here, he said, if we acknowledge that a physical theory of mind must account for this something it is likeness, then he's saying we have to admit that we currently have no clue that I'm saying it's because we've been looking in the wrong place. I think if we look to feeling and we seek a physical theory of feeling, this will account for the subjective character of experience and we will make some progress. This will provide us some clue about how we might go about solving this hard problem. Now my next point, so let me pause at that point. Having said that affect is an extended form of homeostasis, that's my next claim. Let's see if anybody wants to argue the toss or make a comment or ask a question. Yeah, so Stephen, there's no need to raise hand. After each point, we'll take any questions. So go for it, Stephen. And then Dave, if you have anything, thank you. Yeah, thank you. No, I'm really, really enjoying this. I'm interested in this awareness that leads to the homeostatic sort of correction. And if it's where the awareness is with the actual cells or organs struggling in some way to act or enact upon the surprise or be that in the environment that they're encountering at the different scales. So this issue of actually acting or enacting when that action and inaction can functionally correlate up the nested hierarchies, then it's those actions or the inability to act in the way that cells would like that then is becoming the signal of what or where action is failing rather than necessarily a signal of that actually contains the information inherently in it. I wonder what your thoughts would be on that. Yes, I agree. If I'm understanding you correctly, I agree with the emphasis on action. So I'll slightly run ahead of myself if I say this. Obviously a homeostatic deviation can also be construed as a prediction error. In other words, what you're doing has not led to the outcome that was expected. And there are two ways of correcting prediction errors. You can either change your prediction or you can change what you're doing in order to bring about the prediction that you had originally. In other words, when it comes to prediction error, you can either update the prior and have a posterior prediction or you've got to do something differently in order to confirm the prior prediction. So that's where the emphasis is on action. Now, why this is so important in relation to homeostasis is that you can't change your prediction about what your viable bounds are. You know, if you expect to be between 36 and a half and 37 and a half degrees Celsius on the basis of acting in a certain way, in other words, firing an autonomic reflex. And you then find yourself to be at 39 degrees Celsius. You can't say, okay, my posterior prediction is that by doing this, I'll be 39 degrees Celsius. Because if you do that, you're rapidly on the road to death. So the emphasis, so the prediction error has to be corrected on the basis of changing your action policy, doing something different in order to confirm your prior prediction. So the emphasis is very much on action when it comes to these organismic predictions, these phenotypic, the viable states for your phenotype can't be changed. So you can't just change your mind about what you expect will flow from your actions. You have to change your actions in order to bring about the expected or preferred state of your phenotype. Yeah, thanks, can I just add one piece to that? I think that's really helpful is, and also in terms of like multiple scales going down to smaller scales. So for instance, if I actually feel heat in my body and my body might be, my cells are trying to act to find a better state, that can be seen as oppressive and claustrophobic if I then read that as being me getting into an uncomfortable state. It could also be me basking in the sun on a beach, in the same way that if I taste something that's very sour, there's some sort of action at the cell or the organ level reacting to that sourness or that sweetness, which could be like a nice confectionery suite for a child, like one of these gobstoppers, or it could be quite a problem. So I think, I like what you're saying with action. I'm always wondering whether that action piece is, the inaction is the prediction error in terms of how it can go up in terms of these nested hierarchies of physical scale. Yes, so again, I must be careful not to run ahead of myself too much. I mean, inevitably, you must ask whatever question comes into your mind at any point. But I'm aware that as I go down this list of arguments, I'm going to be addressing those points. So I don't want to make all of those arguments in one go. So I will just say that for now, that what you've just asked has first of all, a lot to do with the fact that we have multiple needs, that there's not only one need that they have to be balanced in relation to each other. And also very importantly that we're talking about a predictive hierarchy, that this is what you're speaking about with reference to scale. That what applies at the sensory periphery and what applies at the core of the predictive hierarchy have different implications for the way that effort works. So I'm just saying those very kind of vague and abstract things for now. And I hope that the picture will become clearer as I proceed. Thank you, Mark. Yes, Dave, go for it. Yeah, I don't know how to state this clearly. This is some thinking that I got into from listening to a very recent interview you did with a young man in another part of Cape Town. We're talking about perception. And I think there, and maybe this has already been done. I just haven't gone across the research that there's a larger framework in which these notions of what can be perceived, what can be perceived consciously, what is a sensory modality, what we don't call a sense. It seems that there are on the one hand conditions, we've been calling those homeostatic conditions that you can't do anything about. Your body either adjusts the blood pressure internally or it goes haywire, but you can't be aware of it. So we don't think about this as something that we can consciously perceive and we don't call the mechanisms that are sensitive to blood pressure and core temperature and other things as matters of sensation, matters of perception, matters of consciousness. On the other side, the other side of the mapping, there are single object influences. A mechanism that is sensitive only to blood pressure presents another reason not to call this a sense. An eye spot just tells a certain kinds of worms, there's light around here or there isn't light around here. No directionality, just it's daytime or it isn't, I'm under the mud or