Paul Cisek and Asia Conversation

This is a raw transcript of a conversation between Asia and Paul Cisek that happened on January 23. The recording was started a few minutes after the beginning of the conversation, so the beginning is a little bit abrupt. This transcript was not verified for transcription accuracy. The speakers may not be assigned correctly, and there may be some voice recognition mistakes.

Transcript

Asia Shcherbakova 00:00 Thank you. So in terms of what other things people… So our curriculum essentially covers the entire spectrum of typical movements you may find in exercise but we tend to make them simpler and simpler and simpler. And also in my poster when I talk about the history of the method. What’s interesting there is that it was not developed top-down, like somebody thought, what kind of movement should we teach, but they were just modified, modified, modified so that more people could perform them. And also what’s very important, which is not in the poster, originally the idea was that the method should be consistent so that all the teachers understand all the specification about every movement, all the progression for any potential body type, situation, injury. And the students could come at any moment and practice the same movements with whatever teacher they see. And that’s actually not a trivial task because if you start thinking what does it mean that the movements are the same if people have different capacities like flexibility, strength, skill level and so on. And because of that the logic of movement in this system it somehow emerged from this desire to make the movement understandable. And again, what does it mean to understand the movement? Right? It’s kind of— [INAUDIBLE] Yes. Yes. So very often, movement is— I do have a slide where I try to explain this. A lot of the time, you just have to come up with new ways to explain this, because there is no good way to explain this. So if we talk about the communicability, why the breakdown happens is that, well, in most cases, like in most typical situations in non-pro movement training environments, there was just no requirement for high communicability, right? But the next possible reason under specification, when I just didn’t give you enough information, right? And you just didn’t, you were not told not to do something. But then this is where the perceptual skill topic comes in, the capacity limitation, which we think is a primary bottleneck, that students cannot access sensory information that they need to execute the movement. Like I showed you with this final example. So I do have a few pictures. For example, this movement seems very simple. The task here is to keep moving the blue dot back and keep the spine straight. This is an actual quote from a person I found. It’s very difficult and confusing. It is really hard to tell what my body is doing. Also, what’s very important is in how we move is that we essentially use forward kinematics. So from this perspective, like very often when I see how when there’s a discussion how movement planning happens, for example, there’s the target position, the initial position, the differences computed, right? Here there is no target position. Essentially the next target position is just a little bit next version of this movement. And there are, we tell people, for example, let’s say, we don’t want, I don’t know, their thighs to burn. So they need to stop at certain point when it becomes too difficult, but they don’t know when it’s gonna happen. And not just in this movement, but in any movement for kinematics and plus there are rules when you need to stop. And because of that, it requires much more attention, understanding, and it shows that if we start breaking down movement at multiple places, consistently, many people have no idea what’s happening in their body, Because if you can’t perform an action, it doesn’t mean that you actually know how your body does it, especially when you slow down. And so what was very insightful is… - Steve. - Yeah, so your work and Stephen Grossberg’s work is what I rely on the most. So he wrote this, “The During Moon, Planing in the Armour, even nose could be moved.” This was such an important point for me. Because the way we think about, when we look at what’s happening, the way we can think is that most people when they move habitually, they move these primary points, the rest of the body just falls. So they have very little awareness of how, like if I reach back, how my shoulder will move, how the entire body will reposition itself.

Paul Cisek 04:32 Sorry to interrupt, but are you familiar with the uncontrolled manifold hypothesis?

Asia Shcherbakova 04:36 No.

Paul Cisek 04:36 Because this is, I think, possibly related. So there’s a number of ideas that have been out there having to do with redundancy. That if you ask me to touch that pen, I can touch it like this, but I could touch it like that. So there’s an excess degrees of freedom for performing the goal and achieving the goal. And so the goal is specified, but the excess degrees of freedom are freedom move. And the idea of the uncontrolled manifold hypothesis is in the very high dimensional space of possible movements which there’s always variability in noise, sensory and motor noise. And if you want to try to optimize the movement and really make it accurate, you don’t want to reduce the noise on all degrees of freedom because many of them just don’t matter, right? It doesn’t matter if each time I do this I have a slightly different elbow angle. It doesn’t matter, right? And so the controller can release the certain combinations of degrees of freedom can be let free and the others have to be controlled, right? And so what I think happens there is that if you tell somebody touch your nose or tilt your body, they’re thinking about their hands only as the degree of freedom that matters and they’re allowing variability in the other degrees of freedom. When you’re describing all this, I’m thinking about this idea because I think it’s quite fundamental. And I think when people are asked to do something, they specify, again, they specify certain variables but don’t specify the others, allowing them to modulate. And the ones that they specify very often are going to be visually dominated, right? Visually, and if you describe something as somebody, they’re going to immediately translate into a visual thing. Now if you tell them to keep their spine straight, maybe they can now assimilate that into the target configuration and start thinking about reducing that degree of freedom by letting other degrees of freedom free. And maybe their knees are going to get looser. You know, there’s this concentrate off that happens.

Asia Shcherbakova 06:49 Yes, so what you were talking about in general, yes, but what happens very often is that people are unable to do it immediately. learning is required to do that. So that’s the point. And so the implications here they’re very important in just the entire movement pedagogy for non professionals where you don’t have one dedicated trainer who is just looking up to you and is helping you. And basically so what we say essentially is that we teach people to plan movement in relation to not what they’re used to but all these other points and other points as well, this is just a very simplified picture. But we make sure that people know what all these points are doing in space and in any movement they need to know what each of these are.

Paul Cisek 07:38 So they all become specified targets in the sense?

Asia Shcherbakova 07:41 Yes, yes, so from this perspective, like for example here, like even the nose can be moved, essentially it’s like reaching, but reaching not with hands, but with every point. And at first people are unable to do this, like all these blue points. It’s extremely difficult for many people and it requires many sessions Even within one session people just cannot feel it when they repeatedly do exercises that force them to Pay attention to these points. Eventually they develop these skills. Are they better at some points than others? Like better elbows and knees? It’s well the worst is the trunk. Yeah, obviously but there is a very high variability from person to person based on their background and so on. So, what else do I want to say here? Yeah, so basically one of the problems is that natural movement very much relies on inverse kinematics, right? And we force things more like this and what it does also is that eventually people become much better at our tasks but that also translates to other forms of learning and just daily life because people are able to consciously decompose what they do into these more fundamental building blocks.

Paul Cisek 08:55 And another for example… So joint level kinematics, actual joint configuration, the relative positions and angles of specific joints as opposed to just in space.

Asia Shcherbakova 09:08 We just move step by step. Yes. Yes. So this is another example. This is very difficult for many people. Essentially we want to maintain this rectangle and from standing on two legs we shift the weight onto one leg and then we lift one leg. This is just so difficult for many people because this rectangle becomes completely broken. So it looks like this basically you need to bend shift and lift the leg. But what happens is that people start to shift like this and they are unable to isolate this movement and then when they’re on the leg one leg they cannot tell how their pelvis is and it’s not just balance they just don’t understand if my pelvis parallel or not they cannot control the movement but most important they cannot sense right so this is kind of it’s kind of in this position and we need to strengthen somehow the signal basically my question So if you were to go back to just how you are here, this is based on a situation where the brain has the skill to discriminate between different positions or sensations. But what if they cannot, right? So we don’t have like somewhere in the parietal cortex here. We actually, I don’t know if it is parietal cortex between S1 and the posterior parietal cortex, at which point the down the course, but there is no discrimination capacity. If there is no discrimination capacity, then there’s nothing, nothing competes with nothing. It’s just one option, right? So our

Paul Cisek 10:42

• You can’t shift to the other option. - Yeah, they cannot shift with the other option because

Asia Shcherbakova 10:46 they cannot feel the difference. So from our point, we need to have a slide somewhere about this. Maybe this one. So basically, the developmental sequences like this. So first, we need to increase the perceptual differentiation because without perceptual differentiation there is nothing to choose from and also what I would like to ask about the basal ganglia specifically I spent so much time trying to understand what exactly basal ganglia are doing moment to moment you and the rest of the world yes well yeah those who are interested yes and the thing is that basal ganglia receive cortical input, right? So they do not receive the raw input. What do you think it would be correct that you need to be able to consciously discriminate between certain states if basal ganglia can use that information for discrimination? Not sure consciously. I’m not sure it needs to

Paul Cisek 11:44 be consciously, but probably cortically in order for basal ganglia to get that information. But you know, I think a lot of this has to do with synergies, right? A lot of this kind of… Both the…

Asia Shcherbakova 12:01 So this is about the base of the activity.

Paul Cisek 12:02 Exactly, yeah, yeah. Right.