I'm not under the mud. Whereas even a little more information, more kinds of information qualify this more as a sensory modality. And in things like hearing and vision and smell, we have this very rich many to many. There's many things you can do about, many things you can do about a burning smell and there are many kinds of odors out there. Is this tying to anything you've been thinking about? Yes. So the first thing that you're touching on there is the fact that homeostasis is only part of the story. In fact, the example I gave earlier of the person who finds themselves in a carbon dioxide filled room in a burning building, they need to maintain their homeostatic blood gas balance but there's nothing that their autonomic nervous system can do about that. They therefore need to turn to action in the outside world. So when I said that the person starts moving about upstairs and down and then feels is this working or is this not, is this good or is this bad, that we call allostasis. It's acting in the outside world in order to return myself to my homeostatic palms. So that's the first thing I just wanted to make clear. I know it's implicit and I know I'm telling you something that you know very well but I just want to make that explicit for our participants that the importance of the outside world for these internal bodily states, the bridge from homeostasis to allostasis is how we conceptualize that. So then it moves to the next point which is that think about those kids that I showed you earlier who have no cortex and all they're capable of feeling is their affect and they don't, they can't consciously think I feel like this about that. In other words, they don't know what the about that is. They just feel things and they can't include within their consciousness of the feeling what the thing in the outside world is that contextualizes that feeling. The context doesn't become conscious only the feeling itself. To be able to extend the feeling onto the context to be able to say in effect I feel like this about that incorporates the context within that terrain of uncertainty where one's navigating the one's palpating the uncertainty in order to make choices not only on the basis of blind feeling but also on the basis of what kind of object brings about this change in my feeling and what kind of object brings about that for all of that to be incorporated within the realm of consciousness I think is a further leap and clearly another enormous adaptive advantage. So the first leap is just to be able to feel the consequences of your actions. The second leap is to be able to picture as it were your those actions within the sphere of consciousness and in this way to bring the context into consciousness and extend the realm of choices enormously. Thank you, Mark. And I'll just chime in with a third question. So you mentioned homeostasis is only part of the story and in the chat Brock asked what part of the story is Hormesis? I don't know that word. So a small stress inducing like some kind of benefit over the integrated time horizon like an exercise stress that leads to improved strength. You mentioned allostasis kind of anticipatory movement towards a set point, but where does for example induced stress mild stress and recovery play into this extended homeostatic or generalized homeostatic framework? So sorry, I didn't understand what you meant. So, you know, I think again when one gives an overview of an argument like this one always oversimplifies one sort of has to simplify. And, you know, when I say that homeostasis that, you know, you always have to confirm your prior prediction, that's not entirely true. There are also ways in which the homeostatic range can be extended. The homeostatic ranges, I mean, like for example you know, what I was saying about oxygen you can, if you're a diver, you know you can learn how to hold your breath and how to manage the stress of being out of your viable blood gas range in a way that naive person like I cannot. So there are mechanisms whereby these things can in very narrow limits, these things can be changed. But I think the emphasis there has to fall on the narrow limits. Ultimately, there is an utter limit to your viable range. And, you know, this is the driving mechanism of the story that I'm talking about here. Thank you. Of course, so much more to say and learn and add but please carry on. Thank you. Well, you can see highlighted on the screen now my next point and it might seem like a small point. It might seem like an obvious point that complex organisms require multiple homeostats. And in fact, I've more or less said this already but I want to make explicit. And here is the same slide that you saw a moment ago where I've just added the point that I'm now making. The importance of this point is that we have multiple needs which must be categorically distinguished from one another. In other words, we can't have a continuous variable called need that variable has to be factorized across a number of different categories. So let me be clear what I'm talking about. Imagine you have, if we quantify on a continuous scale how much deviation there is from where you need to be. Let's say, okay, I've got six out of 10 of thirst and I've got four out of 10 of sleepiness. Then all that means I've got 10 out of 20 of total need. And so all I need to do is sleep. I don't need to drink. If I can generalize from that little example, the point that I'm trying to make is that, no, that's not true. You have to sleep and you have to drink and you have to eat and you have to defecate and you have to breathe. You can't just submit all of these and bring down the total number. If you did that, you would die. So we have to recognize that each one of these needs has to be met in their own right. And that means we need to know, we can't just have a total variable called need. We need to know which need are we talking about. And this is why they have to be treated as categorical variables. In other words, this here we're talking about how much water I'm lacking here. We're talking about how much oxygen I'm lacking here. We're talking about how much sleep I'm lacking, et cetera. So that they can be, so that the category of need that the organism finds itself in can be addressed appropriately. The appropriate category needs to be addressed. And I know what I'm saying sounds absolutely obvious, but why it's important is that it's because of this because of the point I've just been making that hunger feels different from thirst and thirst feels different from suffocation alarm and suffocation alarm feels different from sleepiness and sleepiness feels different from pain and so on because