Asia Shcherbakova 12:06 So essentially what we do, we un-chunk. Yeah. Because there is a synergy, so this is a kind of temporal activation of certain movements. And what we’re trying to do, we’re trying to isolate them into more primitive tasks. This is more about action, right? So when we this is about control so in our when we talk to people who are not scientists We say that it’s about sensing controlling and adaptation. So this is more about control This is how we teach to break down actions, but again this and I think this is more related to basal ganglia, but Ultimately if people cannot sense the difference your first need to teach them to sense and you do that by repeatedly making them do these tasks which have constraints on, as you say, degrees of freedom. We usually refer to it as just symmetry, constraints of symmetry, because natural movement usually is asymmetrical. And if we require these only blue points that I showed you to force them under an external grid, then eventually people build better capacity to discriminate between various sensory states. So another two things that also it makes sense to me, but it’s very difficult to find exact kind of proof for it in literature. So one thing is I refer to the mystery of proprioceptive awareness. When talking about proprioception and proprioceptive awareness, a lot of the time people think of it as just a sense of position of movement. I know where my arm is, right? what I’m talking here about is localized sensations in the muscles and so that was also in the poster according to our questionnaire about 43% only are aware of any muscular sensations in the belly of the muscle at rest and so these people don’t even understand what we’re talking about when we ask do you know the sensation in a contracting muscle you know that they don’t know but during

Paul Cisek 14:10 Exercise only about 48 are aware and so this is things like are your muscles co-contracted? They can’t tell whether what you mean by things like are you tightening?

Asia Shcherbakova 14:22 For example when the hand is relaxed like this you don’t feel very much right if you Yeah, use the muscles you begin to feel something right because what what is the name for that something? Oh, I guess proprioception. No, but the sensation. How would you know the sensation you’re contracting muscles? So there’s no good word for it. There is a paper which talks about, it refers to something like that as a sense of effort. And also, well, I think the best explanation for where the sensation comes from is probably it’s the constant feedback from the intrafusal fibers. So not the muscle spindle itself, but the intrafusal fibers. But that sensation has no name, there’s no scientific literature which actually talks about quantifies it somehow, names it. I’m familiar with Uwe Proski, have you heard of it? No. Well, he’s especially has an infoproception, he has written 300 papers infoproception. When I asked him about that, he said that he does not feel anything in relaxed muscles. So he is not aware of any literature which talks about constant sensations in muscles and that my best shot would be the blind people because Because this this person here with the highest intensity of the sensation at rest was blind interesting and yeah, and also In relation to this sensation

Paul Cisek 15:53 Well, relaxed muscles don’t really get much of a signal though from the spindles certainly, because they’re unloaded, right?

Asia Shcherbakova 15:59 They are unloaded, but the intrafusal fibrosis, the spindle always has to be tight enough at the current length in order to be able to signal the length. This is what is controlled through the contraction of the intrafusal fibrosis. There’s constantly efferent signal going up. And if you talk to dancers, dancers call it muscle tone. Yeah, exactly. With their body a lot they just say oh yeah, I feel the muscle tone all the time and just when I contract my muscle

Paul Cisek 16:25 I feel more of it. Yeah, and yeah, that was big part of my PhD thesis with Grozburg Bullock was was how can the corticose system? Make use of that that information and exactly as you say one of the things that has to happen is the Intrefusal muscles have to keep the spindles Loaded enough or you know, yes

Asia Shcherbakova 16:48 Yeah, so I think that that can be consciously perceived as the sensation and In some cases, this is very interesting to second words that so this this picture We can catch a moment through this training because let’s say for example somebody never used The muscles of their back and separation. There’s no need for them to sensory separate the sensations But if we train in this very symmetrical framework, eventually these muscles, sometimes you will use only one side or another side, and at some point they split. And when the sensation splits, it feels like it shifts under the skin. So this is something distinctly different from the sensations in the skin, and you can sense it for maybe like a couple of days, but then it just, the novel…

Paul Cisek 17:35 Oh, I know there’s remapping. Remapping that happens in cortex.

Asia Shcherbakova 17:40 My explanation is there are examples of when you split the fingers in the syndactyly when people were near you. So when they split the remapping happens within a few days and then they can use them separately. So essentially this is kind of like that but we separate muscles not surgically but functionally and

Paul Cisek 17:59 you can really feel the difference. So training essentially. Well it is training

Asia Shcherbakova 18:03 but what’s interesting is that the sensation is very much linked with how you use it. So you don’t have the separation of sensation if you always use the muscle separately. I mean it makes a lot of sense of course, but this is what needs to happen so that people can better control their movements and that kind of goes hand by hand. It’s very difficult to say what’s to separate the motor and the perceptual here. Again this is why all your work is essentially about that. Yes and so in just a second the last thing I wanted to ask. So, Bezel Young what did I want to ask? Well, so in terms of where this sensation sits, because 3A receives the input from muscles and joints it makes the most sense. This is also what what Prowsky said that this makes the most sense, that it would be here, but would it be perceived as a conscious sensation? From what we see and overall just thinking how it’s organized based on the available literature, it seems like if you teach people to attend to those sensations more, it is a really good step towards being able to feel the difference in movement. So even though being able to tell how my spine is, it’s about position. It’s about how different joints are configured in relation to each other. And being able to feel muscles as localized sensations, it’s a completely different thing. Those are different perceptual events. But they inform each other. And the way we do it, so if I go back to those movement patterns that evolved, essentially It’s not just about making movements simple and break them down. We also have these principles that you can think of them as that you can take any movement, which can be performed without momentum, and you apply these six principles to them. So that would our method would be. And distributed activation and micromovements, they often go together. Distributed activation is that we want to contract as many muscles at the same time as possible in any movement. And micro movements is that we keep repeating the movements and tracing them. Essentially what it looks like is that, let’s say for example, instead of just standing like this, what I will do is that spread the fingers, draw the shoulders down, pull the arms away, pull the rib cage back, tug the pelvis and also with the legs pulled away. So when I do that, it activates almost every muscle in my body. And I will add micro movements by just adding this kind of small adjustments. Keep doing that. And we do that in any movement that we do. So if you do a session for 60 minutes, you will be doing that with any movement. And based on your knowledge of neuromuscular system, what do you think it should do to the brain if you do this type of training with every movement?

Paul Cisek 21:12 That’s a good question. I mean, I kind of think that the cerebral cortex is kind of a patchwork of adapted high-level synergies for things like certain kinds of… This is a somewhat controversial view that I used to not believe, but I’m now kind of getting into it. the idea that the frontoparietal system is essentially a kind of like a higher level hierarchy of synergies. So you have synergies which just move a joint, right? But then at the cortical level, and you can still do that, you can still control just a joint. But the maps are organized in terms of kind of useful movements, useful ecologically relevant movements. So things like for like one thing that that’s been demonstrated very clearly in many species is hand-to-mouth movements. Yes. Right? So if you look at the work of Michael Graziano, you’ll see this. What happens is if you if you just stimulate activity in particular patch of cortex, what you get is a something like the hand closing, turning, the head turning to the hand and the hand moving to the mouth and the mouth opening. So it’s like the whole thing is like a unit of very of course relevant movement. And there’s kind of a patchwork of these high-level synergies that as you move around the cortex you get different movements, you get defensive movements which are kind of like this, or you get locomotor activity. And there’s a kind of a consistent map across the whole frontoparietal cortex, with lateral movements being more about the hands and face and mouth, mostly about feeding, very medial, mostly about locomotion, and then in between all the things, in between, which are things like reaching out to get something. And that’s what you get in monkeys. You essentially can activate parts of their behavioral repertoire, which are highly trained. And this would predict that there are certain movements that are easy, certain movements that are hard, not because of any real biomechanical constraints, but simply because you have specific systems for it. And so it’s almost like when you try to do any movement that’s near, it becomes one of those movements. It’s like you fall into the particular pattern that’s the common pattern. very hard to snap out of and separate them up. But with training you can, right? So you imagine that an animal that’s trained to do a particular type of movement probably develops somewhere within that map a constellation of patches that can be, that sort of work together to produce that movement. Let’s add on a pole vaulting, which is a very strange combination of movements, right? That’s extremely hard for any of us to do, but presumably other pick-level pole vultures, it’s a smooth single thing. You know, it’s like one thing. It’s like bringing food to the mouth for us, right? And so I’m wondering whether what you’re showing here on that paper where you show the stuff shifting under the skin is that idea that the cortex is being forced to not fall into the traditional pattern. So it’s not really, so like, I don’t know if you were implying this, but the idea that there’s certain habitual things that you fall into, it’s not maybe so much that it’s a habit, so much as it’s like a really large attractor in the space of possible movements. You know, it’s like any nearby movement you will just get pulled into doing that particular pattern and it takes practice to sort of change that well, you know, to make you not be able to stay out of it and not fall into it.

Asia Shcherbakova 25:26 Yes, exactly. And when we talk about habitual movement, another way we use the word movement of vocabulary and we specify exactly that basically you are really good at doing what you needed to be doing up until now.

Paul Cisek 25:38 Yes, so I think what you call vocabulary is what Graziano would call repertoire. I mean, it’s the same thing.

Asia Shcherbakova 25:42 Yes, it is the same thing. Then the practical question, how do you separate that? And so this is why also this is another thing that I want to run by you, distributed activation with micro movements. So what I understand about basal ganglia with the direct and indirect pathway is that when you perform a movement, other movements need to be suppressed, right? Yes. And then if we think about what you just said about when it’s very easy to default to a habitual movement. From what I see in the classroom and based on what I understand about the basal ganglia, what I think happens is that when I, let’s say for example the movement is like this movement is difficult. So what do we say? Okay pull the legs, sorry, pull the legs away. When I pull the legs away, it activates all these muscles here and here as well. I draw the shoulder sample, the rib cage back, here is all I’m already consciously using these muscles for something and this makes it easier to perform this movement in a controlled way because the muscle is already being consciously used. So essentially we create this field of distributed activation and somehow within the field it’s easier to perform a new combination than it is if all the muscles are relaxed.

Paul Cisek 27:00 So you mean because you’re You’re preventing the suppression of all those moves.

Asia Shcherbakova 27:06 Exactly. Yes. Yes. And so, yeah, I agree with that. So, I have this, this is like a, it’s a rough draft. This is not completely final thing, but when I’m trying to explain how we instruct, we call it while, while not if do algorithm. So, people think that when I instruct and I tell you to do movement X, you just do X. What typically just happens is that I tell you X, you interpret it, you interpret it in your habitual way, your by-stores, your habitual way, and you do actually something completely different. This is by the way the best situation where you can observe it is yoga in the park. If you see what people do in the park when they do yoga, it’s completely not what their instructors do. Oh really? Because it’s an environment with a lot of distractions and people feel kind of self-conscious. So they just do completely different things. And so what we do is that we constantly have in mind these types of instructions. First of all, we do a lot of things all together, simultaneously. So we keep saying while you do A, B and C. So you do A, while you do A, do B, while you do A, B, do C, while you’re doing A, B, C. You will do X until X1 occurs, but make sure that X is not Y. So this thing about this knot, it makes people aware at least that there is a possibility to fall into this habitual pattern. And then we also say if you send something then there will be a branch of instruction and for any movement we do this So this is like a like a sub like a surgery of movement goals essentially and also in relation to This what I found very important. So here is this about reaching so it’s about object identity But when you instruct movements, it’s not just directly goes to You have an interpretative step and but essentially I think it also is somewhere here in infotemporal cortex where you Identify what the movement is the category. It’s called recognition category, right? You need to recognize what is the movement you’re going to do and then that will inform your preferential

Paul Cisek 29:18 You know, that’s probably not this though. It probably won’t be in ferrata. Okay about visual objects, things in the phobia. It’s not going to be… So there’s lots of things that can inform the competition that goes on here. And many of them will involve these pathways, but many of them will involve completely different pathways. So I think in this particular case, where it’s really kind of a proprioceptive goal, or at least not a visual goal, it probably won’t be that particular region here.

Asia Shcherbakova 29:57 Maybe not this particular region, but still the categorical for this name.

Paul Cisek 30:00 I think it’ll still wind up creating the appropriate bias in frontal cortex. Any kind of explicit, conscious goal is probably going to be coming through here, but it might not be drawing on information from here, it could be drawing information from other places.

Asia Shcherbakova 30:20 it probably will be a little bit closer forward because it’s not just going to be an object identity, but it will be an action identity, a category for a particular action. For example, one thing that happens a lot, we instruct like this, so draw the shoulders down, keep this relationship, as you draw the shoulders down, spread the fingers, bring the arms diagonally forward. So I just want to isolate this movement. This is the movement that I want to see, but when people see it for the first time they just think oh I just need to bring my arms up and so they like like this right and then what happens is that it means that the shoulders are lifted and the chest is up so you need to move it back and that is when they just misinterpreted my instructions this is why it’s so important to say that you do X and not Y but even still even if we specify all this a lot of the times Sometimes they cannot tell whether what they do is actually what I ask them to do, they cannot feel it and ultimately it’s over, session after session we need to build these perceptual skills because they just cannot sense it. And overall it’s very difficult to talk to people about perceptual skills because first of all non-specialists don’t understand what perception is, for them perception is point of view. And then when you say, “Okay perception is a skill, okay what?” don’t understand that. And yeah, the idea that you can also censor discrimination, if you say we need to become better at censor discrimination, becoming better at discrimination doesn’t sound very good and so on and so on. So it’s just a complete disaster when trying to communicate it to non-specialists.

Paul Cisek 31:58 The people that you work with are like they they come to you because they want to get more skilled in something like dancing or usually martial arts or

Asia Shcherbakova 32:09 usually people come to us when they were failed by multiple other movement

Paul Cisek 32:16 practices that they do but what were their goals in their healthy people I mean obviously people trying to recover from so a lot of people start thinking

Asia Shcherbakova 32:26 about this stuff when they had a problem so either they had an injury they work with an injury or they for example sometimes dancers athletes when they’re young they don’t have to think about anything but then the older they get the more they like okay I cannot do this anymore but I don’t know actually I have to take care of my body so those people but also people for

Paul Cisek 32:45 example for me they want to improve their performance they want to understand the technique

Asia Shcherbakova 32:49 because very often in many places usually not the high high high level but like maybe mid level when you want to improve but you’ve hit a plateau and your instructor in your movement practice of choice cannot help you. And for example for me you know having the fluidity like this in hands and arms and dance the teacher would show this and my arm would just not do it and how do I do that and she cannot help me. So through this work with these with the distributed activation and micro movement It’s very rigid, very symmetrical. It’s very interesting that you can get this control and fluidity. Yeah. - Anything is because you’re training

Paul Cisek 33:33 the perception of the movement.

Asia Shcherbakova 33:35

• Yes, perception of the movement. So I think, what do you think about the ability to tell what your spine is doing right now? Where is it between the sensory and posterior parietal cortex?

Paul Cisek 33:48

• Yeah, well, I think a lot of that is coming through the proprioceptive pathways, more into closer to the central sulcus, it’s not gonna be too far back, it’s not gonna be in those visual regions. Like the really posterior parietal cortex, like really posterior like area seven is probably not gonna be yet. I think it’s gonna be going up much more into, right into that post-central gyrus. So areas five and three A and such, I think it’s going to be pretty direct.

Asia Shcherbakova 34:21

• So when I think about how to organize, basically the more you go there, the more abstract it becomes, and also it fuses with the visual.

Paul Cisek 34:30

• The more you go this way, it’s going to just become very visually driven, dominated by vision, right? Here you get all this stuff that integrates vision, proprioception, auditory, vestibular is very important. And so in this whole part is a very, very multimodal, right? And that’s probably where all this kind of stuff that you’re studying is going to unfold. Over here, it’s gonna be really just the stuff that’s visual, really just the stuff where they are either looking at the target to which they’re gonna put their, whatever joint they’re reaching with. And maybe observation, imitation has some, as people have discussed, imitations in those regions, but I think the thing of interest in your, in the kinds of cases you’re looking at, is probably going to be extremely multimodal, and it’s going to be like you have visual vestibular, proprioceptive, and even auditory information, and you’re going, you want to train yourself to pull out the right pieces through the movement. In your case, it’s probably mostly proprioceptive, vestibular for balance for sure. You know for dance auditory we’ll set the beat and things like that. So there’s certain, I think what happens is this region can parcelate out which combinations of multimodal signals it uses. So proprioceptive from the spine and the ankles and the arm. If that’s useful for the movement, it could learn to sort of pick it out. So I think that that probably more than anywhere, this region here. Prefrontal maybe not so much because prefrontal tends to be a bit further away from the drawing. So in relation to the

Asia Shcherbakova 36:21 prefrontal cortex, so if you look at this framework, like what fails, I feel like prefrontal cortex is just, it’s really about understanding instructions in in this whole context. So it’s basically if I specify more, preferential,

Paul Cisek 36:35 Yeah, it might be good at evaluating, like whether you’re doing it well or not. But the, sort of the perceptual learning, I think, is going on in the parietal, I think.