each one of these feelings, the quality of the feeling tells us which need is at issue and categorical variables are distinguished qualitatively. So we're not only talking about valence, goodness or badness and arousal how much or how little. We're talking also about qualitative categories. And I think that's important when we come back to the whole point of what I'm talking about, namely qualia, namely the qualitative stuff of what it is like to do anything, be anything, that those qualitative distinctions, it's of the essence of consciousness that it has qualitative differentiations. It's not just some total quantitative, some total continuous matter of need. It's a matter of different categories of need which therefore are different qualities of need. And I think that's important in terms of understanding what an affect is. An affect is a state of the organism registered by the organism. In other words, it is intrinsically subjective. It has intrinsic value, valence, goodness and badness which has existential consequences for the organism. And in addition to that, it has quality. It is inherent in the nature of affects that they are qualitatively distinctive from each other. And I hope that this very simple functional mechanistic reasoning that I've given you as to why that needs to be the case, why that must be the case helps to make clear why affects take the form that they do. Remember what we're doing here is addressing Nagel's question about why is there something it is like to be an organism and how does it come about that there is something it is like to be an organism for an organism. That's something it is likeness is not just a valence, goodness and badness but also a qualitatively differentiated state, subjective state of the organism. So I think that by reducing the mechanism of affect to these essential features, we begin to see why it takes the form that it does, why it is a necessarily conscious something in the sense that aren't the rudimentary sense that I'm defining consciousness. So that's why this point is important. That's why I wanted to make the point clearly that complex organisms require multiple categorical homeostats. They need to be treated as categorical variables. Therefore, they need to be qualified. Each of these different factors of need, each of these different categories of need have to be qualitatively differentiated from each other. And this is, I think the ground zero of where qualia come from. Now I'll go on to my next point which is that these different categories of need must be prioritized. Again, I think this is in a way an obvious point but I want to just dwell on it for a moment to draw out some of the implications of this. I said earlier that our autonomic homeostats run, well, autonomically most of the time. In other words, there are automatic automatized predictions which we call reflexes that are, when the organism moves out of its viable bounds, then it fires these reflexes which bring it back into its viable bounds. And then I told you that not all of our needs can be dealt with that way. Some of them, as we move outside of the range of what the autonomic reflex can achieve. Like for example, when it comes to thermal regulation you start to perspire. If you get too hot, this is a reflex. You start to breathe more shallowly and rapidly panting. In other words, that's a reflex. These are autonomic responses to overheating. Then they reach a certain limit. They haven't cooled you down adequately. So then you've got to do something. You've got to do something like leave the kitchen. And that is what I've just said applies to one of your multiple homeostats. In other words, one of them now requires allostatic action. You now need to do something in the outside world. And clearly you can't do something in the outside world in relation to all your needs simultaneously. So all the time you're sliding, your hydration, your water in relation to salt content of your body sliding all the time. It doesn't mean that you're thirsty all the time. You're burning up the sugars in your adipose tissues all the time. It doesn't mean you're hungry all the time. So the question is, what gets prioritized? And there must be some prioritization because you can't do everything at once. When it comes to voluntary activity in the outside world, you have to prioritize, there's an action bottleneck. And the point I'm making here is first of all, just simply that, that you do have to prioritize. And secondly, that what is prioritized is what becomes conscious. In other words, as I gave in the example of suffocation alarm, that business carries on unconsciously until it becomes prioritized. At that point, it forcibly intrudes unconsciousness. My priority now is I need oxygen. And that's what you feel. The other needs, this is the other point I'm making, the other needs don't disappear at that point. They continue to be regulated, but they are regulated automatically. So the prioritization of a need brings that need into the realm of palpating uncertainty. In other words, into the realm of feeling your way through the problem. And this I think is an important part of how feeling works. So let me here, remember I said I was going to talk about the periaqueductal gray. So, and I was going to tell you how this plays as important, if not more important role in the basic machinery of consciousness of the upper brain stem than the reticular activating system. The periaqueductal gray, which is just a 14 millimeter long structure, columnar structure around the central canal of the midbrain. All of our homeostatic mechanisms, in other words, all of the multiple homeostats, and we have many, all of the multiple homeostats, all of them send their residual error signal to the periaqueductal gray. The periaqueductal gray is like a final common pathway of all of these homeostatic error signals. And it seems that this is where the prioritization must be going on. The determining which of the current error signals is the most salient then gives rise to, and this is an argument I must again make clear. This is not an argument of, this is not my own novel insight. This was beautifully argued on comparative anatomical grounds by Bjorn Merker in a brilliant paper in behavioral and brain sciences. So he recognized that the periaqueductal gray, together with the superior caliculi that I showed you earlier, and the midbrain locomotive region, they form what he called a midbrain selection triangle. In other words, this is where the effect, the need that is to be prioritized is selected, and this gives rise to a feeling. The feeling being, as in the example that I gave with suffocation alarm, the need that is prioritized, the need that is currently most salient, the need that is going to now color your consciousness, qualify your consciousness in terms of its most rudimentary property, in other words, the feeling, the effect. What state, what organismic state am I in? I'm in a state of respiratory distress. I'm in a state of suffocation alarm. That's what I feel, why? It's because there, this is where choices need to be made. This is where the creature needs to feel, the person in that example needs to feel their way through the problem. The other needs remain, but the other needs, the non-prioritized needs are not raised to the level of feeling, and I hope it's clear why. And we need to understand mechanistically, how does all of this work? And that leads to my next point. So before I go to that next point, let me just pause again in case there are any comments or questions about this prioritization function performed by the periaqueductal gray, which is what determines what affective state you're going to be in from one moment to the next. In other words, which need is going to qualify your affective state? Thank you, Mark. Dave, Ben, Stephen. Yeah, I just want to throw out their, in Sylvan Tompkins' later work, that is after 1970, he came to elevate, reset to the status of a fundamental emotion. And I think it's very relevant to this, the notion of suddenly shifting from one conscious emotion or one conscious activity or world attitude to another. I believe Iyak Pongsep drew pretty heavily on Tompkins and the reason that he doesn't cite him in a lot of detail is that he accepted so much of it. Kind of like what you said about, I agree almost totally with, oh, the Britisher, whose father's from India, I'm sorry, I'm blind from only spending, he had a dialogue with him just a few months ago. Am I completely wrong about that? And Neil Sith, are you talking about? Yes, thank you. Am I completely wrong? Was Tompkins simply agree with that? No, no, I agree with that completely. Tompkins, the reason that Pongsep doesn't cite him, or barely ever, he cites him usually in the beginning of a kind of a general discussion, just to say this is the tradition I belong to and then carries on building on his shoulders. I agree with that very much. Thank you, Stephen. Yeah, I liked when you talked about palpating uncertainty and the salience around that. So that's a nice term. And I think that relates to what is going on, so to speak. So what about what's going on? And then now what happens next? And I'm wondering though as well, if the idea of where, which I think is often left out, there's a lot of categorization of what things are, but where is it happening? That makes a difference in terms of how significant it might be in terms of prioritization. So I was wondering how much the risk of awareness of some sort of disequilibrium happens could be featured in this. And maybe that's where the physicalization and the spatialization becomes much more significant than when you're just dealing with the brain. Yes, so well, all of these homeostatic mechanisms, of course, the physiological processes that they regulate are widely distributed. But the control centers of these homeostats are for the most part in the brain stem broadly defined. And I say broadly defined because there are some that are in the medulla, oblongata and pons, but there's some that are in the midbrain. There are some that are diencephalic. For example, I mean, just mentioned the most outstanding example, the hypothalamus is full of homeostatic control centers. But, so hypothalamus, circumventricular organs, the parabrakeal complex, the area postrema, the nucleus solitaris, these are all homeostatic control centers. But of fundamental importance, and I think it's not sufficiently recognized, although Merck has certainly made much of it, that all of these nuclei that I have just enumerated, they in turn project to the periacuductal gray. The periacuductal gray is, in my view, a meta-homostat, it's sort of the control center of the control centers. And that the essential function there is a prioritization function. What's, and by being able to physically locate these mechanisms, bearing in mind where I started, I started my answer to your question by pointing out that ultimately these homeostats regulate physiological processes which cannot be localized. They are distributed processes, paroxalons. But the center, the control center can be localized. And those control centers, they are numerous ones, and they in turn, they all send their residual aerosignal to periacuductal gray. That is really important. Now, what that enables us then to do is to test models like the one that I'm describing to you. What happens when one or another of these individual homeostatic control centers is lesioned? What happens when the periacuductal gray as a whole is lesioned? What happens, I might as well just insert that here, is you get a persistent vegetative state. In other words, you get that condition that Magun and Maruzzi led us to believe is a theoretical possibility. In other words, blank wakefulness. Remember, I said earlier in my summary of what I said last time I spoke that the idea that the reticular activating system provides merely a quantitative level of consciousness and not any quality or content. That's fictional state of affairs of blank wakefulness. The thing that the reticular activating system was supposed to be producing, which as I've showed you is not the case because all of those different lines of evidence that are summarized show that what the reticular activating system is producing is anything but without quality and without content. That this is in fact the foundational form of the qualities and contents of consciousness, namely the different affects. If you lesion the periacuductal gray, then you get that fictional state, that artificial state of blank wakefulness. In other words, these patients show non-responsive wakefulness. They still have the autonomic sleep waking cycle. In other words, they wake up in the morning and go to sleep at night like those hydran and cephalic kids that I showed you. But unlike those kids, they show no emotional, no affective, no response to their situation and no intentionality either. And so that's what you would expect would occur if the periacuductal gray is where the affects are actually being generated. Because remember, that's what I'm saying. I'm saying that the prioritization of a homeostatic error signal is the feeling of that signal. It is the rendering conscious of that signal. And that is what we mean by feeling something. It means now this need is not just a need. Now this need is a drive. Now