Asia Shcherbakova 36:45 I mean, this, well, this is also how I think, and yes, but also I wanted to ask in relation to, I don’t have a picture here, but I’m sure you’re familiar with it. Sometimes they show what you’re talking about the behavioral repertoire, where you have like an area here here and they’re kind of connected here and here they connect right? So that’s it. And you call them idiosyncratic maps. Oh yes, maps here. By the way, where does it come from?

Paul Cisek 37:11 Did you come up with this term? Well idiosyncratic just because that’s the way that’s scribing, but the work that it’s based on, you know, I can show you one paper which I found very beautiful. So a lot of my view has become modified from certainly from those papers, because I’ve been reading much more about the evolution of the brain. And so I’ve been reading more broadly. But there’s this very nice paper from this group. It’s actually Camini really, Roberto Camini, who’s leading it. And I’ll just show you a couple of pictures. What they’ve suggested is that if you look, and this is mostly based on monkeys, right, but it will apply to humans, is that there’s certain connectivity patterns between different regions. And they go through a great depth. But let me just summarize the take home message. The take home message is that there’s certain circuits, sort of what they call modules, for different kinds of activity. And they’ve sort of mapped this out in some detail. And this is what I was calling the, you know, this is what I think a lot of the Graziano stuff, and he let a lot of these ideas about action maps is, is that certain circuits that are really involved in certain kinds of activities, like reaching tool use and object construction, when you’re sitting in your butt, relatively stable, and just looking down and doing stuff with your hands, which of course primates are fantastic at this, especially humans, right? But there’s other modules like locomotion, which are completely anatomically, or largely anatomically distinct, even though they still use the same effectors, there’s a separate map in the singulate motor areas that has a separate representation or separate connectivity with the arms that’s used for locomotive with the arms. Another for grasping, I think. This is grasping and observation and hand mouth, a lot of hand mouth synergies. Anyway, the point is there’s anatomically distinguishable modules and the thing that’s actually remarkable about them is that they have their own projections of spinal cord. they don’t all have to go through motor cortex, right? So we think of motor cortex as like the final common pathway, but it’s not really. And we’ve actually known that for a long time. There’s just nobody really kind of appreciated the meaning of it. So the point is that these things exist. And what I mean by pediosyncratic is that if you’re sitting on your butt and manipulating stuff, then the kind of information you need about space is peripersonal space. That’s what matters, right? Peripersonal in front of you, distant stuff is not relevant to this behavior, right? And so, and there’s a lot of work on peripersonal space perception or something is in reach or out of reach, right? And so there’s a kind of a system that’s been adapted to using that information and not other information. And in that sense, it’s idiosyncratic, right? Whereas the medial locomotion system, it’s going to be all about distant space and obstacles and trees or whatever and navigation, right? So it’s gonna be very different. They’re gonna have different, like in this, for its competition idea, I’m really just talking about one of those things, the reaching one, right? It has its map, it has its, you know, hills of activity in that map, but the locomotion map is gonna be rather different. The grasping map is gonna be all about object shape. It’s gonna be very non-Kartesian because it’s going to be all about, you know, is it curved? Is it thin? You know, it’s going to be very much in the case in the sense of object space. So the grasping network which is even more lateral is going to probably have again an idiosyncratic space. So there’s ventral premotor and anterior interparietal that are going to be very different than dorsal premotor and media entrepreneur.

Asia Shcherbakova 41:30 The way that I understand this idea is that you have a type of actions which are associated with particular body parts and also particular space, like for example, experience or before the foot when I step I need to also. So maybe there’s like this space is not doing really well. But they are idiosyncratic in a way that they are not really connected.

Paul Cisek 41:54 They’re quite different and they’re kind of really useful for certain kinds of activity, but completely useless for other kinds of activity in a sense.

Asia Shcherbakova 42:05 But how do you think it connects to this idea that Grossberg expresses is that if you want to move something you can move something even if it’s your nose?

Paul Cisek 42:14 Well I think in his case, I mean it’s in the context of what they call the direct model. He probably talks about adaptive resonance theory, but the model that’s actually more relevant for this is what they call the direct model. It’s actually a really, really great model in my opinion. The model essentially has a little bit to do with what I was saying, the uncontrolled manifold idea, that because you have an extremely redundant motor system, you can accomplish the same thing in many different ways. And the way that model accomplishes it is essentially kind of like it’s based on Jean-Pierre Gé’s idea of the action perception cycle, that babies learn to control their bodies just by trying stuff, right? So they wiggle and they see— Yeah, circular reactions. Exactly, circular reactions, exactly. That idea, and I think it’s an extremely powerful idea, is implemented within that model to say, “Well, if you want to reach this thing, what you need to do is you have to make movement through space that drive you towards that thing and it will become a straight line, just it’ll just emerge to be a straight line and it’ll use whatever degrees of freedom are available to you, right? And the nice thing about the model is because it’s a feedback control system, if you suddenly let’s say block the elbow, it’ll use the remaining degrees of freedom to get there, right? Like a human will do, Right and if you give them a tool if you give the model a tool it immediately will reach with the tool It’s just learns the whole thing because it’s all a kind of a clever Mathematical trick for dealing with redundant systems, right? In fact, I can show you simulations that I do in classes of this model because I think it’s a nice demonstration of Of the power of these kinds of models And so the point is if you define the effector to be the finger then the finger will land on the target If you define the effector to be the wrist, then you’ll point with your wrist, you know, and it could be the nose, right? And so I think that idea means that Because you have this system that can land on a target with the redundant joints you have If you define the target in some novel way It’ll be able to deal with it, right? And so this this may be again, this I think is relevant to the uncontrolled manifold and this idea that you’re pointing out There’s certain parts of the body that you’re very good with other parts that you’re not so good with But if you train then you can reach with your elbow You know or with or with your nose or your knee or more you could you could constrain the target to be Not just a position but rather configuration Like your spine has to be straight, you know, in that sense. There’s certain constraints that you could ask them to do and again it I mean mathematically, I think the model could probably learn it Because it’s essentially learns from errors if you define the space of errors in a particular way it’ll learn to minimize that

Asia Shcherbakova 45:29 As you were just speaking I’m roughly aware of These theories that you’re talking about but just as you were speaking actually thought that reaching, it’s true that, for example, if I wanted to reach here with my elbow, I don’t feel like I need to learn, I can just do it. Yeah, that’s right. With my foot I could do it. But also, as you said, mathematically, it’s quite simple because just one point needs to move somewhere or define one point. But what we do is that it’s not like one point has to go somewhere. Many different points has to agree with each other in an external, allocentric, essentially, or I don’t know if it’s allocentric, but in the external grid line, 3D grid line kind of space, and they all have to agree with each other. So it’s actually a completely different task, right?

Paul Cisek 46:19 Well, I’m not sure though, because, you know, reaching a, let’s say I want to just touch this thing. Well, I could touch it in any place, right? So that’s an unconstrained, unspecified target. You can tell me, touch the thing, right? I can touch it anywhere, right? Yes. But I still have to control the degrees of freedom so that I don’t break my finger or land over here, right? So you’ve constrained the target to be this horizontal space, right? You could constrain the target to be a 2D space like this, right? I just have to touch the window. Or you could define a particular point, right? Or you could define a particular point with a particular joint angle. So you could be even more specific in your request of what movement you want me to make. One is, you know, touch the window, one is touch the window at a particular point, and then even more specific touch the window at a particular point with a particular elbow angle or a particular body posture. So I think the system as a whole handle multiple levels of ambiguity of which are the degrees of freedom, the external degrees of freedom, and the intrinsic degrees of freedom that are part of the target and which ones are free to go wherever you want to go.

Asia Shcherbakova 47:37 But as long as you can actually track what’s going on in terms of whether your performance is actually close to your target, right? Because what we work with is that we, okay, we say so it’s not just one point that you need to move, we have several targets and we need to move several targets which are all kind of connected to each other in different points and the problem that people have is that they can’t tell whether their targets are in the right position.

Paul Cisek 48:04 Yes, so what you need is you need them to have the right expression of what error is, right? And if error is just the endpoint of the finger then they’ll just do that and I’ll leave the other joint street. But if air includes a measure of where the elbow is, then they’ll put the elbow in the right place. So I think this is where the perceptual training comes in, is that that part of the parietal cortex, which is sort of adapted to the sort of idiosyncratic action type that they’re doing, it’s sort of measuring your state. And I don’t know, maybe it could be your state, but it could also be your state with respect to some definition of target state, right? And there’s certain dimensions that matter and certain dimensions that don’t matter. And so what happens I think in the natural situation is many dimensions just don’t matter to us, right? And so we fall into the solutions that get the target, but everything else is just like habitual, that’s the attractor that you fall into. And then through training I think you could sort of force a more specific set of constraints. And again if people are able to measure the state then they can do it. If they can’t measure the state then they won’t be able to do it.