this need is driving my voluntary behavior. And that is why it becomes conscious because as I said to you earlier, voluntary behavior can be defined as the making of choices as opposed to the executing of automatized predictions. Just to add two notes on that, when hearing Stevens questions about spatial localization and how that can refer to brain localization as well as like in the peripersonal space, it made me think about how awareness, a, awareness, it could be without awareness. Awareness could be spatially dislocated or awareness might be a highly spatialized percept. And then the other note that I wanted to make was as a researcher of distributed physiology in the USocial Insect Colony, this extended homeostatic perspective leads to many interesting connections when we think about extended social homeostasis as well. But just two notes and we can carry on. Very good points. So you're saying we can carry on meaning I can carry on? Whichever way you'd like to. Okay, great. So I'm now going to move to my next claim, which is, I see we have half an hour left, so I mustn't, it's really quite amazing because you are so generous with giving your speakers two hours, you lull us into a full sense of security and thinking, well, I can just elaborate as much as I want to, there's plenty of time. And now we're down to our last half hour and I've still got three points to go. So I'm going to- It's like no one goes there anymore, it's too busy. Yeah. Well, again, grateful for the time that you're giving me. So let me move to this point. And this has everything to do with the relationship between affective consciousness and cognitive consciousness. I spoke earlier about, I feel like this about that, in other words, the incorporating of the cognitive domain within the sphere of feeling. And so this is the point that I'm making now. This is a drawing, a diagram that comes from a paper that I wrote with Carl Friston. And here, the important equation is, is the third one. And here, this diagram basically is just trying to spell out in a visual form what these equations are saying. So when I said earlier, that we have two ways of reducing prediction error, when a prior prediction does not lead to the sensory state that's expected, then we can either change our perception. In other words, we can change the prediction or we can change our action. We can do something differently in order to bring about the prior prediction. So that's what these two equations describe, that we have a generative model which generates predictions as to what sensory state will flow from our actions. And so here, there's an action in the external world which is predicted to bring about a certain sensory state. And to the extent that it does not bring about that sensory state, in other words, the difference here is the prediction error. And of course, the prediction error is then used to update the generative model in order to give rise to better predictions in order to better maintain your expected sensory states. I need to emphasize here because this was not the case in the early days of the predictive processing paradigm that Carl Friston and his colleagues unleashed upon the world. Sensory states do not necessarily mean extraceptive sensory states. So in the examples that I've been giving in this talk so far, this would be the equivalent of, for example, core body temperature or blood oxygen level, et cetera. These, the most important sensory states for the organism are its viable states in terms of its homeostatic expectations. So please remember that when I speak about actions in the world, the world that we're talking about here is the, can be the visceral body and it can be the external world. The mark of blanket, the body is as much external to the blanket as the external world is external to the blanket of the nervous system. I think we've got some, we're in somebody's kitchen here, which you must put your mic, whoever is doing their dishes there. So what this equation here, which is written in words here, the rate of change of precision, which is omega in the equation. The rate of change of precision over time depends on how much free energy changes when you change precision. This means that precision will look as if it's trying to minimize free energy. The rate of this free energy minimization process is the difference between the inverse precision and the sum of squared prediction errors. So that's this equation here put into words. And it's foregrounding the central role that precision plays in minimizing prediction error. So I want to be clear that this is precision modulation is in physiological terms, the modulation of post-synaptic gain. So the message passing going on here between error signals and prediction signals is modulated by posts. It's the reticular activating system's modulation of that message passing. So it's the increasing or decreasing of the gain on the error signals. That's the role of precision modulation, which is just the same thing as to say that is the role of the reticular activating system. It is modulating the gain in the message passing to speak physiologically, to speak computationally. It is a matter of increasing or decreasing precision values attached to the predictions over the errors. So to put it into different words, if things are turning out as expected, that's good. If uncertainty prevails, that's bad. So increasing confidence in a prediction is good. In other words, increasing precision in the prediction is good. Increasing confidence in an error signal is bad. In other words, the more uncertainty, the more you become clear that things are not turning out as expected, to that extent, of course, your confidence in your current policy is reduced. In other words, the precision in your current policy is reduced, and that just is bad. So the goodness and badness, the pleasure and unpleasure function of precision modulation that I described to you earlier is it has this enormously important contribution to make to the whole of this mechanism by determining the influence of the error signal over the predictive model. So to the extent that precision is reduced in the error signal and thereby confidence is maintained in the policy, to that extent, the error signal will or will not have influence over the parameters of the predictive model. And so this is trying to illustrate the crucial role that APEC plays in this whole predictive mechanism. So this is, as it were, the role of the reticular activating system of the modulation of the message passing that goes on. These are, as it were, the synaptic transmission mechanisms, and these are post-synaptic modulatory mechanisms. And so I'm just