Asia Shcherbakova 49:22 So that’s the part that’s tricky. Grossberg has a very interesting idea so when he talks about what’s conscious and what’s unconscious he says that any conscious state is a resonance state. So he says that essentially movement is not a conscious process because it’s about the vector moving towards the zero and it’s the mismatch rather than the resonance. One thing that whenever I read this I think overall just makes so much sense I think it’s very beautiful framework but it’s specifically in relation to movement it feels like he’s actually missing the awareness of position as a conscious state. Because if I feel something in my muscle, I am aware of it. How is it different from me seeing a collar? I believe there is resonance, but this is a type of resonance which we don’t talk about, right? So this is probably why it wasn’t on his radar and it doesn’t seem right now that there is a clinical significance, although from my work I’m sure there’s clinical significance, just a whole domain of knowledge that is not mapped and people don’t realize how important it

Paul Cisek 50:32

You know, the conscious unconscious thing is a little bit of a kind of a black hole for theories, right? Because there’s always this bit that we really don’t understand, conscious or an unconsciousness. And I think we’re probably pretty far from understanding it. I think Grossberg has got a proposal for it. Lots of people have proposals. But the thing is for you it doesn’t actually matter how to explain the mystery of subjective consciousness, right? For you what actually matters is a much more pragmatic thing. Are people able to control certain aspects of their movement or are they not able to control? Whether they’re subjectively conscious of it is a…

Asia Shcherbakova 51:16 No, but in order to be able to control it you have to be in some way to be subjectively conscious and where consciousness comes. I’m not talking about something…

Paul Cisek 51:24 But it doesn’t have to be subjectively, right? You can ride a bike without understanding what you’re actually doing. Because the only thing that you’re controlling, actually sort of consciously controlling, is the direction your bike is going. And all the other movements, all the strange, very unintuitive actually, biomechanics of riding a bike, are being controlled beautifully. By like the cerebellum and other parts. And it’s just, it’s like riding a bike.

Asia Shcherbakova 51:55 Yes, which is normal and just completely how it should be most of the time, unless you actually have a problem that you’re trying to solve for, you want to consciously control certain aspects of these movements and you don’t have consciousness. So when I talk about consciousness, I’m not talking about some universal theory of consciousness or what is consciousness. I’m just talking about whether or not you are aware about a particular distinction in some sense or modality. Right. Right, and so that is, yeah, that is something that is…

Paul Cisek 52:24

• Sure, a little bit like what I was saying about error, that you incorporate certain variables into your evaluation of how well you’re doing, or your faith.

Asia Shcherbakova 52:35

• For reaching, the conscious awareness is vision. When I reach, I constantly have the visual feedback. If I close my eyes, I have no idea how close I am until I touch, see, like I missed it. But when, with proprioception, We don’t always, what is the feedback for proper perception if we don’t use the visual feedback? So that’s the thing. And when somebody is doing it, for example, when I’m telling them, as you reach back, you want to have a completely straight line out with the leg, your upper body, everything has to be completely straight. They’re looking here, they don’t see their body. How do they know? Are they like this? Are they like this? They cannot tell. So they need to learn to tell only from proper sense of feedback. back. So part of it we call spatial awareness, how is my body oriented in space and then what we call proprioceptive awareness, proper or localization of the sensory whatever, is whether or not they can feel a sensation in the muscle. And we find that when you… these two things they grow at the same time. The more you can control the muscles, the more you can feel the sensation separately in the muscles. And also just to go back to this idea of idiosyncratic map. Also, one of my ways to potentially explain it is that by doing all this, by enforcing a unified external grid line or movement on all these constraints and seeing that you need to make sure that all these points at the same time, they fit certain grid, which is very symmetrical, we are actually teaching the idiosyncratic maps to talk to each other so that they can plan movement all together and agree with each other irrespective to what one of the other is doing. And for example, for reaching there is a separate map, as I understand it, to bring a hand and a separate one to grab and they know how to collaborate really well, right? But they’re still separate systems. So what we are doing is that we’re teaching all these separate maps, maybe we’re developing new maps, but also the ones that we have, we refine them and we make them

Paul Cisek 54:43 communicate, right? And also… Yeah, like eye-hand coordination, right? There’s the gaze system and the reaching system, they work together extremely well because they have to. Yes, because they have

Asia Shcherbakova 54:53 to and they had to from the first days of life. So what the way I see it is what we’re doing is that we’re just through these circular reactions, micro movements, you did it and then you keep reaffirming, reaffirming, this is the closest we can get to these circular reactions in adults.

Paul Cisek 55:07 And you’re creating them, but you’re creating new sort of constellations that have to work together.

Asia Shcherbakova 55:13 Yes, yes, yes. And we reconnect them and we teach you to, yeah, to have conscious access to how they are. Because most of the time, for example, for reaching, when I’m reaching, I don’t need to know what the rest of my body is doing. Right? But in our case, our tasks are required that every part of the body knows everything that is doing. Also going back to what you said that there are many, we showed the paper there showing that there are sometimes pathways going directly into the spinal cord that not everything happens. So that’s essentially also part of what we’re fighting with because there are various, for example, right knee reflexes which are kind of below cortex and also just the way the spine is it’s innovated. When I was looking at how we just all these movements, why are they so smooth? There’s for example something going, descending pathway and it branches. And let’s say it connects to three or four or five different vertebra at the level of different vertebra, controlling the muscles around that. And then you have another pathway which goes lower, but it overlaps like this, trance and ventral.

Paul Cisek 56:26 It’s the same axons that branches.

Asia Shcherbakova 56:28 The same axon that branches. Here you have more density than here. Because it’s all like that, it’s almost like the diffusion in the sense of liquid would diffuse into another liquid. So this is why fluidity is possible. Because as the signal goes down, one point will receive a lot of signal.

Paul Cisek 56:48 And then there will be a gradient of fluidity.

Asia Shcherbakova 56:50 Yes, so this is why all these muscles around the spine are able to achieve such fluidity. But a lot of this, So basically that fluidity, the way that it all connects in the spinal cord is part of the infrastructure that allows the movement to be fluid. This is why it’s so difficult to only turn the torso because you have to somehow break something which is under the cortex. This is why it’s so difficult. And this is why, for example, what we do with distributed activation when you pre-activate the muscles, This overwrites the subcortical commands.

Paul Cisek 57:28 Yeah, this is actually one idea for something that primates had to do in order to really have good hand movement. It just was not possible to do it through the old pathways, because the old pathways are so sort of hardwired to do certain kinds of things like for walking, that primates essentially evolved a direct path from the motor cortex down to alpha motor neurons to control the hands because the old pathways just you couldn’t break sort of the synergies that were already embedded in them. And so to do things like play the piano you have to do it directly. You have to have a separate controller for the muscles. So the cortex of course is extremely plastic and how it can change. You know the phenomenon that you talked about with the splitting, that’s something we could actually experimentally test with TMS. So with TMS we can look at sort of the muscle output of a particular site in the cortex. And you know, there’s a very classic study that showed, a very nice study that showed, had people sit still and they would find a spot in the brain, in the motor cortex, that if they stimulate magnetically, it makes their finger twitch, right? was a twitch of the thumb in a certain direction, I believe. I might get the details wrong. But anyway, so the twitch in a certain direction reliably. And then what they did is they trained them to just do this for like ten minutes, right? And then when they twitched, the same very same location induced that movement, right? Because the map is very plastic and it’s constantly adjusting, maybe not at the level of synaptic connectivity, but in terms of sort of activation and whatever. And so, you know, essentially you can observe the remapping through training. And so something like this phenomenon where people suddenly have the spine modules split into two modules could actually have a sort of a cork or spinal correlate. Well, maybe not that one. But something with the hands and the arms probably could in terms of showing how the training breaks up the synergies into subgroups.

Asia Shcherbakova 01:00:10 I think with hands because hands it’s a really fine more control that will probably not happen with humans unless you keep them in this state for 20 years. Oh yeah? I think what’s happening over there in our case is that the separation happens for gross Musculature which normally people don’t have fine control of this is why we are able to even get this This is a conscious sensation that it shifted like by this much. So you actually need

Paul Cisek 01:00:47 Yeah, maybe you’d work better for those muscles

Asia Shcherbakova 01:00:49 You need bigger muscles and also in relation to whether or not you can induce something like that with transcranial stimulation. I’m not sure because I feel like my intuition is that you actually need the more perceptual pairing and

Paul Cisek 01:01:07 Well, no, but it’s not the stimulation that induces it. It’s the learning that induces it.