wanting to link these formalisms to the affective role that the reticular activating system plays. So just to go back to the statement here, the mechanism of perceptual and cognitive consciousness is precision modulation of allostatic prediction errors. In other words, it is the modulation of confidence in a current policy. In other words, the perceived consequences of a current... So as a particular need is prioritized, that need is felt. This generates a category of predictive policies. In other words, this is what I do in this situation in this context that I find myself in. This is what I expect the consequences will be. In other words, these are my expected... And there's an expected precision attaching to the error signals, of course. All the other domains of need. Remember what I was saying earlier here about the needs must be prioritized. The other domains have monotonous precisions. So the crucial mechanism in terms of voluntary action to use the phrase that one of you said you liked earlier is the palpating of the uncertainty. In other words, the palpating of the precision and the adjustment of the confidence in the current perceptual and cognitive, in other words, the allostatic aspect of what I must do about this need state. Because there's a state of uncertainty that's been prioritized there, there's a changing of one's mind on the fly. This is what this mechanism makes possible. So the effects are, as it were, a drive for or a demand for predictive work, for mental work. And this, the modulation, the palpating of the confidence in the policy over the error signals is the predictive work so demanded. So in other words, the work of updating one's policy, of changing one's mind, it is all of it underwritten by the effect of demand. So the effect of demand for predictive work is what gives rise to the predictive work itself, which is the palpating of uncertainties in the current policy and the sensory states that it gives rise to. And so this is changing your mind. In other words, voluntary action, in other words, the capacity for choice, this is the crucial role that precision modulation plays in that process. So that's, oh, okay, good. So I think, let me just pause for a moment to see if there are any questions about that because I wanted to illustrate it with a case, which I think is going to pretty much take us to the end of our time. Before I launch into this last case, let's see if anybody has any questions or comments about that mechanism. I think we'll go to the case. And then there's always so much to unpack with defining the variables and understanding what the edges mean and so on. But I think the case is a good way to close this. Okay. So this is a patient of mine, Mr. S. And he had a meningioma here at the base of his frontal lobes. It was an olfactory sheath meningioma. And it pushed on his optic nerves and as a result of that, his vision was impaired and this is how the tumor came to attention and it was successfully surgically resected. Now, because of its location in relation, in fact, to the optic nerves, there was some nervousness on the part of the surgeon to remove it in its entirety. In fact, he felt it was not possible to remove it in its entirety. So he left a little nub of the tumor and this re-grew and so the patient again noticed these visual difficulties, returned to the surgeon and the operation had to be repeated. Because of scar tissue, the second operation is always trickier than the first and unfortunately in the second operation there was a bleed and that bleed was into the basal forebrain nuclei. Small bleed, but the basal forebrain nuclei are crucial. These are the upper end of the, remember I was saying earlier that these different parts of the ascending arousal mechanisms of the brain that they are the source nuclei for these different neuromodulators and I spoke of dopamine and noradrenaline and serotonin earlier. Well, these nuclei are the source nuclei or there are other source nuclei for acetylcholine here too, but these are very important source nuclei for acetylcholine. Remember all of these, what I was saying earlier about these neuromodulatory systems is that they are modulating postsynaptic gain which is just the same thing physiologically as to say computationally they're modulating precision. That's what they do. They up and down regulate the precision in the message passing and this as I said earlier, it dictates which messages are going to be selected and which not. So it plays a crucially important role does precision modulation in cognition and I'm now wanting to show you this relationship, the relationship between the affective mechanisms that I'm talking about the affective functions that are performed centrally by periaquiductal gray and then how this gives rise to the modulation through these arousal systems of these different neurotransmitter systems and I'm going to now show you how this worked in this case. Acetylcholine modulates the confidence in error signals and so let me just quickly tell you a little bit more about this patient. He was 56 years old at the time of the second operation and as a result of the damage to the basal forebrain nuclei from the second operation, he woke up from the surgery with a condition called confabulatory amnesia. So although the tumor was resected successfully and the visual problem was corrected, he now had this devastating new condition called confabulatory amnesia. Confabulatory amnesia, the patient is, it's not only that they are amnesic, in other words, that they're unable to remember, particularly recent events, but there's also quite a long retrograde extension too. In other words, there's quite an impairment of their retrograde memory too. Not only do they have this, but they also are not aware of their memory problems. And so when they try to retrieve a memory, the memory that they retrieve is not the correct one and they don't realize it's not the correct one and so they have what appear to be false memories. These are the confabulations. They frequently are related in some semantic sense with the target memory, with the memory that they're looking for, there's some semantic relationship between the memory they find and the one they're looking for, but they can be grossly misplaced in space and time. And so these are what we call confabulations and the patient is not sufficiently critical of these misrememberings as a result of the damage to these neuromodulatory mechanisms. So this patient, Mr. S, just to give you one example, I mean an extreme example, he had his operation in Johannesburg, which is in South Africa, which is where I hail from, but at the time I was living and working in London and so the surgeon who I knew well sent the patient to see me to consult me in London because I was doing work with this condition, confabulatory amnesia. He arrives in London on a Friday and comes to me on the Monday and has no idea that he's in London because of course he doesn't remember the journey, he can't remember anything from one minute to the next. And so I say something about, well, you know, the surgeon, whose name was Mr. Miller, the surgeon referred you to me because of this memory difficulty that you're having, that's why he sent you to London and he said London, what do you mean London? And I said, yes, you're in London, you don't realise it because you don't remember the journey, you know, that's the whole point, this is the kind of problem that you're having and he denies it, he says he's not in London. Now, so that's the confabulatory aspect that these patients, in his memory, he's in Johannesburg and so he believes he is in Johannesburg. So I point out to him, it was winter and it was snowing outside, I point out to him the wintery conditions outside, you never by the way have snow in Johannesburg. So I say, look out the window, he looks at, he's absolutely shocked, but then retorts, no, I know I'm in Johannesburg, just because you're eating pizza doesn't mean you're in Italy, that's what he says to me. So in other words, you don't have to, you mustn't overrate the evidence of your senses. And so that's just an extreme example of what I mean by how these patients, how they're amnesia, it's not just a lack of memory, it's also an excessive confidence in the incorrect memories that they draw up. So that's the background, and I saw this man, the whole point of him being referred to me is because I treat such patients and so I then saw him six days a week at the same time in my art patient clinic at the Royal London Hospital, same time, same place, he came with his wife to the waiting room, I would then go and collect him, take him up to my consulting room, spend an hour with him, sit down to the waiting room, and then I would talk to his wife because he was so full of confabulations, I needed to verify things in order to get some sense of what was going on. And there were certain themes, and this is an interesting thing about these patients, these confabulatory patients, there's certain themes that returned again and again. The patient was an electronic engineer, but in reality he was, but he thought the reason he was coming to me, frequently he thought he was coming to me, because I was consulting him about an electronic problem, or otherwise he thought that the two of us were electronic engineers together working on some electronic problem. Also frequently he thought that he and I were in some sporting team together, that we played rugby together, or we were in this rowing team together. Now his wife told me he had played rugby at university a good 30 years or more before, and likewise he had been a keen rower, but this too was at university more than 30 years before. And so these are again good examples of confabulations, the patient mislocating in space and time a memory that he draws up now and he has too much confidence in that, he too readily accepts the veracity of the products of his own memory search. So that's the background. Now on this particular day that I want to give you a little snippet of a session from, he on this day when I came to the waiting room, he touched the scar on his head, the craniotomy scar on his head, and he said hi doc to me. So this was progress. For the first time he was associating me with medicine and he was associating me with the surgical scar on his head, and so I thought this was great progress. And so when we got into my consulting room I said to him, you touched your head when we met in the waiting room. And he said, I think the problem is that a cartridge is missing. We just need the specs. What was it? A C49? Should we order it? So I said to him, what does a C49 cartridge do? He says memory. It's a memory cartridge. A memory implant. Oh, sorry. I should have mentioned this. His wife told me that he had had dental implants. He'd had serious problems with his teeth. And these had been for many years, but these problems had finally been successfully treated by implants, teeth implanted into the jaws. So when he said that C49 cartridge is a memory cartridge and that it's a memory implant, it brought that operation, that dental operation to my mind. He said, but I never really understood it. In fact, I haven't used it for a good five or six months now. It seems we don't really need it. It was all chopped away by a doctor. What's his name? Dr. Psalms, I think, but it seems I don't really need it. The implants work fine. So I said to him, you're aware that something's wrong with your memory, and he interrupts me. And he says, yeah, it's not working 100%, but we don't really need it. It was just missing a few beats. The analysis showed there was some C or C09 missing. Denise brought me here to see a doctor. What's his name again? Dr. Psalms or something. And he did one of those heart transplant things. So this is referring to another operation that he'd had, which is clearly also being referred to here, which was that he had a cardiac arrhythmia and so he had a pacemaker fitted, a cardiac pacemaker. So he says, so he did one of those heart transplant things and now it's working fine, never misses a beat. So I said to him, which is what I actually thought, I said, you're aware that something's amiss. Some memories are missing, and of course that's worrying. You hope I can fix it just like those other doctors fixed the problems with your teeth and your heart. But you wanted so much that you're having difficulty accepting that it's not fixed already. So he goes on and says, oh, I see, yes, it's not working 100%, and it touches his head. He says, I got knocked on the head, went off the field for a few minutes, but it's fine now. I suppose I shouldn't go back on, but you know me. I don't like going down. So I asked Tim Noakes. Tim Noakes, by the way, is a sports physician in South Africa. He says, so I asked Tim Noakes because I've got the insurance. So why not use it? Why not go to the best? And he said, fine, play on. So that's the end of the snippet that I wanted to show you. By the way, before I show you about that. So this is a case in which due to damage to one of the precision modulating nuclei or sets of nuclei, in this man's case, the basal forebrain nuclei, due to the damage to the acetylcholine modulating mechanisms that I showed you earlier. He has too much confidence in his predictions and too little confidence in the, he