Asia Shcherbakova 01:01:11 But you’re saying that it’s possible to do it within 10 minutes.

Paul Cisek 01:01:15 Well, the stimulation is just essentially probing the map, right? Okay, probing the map. It’s probing the map, it’s stimulating and creating a muscle twitch, but the changes in the map are induced by the training.

Asia Shcherbakova 01:01:31 This is what you’re saying, that there’s a population of cells which is primed towards certain direction and the same population can actually be…

Paul Cisek 01:01:40 Yeah, I think it I mean in the case of the study that was mentioning right class and at all that was I think something like 20 minutes, I mean maybe it’s maybe it’s synaptic changes Maybe there are actually synaptic changes happening and that in that short period of time. It’s it’s There’s no reason why it couldn’t But it could just be that you’re biasing the activity. It’s not just don’t you know Cells and motor cortex of course are tuned to different directions and movement, right? But that tuning is not something that’s simply a property of that cell. It’s a phenomenon that that cell exhibits because of its interactions with all the other cells, including inhibition from other cells. So cells preferred direction in space could shift not because it’s receiving different input, but because some other cell is more or less active. So competing cell is more active or less active. And so this phenomenon of remapping with practice could just be a change in the population, right? It could be a change in the activity of the population so that a particular site now is going to be affecting the musculature differently. In the case of what you’re suggesting, it’s very fast though, right? This shifting under the skin business.

Asia Shcherbakova 01:02:57 No, it’s not fast. It’s not fast. No, the thing is that it happens after…

Paul Cisek 01:03:01 Days of practice?

Asia Shcherbakova 01:03:02 Well, it depends on how often you practice, how much you practice. maybe after a couple of weeks.

Paul Cisek 01:03:08

• Okay, so it’s not a very fast one.

Asia Shcherbakova 01:03:10

• It’s not a very fast one. And it lasts for maybe a couple of days where you actually feel that something is different and then you stop noticing it. You don’t feel the shift anymore because at first it feels different and then it just feels, it just becomes normal.

Paul Cisek 01:03:26

• It becomes normal. It doesn’t go back to the old way, but it just becomes not.

Asia Shcherbakova 01:03:30

• No, but also it’s just a very strange sensation because you feel like something was like this, and now you feel like what was like this is moved. So you feel like somebody opened something on your body and it’s just a very weird sensation, but it goes away pretty quickly. But the thing is, I don’t know whether the speed at which we observe the change is something that is biologically constrained, or it’s more just, or I finally understood the task. - Oh yes, that’s true. - And I made it correctly. And then the change happened. But at the same time, again, from just observing how people learn, there are certain things which are possible to fix just by making people notice certain nuances. And other things, it feels like they require some serious learning. This one, to be able to tell what is the position on the pelvis. Some people have to go through multiple sessions. And within one session, they may not feel the difference at all. And you’re like, okay, we’ll try again. and then the next time you try again it doesn’t work. And then eventually they’re like, oh, finally I can feel it. So usually the moment where you finally can feel it, it just feels like this. But it may take—

Paul Cisek 01:04:44

• And it’s very, a lot of among people?

Asia Shcherbakova 01:04:46

• Yeah, it’s very, very common. Yeah, of course, when somebody comes with a background, like dance or something like that, when they have more control, they usually do it much better, even though there are still maybe certain tasks that they find surprisingly challenging.

Paul Cisek 01:05:02

• Let me ask you this. So you and this system of learning have sort of converged in these principles. Are these principles similar to what dance instructors have been using for?

Asia Shcherbakova 01:05:15

• Some of them, yes, some of them not. So if you look at our principles, maybe like for example, the fixing, separating, isolating, where we isolate just one part of the body, we can, well, actually not. Dancers, they do isolating exercises. They just need to be able to move different parts, different joints separately. They don’t use distributed activation with micro movements to learn that. They just look in the mirror and they train it. We don’t look in the mirror. We never use the mirror because mirror will just completely remove the necessity to use proper setting input.

Paul Cisek 01:05:48 Yes, exactly. And maybe that’s what, yeah, yeah. Dancers probably just rely on the visual.

Asia Shcherbakova 01:05:52 Yes, dancers rely on the visual and we teach people not to rely on the visual and So distributed activation with micro movements is the unique thing. I’ve never seen anything like that.

Paul Cisek 01:06:04 And you think it’s because you’re preventing the suppression of everything else?

Asia Shcherbakova 01:06:08 I think this is it. But another thing is that I think it also increases the amount of sensory input. So you constantly have more sensory input to pay attention to. Because as I said, for the sensations in the muscles, some people only feel them when they’re contracted. So if they don’t feel it at first, then you just don’t feel anything.

Paul Cisek 01:06:29 So you’re creating a richer input so that it’s easier for them to notice.

Asia Shcherbakova 01:06:37 Yes, and also for example, obviously you have clothes and when the muscle is contracted the skin will also pick up something because there are fast adapting and slow adapting receptors. I think when we keep moving for the fast adapting receptors we also keep sending the signal that’s also part of it. So essentially the idea is that you just activate everything and as much sensory input as possible and then give clear instructions about what to pay attention to. And then eventually people build skills which are when it becomes a habit, when you can easily position your body in a certain way and you feel the joints connecting in one line, it becomes a skill to be able to come into that position. And also the, so when also, when that happens, when the symmetry becomes a habit, there is what also happens with perception is that the perception of space becomes aesthetically pleasing.

Paul Cisek 01:07:34

• Yeah, I know, I read that, that actually it becomes rewarding, that’s interesting.

Asia Shcherbakova 01:07:39

• And this is, I think when sometimes people say that, for example, they did dance or yoga or something like that, and they feel more balanced, you know? You can interpret it as an emotional quality, right? just feel at peace with myself. But I think that in many cases it actually means that they are aware of different points which are symmetrically aligned and it creates almost visual like harmony. Right. Right. And also I did this training for a few years and it completely changed my perception. So I also can draw on my personal experience and it It really feels like I’m entertained by how different points in my body are relating to each other. So it’s a very visual experience, even though I don’t see, but I perceive it. When I discussed it with a blind person, she’s also a neuroscientist, so it was a very interesting discussion. She talks a lot about how… She calls space sticky. So for her space is sticky and I also can relate a little bit to that. It also connects the space which you inhabit per personal space to what you use for mental imagery. Because also one of our hypotheses, which are, I don’t know how to test them, that if you can have better resolution of the per personal space, not just here, but the entire space around you, you can use that as a sort of kind of working memory, almost like screen, because if you can place more items, imagine and stably place items onto space, like with a sticky board, then you can hold them more, you know. Yeah, you know, we’re starting that actually with

Paul Cisek 01:09:36 my colleague Andrew Green who’s really interested in perception of body movement through space. And yeah, we’re looking at these kinds of questions of, let’s say, even in the absence of vision, how do you keep track of where things are? And the question that interests me the most is, do the affordances, are the affordances sort of low, certain circuits like the reaching system, presumably the affordances of this cup is a reachable affordance right now, but it should be completely different at this point, because it’s not affordance, I cannot reach it. And we’re actually gonna do the experience where we can see that as we, let’s say, move the animal, this is gonna be recordings of monkeys, where objects enter into that peripersonal zone, do we now see activity appearing and the whole competition changing because there’s a new affordance, right? I mean, this is a bit far from what you’re talking about, but the idea that there is a, again, this sort of specialized map of things you can do, but it really only goes to a certain distance for certain things.

Asia Shcherbakova 01:11:00 Because you know how much you can reach. Exactly. You have a paper about navigating the landscape. Yeah, that’s right. There’s a hierarchy, right? Exactly. If I’m away from the cup, basically I perceive it as an affordance, but I know that to unlock that affordance, I’m just going to step away, but I need to actually move, and I will…

Paul Cisek 01:11:17 Yeah, exactly. That’s exactly it. So you perceive the predicted affordances of reachable cup up if you exploit the affordance of moving forward. And this is, we think with Giovanni, who’s the co-author of that, that this is kind of what really gives you— allows you to build intelligence on top of sensory motor control. You know how your actions affect your next state and what actions that makes available. And some of those things will then lead you to whatever is some abstract goal.

Asia Shcherbakova 01:11:48 So it’s a very kind of grounded sensory motor, grounded in action, example of something more abstract, like, for example, a degree, you know, exactly. Exactly.

Paul Cisek 01:11:58 Yep. You work so you can deposit money in the bank so you can buy food so you can eat food. But I need to spot food. And you know, there’s a paper which I

Asia Shcherbakova 01:12:10 sometimes go back to it and somehow I still don’t fully understand it about, it asks a question, are all the allocentric maps ultimately egocentric. Are you familiar with that? No. Well, I mean, it’s just if you Google something,

Paul Cisek 01:12:25 I mean, I can find it and send it to you. Are all alocentric maps egocentric? Yeah,

Asia Shcherbakova 01:12:29 so basically, when I think about… Do you remember what it was? I don’t have internet, but if you Google something like that, all alocentric maps, egocentric, I don’t have it on my computer

Paul Cisek 01:12:43 somewhere. Yeah, here we go. By Philemon. Oh, I know Philemon. Flavia Philemon. So the question

Asia Shcherbakova 01:12:59 sounds very interesting, but somehow the, every time I go back to this paper, I don’t fully get like I really understand it, and it doesn’t really change much how I think

Paul Cisek 01:13:12 about things but… That’s an interesting question to raise certainly. Yes it is a

Asia Shcherbakova 01:13:17 very interesting question to raise and when talking about allocentric navigations more about well navigation, hippocampus right and then egocentric is more about this per person space. Yeah more parietal. Yes and each body part has a per person fist. But what if we can superimpose them and yeah but at the same time I also see studies that show for example that there is hippocampal activation in reasoning?

Paul Cisek 01:13:45 Yes, yeah well the hippocampus, yeah it’s almost like you use that map for lots of things but you know the region here that I searched for, just searched for whether she mentions it is retrosplenial. Yes. Which is a very important part of this network for this retrosplenial right here. Yes. Which is huge in rodents but not so big in monkeys, but anyway, it’s this, where’s that paper that just now? This is, yeah, I don’t know where she’s going to go, but it’s kind of egocentric, but it’s kind of in between the allocentric maps of the, anyway, I should take over this paper. This is mostly human imaging, eh? It looks mostly like human imaging. I know Phil Emelman isn’t an imaging person.

Asia Shcherbakova 01:14:35 But it’s also possible that there are different strategies because for example, allocentric in a way, you need allocentric to map affordances, egocentric affordances on an allocentric map in some ways. So it’s like a bigger map that can include smaller maps and you know, okay, from here if I can get to this point in this map, essentially it’s like, okay, I am here, I need to be here, this point here then at this point I can pass the activity to… An egocentric thing, yeah.

Paul Cisek 01:15:09 Yes. Yeah, I mean my guess is there’s lots of allocentric maps, just like there’s lots of egocentric maps, but I mean the hippocampus is sort of the one that’s been really studied well.

Asia Shcherbakova 01:15:22 So it’s easier to study.

Paul Cisek 01:15:24 Yeah, yeah. It’s very rich, certainly. Anyway, well, I mean, I think I understand better what you’re doing and why you think it’s related to the stuff that I’ve been thinking about and people in my field. I don’t know if I’ve answered any questions.

Asia Shcherbakova 01:15:48 Well, it’s certainly…

Paul Cisek 01:15:49 Because you had some specific questions.

Asia Shcherbakova 01:15:51

Yeah, my specific, well, I mean my specific, okay, let me look at my specific questions. I think the first two questions are maybe more related to the foreign competition than my last two questions. >> Yeah, here you go.

Paul Cisek 01:16:05 Could you low level, I remember some of them I could answer, you know, could low level musculosensiton at the rescue, useful sensory evidence for action selection? Yeah, sure, I absolutely believe that’s the case. But I think you’ve already pointed out that the co-contraction is useful because that makes it more available. Anyway, go ahead, you wanted to ask a specific thing.

Asia Shcherbakova 01:16:30

• Well, yeah, just looking at my specific question, the first one was about the perceptual skill training. If you think about this, what happened with this method then for 10 years we were optimizing for communicability of movement. And what emerged is the focus on perceptual skills.

Paul Cisek 01:16:52

• Yeah.

Asia Shcherbakova 01:16:52

• Does it make sense to you?

Paul Cisek 01:16:54

• It does. I certainly wouldn’t have predicted it though. I wouldn’t say that my hypothesis would have necessarily predicted that. - But because you’re— - It’s only because I haven’t thought about that.

Asia Shcherbakova 01:17:08

• Your hypothesis is about what happens in this very moment in this brain, in this state that is it. - Yeah, exactly.

Paul Cisek 01:17:13 Essentially this is asking about learning. Yes, right, right, which I kind of avoid learning, you know I mean not because I don’t think it’s important, but just because you know well it’s a completely different topic Well, no, I don’t think it is a completely different topic. I just think it opens up It’s a different dimension. Exactly. It’s like I can only handle so much, right? But it is actually something we’re studying more much more now Even at the neural level, but the thing is Yeah, I mean the idea that perception The ability to perceive makes it possible for you to do both the specification Better and the selection of course I was thinking but the specification is yeah, I mean that’s part of it too

Asia Shcherbakova 01:17:54 I think but also for selection in order to select between two things you need to be aware about yeah The distinction between the two things so this is why they all go hand in hand. Yeah, I

Paul Cisek 01:18:06 Mean I guess yeah, I mean I guess there’s a parietal part of the learning which is the having the information and sort of separating out the pieces into the pieces that are more appropriate for the task that you’re trying to learn.

Asia Shcherbakova 01:18:20 The pieces that are useful for the task because if we were just aware of everything at the same time that would just be over. For example in the back muscles you just never need to until you start doing certain types of training to be aware of them as separate.

Paul Cisek 01:18:37 Right? Yeah, you can use them as one big thing. But I think there’s also a prefrontal part of the learning which, as again, is kind of like the rule, you know, which things are almost like telling the rest, the frontoperital system, which activity patterns you should try to have, you

Asia Shcherbakova 01:19:04 You know, like, you definitely need to have some higher level directives about what you should train. You need to be able to realize, OK, this is useful. I’m bad at it. I need to do more of it, right, rather than, oh, I’m just going to do what I’ve always done. Yeah, right. So there is some, just basically what structured learning is about. You’re bad at something. You need to get better at it. One more thing I remember I just have. I’m wondering if you’re aware of this paper. this one. I know it’s about learning in the center, not really into learning. I think this paper is computational. So it’s, I don’t know how it really applies to actual animals, but this is something that is bothering me a lot, just the role of the basal ganglion, both learning and moment-to-moment movement control. And it essentially is talking about how Once the movement is learned, you don’t need basal ganglia anymore. And also from Parkinson’s studies, there’s also this idea that it’s better to have no input from basal ganglia than having a noisy one.