does not up the game on the areas. Say, looks, I say to him, his predictive model tells him he's in Johannesburg. I say to him, no, you're in London. Look out the window. He looks out the window. He sees snow, you know, clearly not something that could possibly be associated with Johannesburg, but he sticks with his prediction. He says, no, I know I'm in Joberg just because you're eating pizza. That doesn't mean you're in Italy. And likewise, in the case of these memory processes that I showed you in that session, each time that he starts to feel the unpleasure of the mounting error signals, the mounting, the things are, so just think about it. He has, he touches his head. He's on the brink of being aware that he's had a brain operation, that the brain operation has resulted in loss of memory. I hope you can see that with him saying that he says, we just need to order the specs for this module that's missing. What does the module do? It does memory, you know, so touching his head, speaking about operations. So he's on the, he's on the brink of recognizing that things were not as he expected, that things are in fact quite different. My point being that this evokes feelings, that this is bad, this is a panic inducing situation. And because this man has damaged to these precision modulation mechanisms that we've been talking about, that he, what he does is he simply up regulates the, or maintains his confidence in his prediction. And in this way maintains his emotional equanimity rather than allowing the unpleasant affect to dominate and the unpleasant affect to update, to up regulate the error signal, and which would normally update the predictive model. So I'm hoping that in this case, you can see something of the role of affect, the role, the feelings involved in this man's case in relation to his predictions and the prediction errors. In other words, the cognitive business of what he perceives and what he believes and how what he perceives changes what he believes and the role of feeling in all of this. I thought this was a succinct case to be able to illustrate, illustrate all of that. The, because it's only one case, I thought I should just show you some of this. These are several papers that we've written on showing how these mechanisms work in confabulation. We, it's not just this one case. We were able to show that affect regulation, that the confab, that confabulation has a wishful quality. In other words, it has a down regulating of error signals quality and an up regulating of predictions quality. And we were, we were also able to show that by analysis of transcripts of cases like this, we studied many cases like this, how the affect actually improves with each confabulation. So there's an increasing negative affect followed by confabulation, followed by an improvement in the affect. So I see I'm right at the end of my time. I just want to, I don't have time to go into this point. I'm sorry. This is, this is not such a fundamental point. It's just a slightly different way that I see the predictive hierarchy. If we, if we take seriously that what we're talking about here, fundamentally, we're talking about homeostatic systems and that the most important predictions have to do with maintaining your, your phenotypic prior preference distribution that this has implications for how we conceptualize the predictive hierarchy. But as I say, we don't have time to go into that. And then the last thing is just to say that these models that we, that I've, I've together with my colleagues derived from the study of, of the neuroscientific evidence on the basis of this well-known statement of, of Richard Feynman's what I can't create. I do not understand. We, we are trying to instantiate these mechanisms that are described to you earlier in an artificial consciousness. I'm working with a group of really great guys, physicists and computer scientists and applied mathematicians. And we presented our preliminary findings a few months ago at Colfriston's theoretical neurology meetings. And so watch the space. We are, we are, we are saying that if we have identified the mechanism, the causal mechanism where, whereby effects are generated, what they actually do in relation to a self-organizing system, then if this really is how effect is, is generated, if this really is how it is caused, then we should be able to instantiate it in a, in an artificial system. And that's what we're, what we're trying to do at the moment. I'm sorry to have rushed at the end, but if you'll indulge me, I will wait for, if there are any questions, comments. I just wanted to show you my, so this is, I'm saying these questions can be, that they look quite different if we look at it through the lens that I've tried to encourage you to look at it with me through, in other words, through the lens of affect. These questions, why and how is there something it's like to be an organism? Something it's like for the organism and questions like, why is the performance of these functions accompanied by experience? This is Chalmers' question, why doesn't all this information processing go on in the dark, free of any inner feel? I'm saying that these questions, I've tried to provide you a clue as to how we might be able to address them differently. Ooh, what happened there? I've lost my, oh my, good, okay. Ah, there I am, yes. How we might be able to address them differently by looking at them through this lens. Sorry to have gone on, my word, I can't believe I've spoken for two hours again. But if you can allow me, I'll be happy to wait for any further comments and questions before we end. As we have actually a whole document of questions and many questions in the chat, I think it's a perfect place to end the live session and whenever the affordance presents itself, we're always happy to have a .3. The numbers keep counting and we'll keep hosting the sessions because this was very fascinating and I hope people got a lot out of it and very tantalizing. Thank you so much for inviting me and apologies for going on for so long. I really enjoyed what little interchange we had over these four hours I really enjoyed. Thank you so much. Maybe one nice future event could be picking up on those last two, which are extremely interesting and important and then starting with those two, with the .1 and the .2 as prerequisite and then sharing more about those two when the time is right and having that as more of a full discussion. Very good. Thank you, Mark. Thank you, Stephen. Thank you, Dave. Thanks, Stephen. Thanks, Dave. Thanks, Daniel. Peace out. Thank you. Take care. Bye. Bye. All right. Thanks everybody for watching. Hope you enjoyed that. See you later.