Paul Cisek 01:20:16 Yeah, so the basal ganglia are fascinating but frustrating because despite the fact that we’ve studied it so well, there’s no unified consensus in the field about what they do I mean, I think everyone will agree that they’re involved in reinforcement learning. There’s enormous evidence for that. But like, in what context and in what sense, like, and the thing that I’m always very interested in is the decision making. To what degree are they involved in decision making? And, you know, we’ve… Many people have recorded in the Basel Gangliet during decision making tasks. And the results are kind of contradictory. victory. You see a lot of decision-like activity, like it goes up with reward or indicates which choice a mouse or a monkey will make in this triadum, for example. And so it looks like there’s activity in there that’s involved in the decision, but a lot of that activity happens right after the movement has already started. So it can’t be causal to the choice that was made. It’s just too late, right? The other thing, which is, so in one sense there’s activity that looks like decisions, suggesting there are involved decisions, but then it’s kind of late so it looks like maybe they’re not, right? On the other hand, if you stimulate the basal ganglia, you can bias an animal’s choices to go one way or the other. And yet, if you completely eliminate the output, there’s no deficit, you know? And so it’s extremely confusing. I think the answer has to do with the fact that it’s a very recurrent system, it’s a very interconnected system where, and at least our studies in the basal ganglia have suggested that we focused on the output nuclei and we don’t see decision activity there until the decision has essentially been reached in cortex. So the basal ganglia, at our task cannot be responsible for contributing to the unfolding decision and cortex and even the decision that is reached in cortex happens before the basal ganglion really reflects anything, the output of the basal ganglion. So we think that the inputs the basal ganglion are reflecting the decision value because the cortex is sending that there, right? But the basal ganglion itself is it’s almost our proposal is that it’s involved in the commitment but not the not the choice, right? It’s not gonna lean you one way or the other. It’s going to sort of open the gate when you’ve leaned enough. You know, when you’ve leaned towards one movement or another, it’ll say, “Okay, go ahead with it.” Right? And the reason why you stimulate in it causes your choice because by stimulating the basal ganglia, you’re sort of shortcutting in the system. You’re creating the illusion to the basal ganglia that the cortex has made a decision. And then it comes along and says, “Okay.” But if you entirely inactivate the basal ganglia, you can show in at least many conditions no deficit. But the reason for that is because the cortex can in most conditions make the decision on itself. And the basal getting this really only critical for when it’s difficult to make that decision on yourself. So for example, if you’re in a state of ambiguity, you’re not sure which way to go, you need a little bit of a push and that positive feedback can create that push. So it’ll take a small difference in evidence for doing one thing or another and turn into a big enough difference that you’ll commit. right? But if you inactivate it you can still do movements which are easy where it’s really clear, right? And you can see this in Parkinson’s, right? A Parkinson’s patient is going to have a really hard time doing something telling them go lift your arm, right? Because there’s not enough kind of external motivation for doing and the internal motivations lacking because of their deficit. But if you throw them a ball then they’ll reach it right away because there’s an external signal that’s making that movement just so much better than the rest. It’s unambiguous and it just does it, right? So, anyway, so we’re working on this, but you know, there’s always a piece of data that, you know, whatever theory you have, there’s always a piece of data that just makes that theory suffer. But I’m actually kind of wondering whether this is potentially, I don’t know what you to think about this. But the idea that it’s kind of like a, again, kind of like a multi-level selection of synergies. Again, thinking of synergies in a very, very broad sense. Synergies from the point of single joint and synergies from a whole body movement like bringing hand to the mouth or walking. In a sense, Bézol Ganglion may be doing a very, very high-level selection of behavior type. You know, like, now is a good time to reach for something, stay in your local environment or reach for something, versus get up and go somewhere else, right? Essentially by activating those different maps, right, like for reaching and peripersonal space, the lateral map, the medial map for locomotion, essentially, basal ganglars sort of say, “All right, now is a good time to walk, now is a good time to stop walking and reach,” you’ve landed in a place where there’s lots of good stuff. So if you think about foraging, foraging is a task of you pick berries in some bush and then you’ve depleted the bush, time to go somewhere else. So the basal ganglion is sort of saying, all right, you’re getting good results, keep doing what you’re doing. Now, well, it’s not so great anymore. Why don’t you go and do something else? And so it’s activating sort of like a high level selection between different behavioral systems. But then maybe it’s within each of those behavioral systems, it can also modify a little bit, like which way you walk or which arm you reach with or something like that. But I don’t know, again, like I said, there’s always a piece of data which screws up any viable theory. And anyway, but I mean, if you’re interested in we’re doing that right now, asking that exact question, well, we’re setting up.

Asia Shcherbakova 01:26:44

• What base again we are doing?

Paul Cisek 01:26:45

• Yeah, well, we’ve been asking that question for a long time, but we’re doing this study where we’re doing it already with humans, with virtual reality, where we have them do this apple picking task. You ever go apple picking in the fall?

Asia Shcherbakova 01:27:01

• Yes.

Paul Cisek 01:27:02

• You know, you do exactly that. You go to a tree, you climb up the ladder, you pick a bunch of apples, put them in your bag, and then you can’t reach anymore, or they’re no good. Why don’t you go to the next tree, right? You’re switching between these two different behaviors, right? But within each behavior, and I mean the behaviors of reaching versus walking. There’s also decisions, like which apple you reach and which tree you walk to, right? And so the hypothesis is that the which tree you walk to decision is done in the cortical system involved in locomotion, that medial system, retrosplenial and singulate motor areas. The apple reaching system is done in the premotor and parietal, peripersonal space reaching system. And the shift between the two behaviors is basal ganglia. And so in the past, we’ve always done experiments where we recorded monkeys while they were doing reaching. So we didn’t give basal ganglia, I think, the opportunity to show what it does. And so for that reason, I don’t think we’ve seen it. We didn’t see good decision-related activity in the basal ganglia. But I think in this task, which is a virtual version apple picking, we should see Basil Gangler doing that shift. When the monkey stops grabbing the apples in the tree and moves to a new tree. Now in our experiments the monkey was not actually going to be walking, at least not first experiments. He’s going to be sitting in a chair with goggles and he’s going to press pedals to go in that another tree. We’re doing that with humans now, it’s rather cool. But with monkeys it would be, we can actually in neural activity. So anyway, I’ll let you know, you’ll have to wait probably about three or four years

Asia Shcherbakova 01:28:48 before we know anything. But basically, ganglia, because there are loops essentially going through any of their lung, right? Yeah. The loops go through the cortex. Is it not possible that at every step, and because the cortex would go from concrete motor to more and more abstract actions, is it not possible that it kind of selects actions on different levels of abstraction depending on which point we’re talking about?

Paul Cisek 01:29:15 Exactly. The basal ganglia part that projects the prefrontal cortex is probably more high level goals versus more low level goals. But the thing is the projections from the basal ganglia, so you probably know this, there’s a massive convergence from cortex down to basal ganglia, something like 30,000 to one, right? And then the basal glenolia projects the thalamus and the thalamic cells that project back to cortex have these gigantic axons that just activate whole large patches of cortex. So to me it seems it doesn’t have a lot of resolution for controlling the finer points of a movement. It’s more about, I think—

Asia Shcherbakova 01:29:55 Well deciding which movement to go and then specification happens.

Paul Cisek 01:29:59 Exactly. Exactly. That’s the way I’m thinking about it now. a little bit different than what’s in that 2007 paper, although I could be wrong because again there’s evidence to contradicting this. The idea is that it’s sort of selecting out broad classes of activity which are then specified and compete in the correctly chosen idiosyncratic

Asia Shcherbakova 01:30:21 map. But ultimately, so as you I think also we write somewhere that the idea is that you cannot do completely different things at the same time.

Paul Cisek 01:30:34 Right?

Asia Shcherbakova 01:30:35 So it’s, and also as earlier we were talking about how there are more primitive parts of the nervous system, which primates have to develop something high level, right? For what about the simpler animals? The basal ganglia, do they have stronger function?

Paul Cisek 01:30:55 Yes. So if you look at outside of mammals, the projections on the basal, really the only thing, well, I’m simplifying, but in outside of mammals and birds, there is the major role of the entire forebrain really is more to modulate midbrain control systems, right? So there’s midbrain control systems that have been around for millions of years, hundreds millions of years that kind of run the show for like reptiles and fish, right? And the forebrain is mostly modulating those systems through the basal ganglia, through projections downstream from the basal ganglia. There’s not much of a closed loop. There’s not much evidence of that. And so the entire forebrain is more of a modulation of what kind of thing the midbrain should do, but the midbrain runs the show, right? And now in mammals that’s changed dramatically, although it’s still true that the midbrain does a lot of the driving show and even hindbrain and spinal cord.

Asia Shcherbakova 01:32:01 Oh, we’re not aware of it.

Paul Cisek 01:32:03 We’re not aware of it, yeah, exactly. But because the, and I actually think it’s because we became nocturnal. It’s because the mammalian branch sort of lost the battle for the diagonal world against reptiles and became nocturnal and relied less on the midbrain systems because they were they were that was the visual thing but vision is not that useful at night right so they became much more semi-sensory and proprioceptive you know whiskers and of course olfaction right and all that was going on the forebrain and so that part became more important and developed its own projections much stronger projections down downstream to control the musculature and sort of sort of as vision became, as the sort of midbrain controlled vision behavior became less important, the somatosensitional factor-driven forebrain behavior became more important and the balance shifted to the point where mammals now, even the visual control is, for many activities, is through the cortex. And then, you know, of course, once you’re there, when primates develop, right, that’s That’s what they’re working with. Primates which returned to being extremely visual didn’t do it really so much by elaborating those old mid-brand systems. They did it in cortex. The cortex of a monkey is dominated by visual information. Everything we do is so highly visual. Everything depends. All those frontoperial systems are really very much dominated by visual information.

Asia Shcherbakova 01:33:42 This is a little bit of my feedback from my field to your field and not just your work, but just overall whenever I pick up papers about this, I see there’s just so much emphasis on vision. I think it’s because just scientists, they also themselves personally rely on vision so much. Very often when I read a paper and it talks about vision, I just think, okay, like I forget about the vision. I just think about the…

Paul Cisek 01:34:10 No, you’re right. There’s a huge neglect of all the other things that are important. Yeah, and in particular proprioceptive, vestibular, and inter-acceptive. Right? Yeah, no, that’s true. I mean, not to say that the parietal, that the primate cortex isn’t dominated by vision,

Asia Shcherbakova 01:34:31 but the study of the primate cortex is even more dominated by vision. I mean the whole thing about the Dwarthels to an eventful stream, it’s all about vision. And I mean sure, yes, I mean within this framework, yes, it’s about vision, it starts in V1, but what it is fundamentally talking about is that we have a system for identifying stuff and we have a system for interacting in space with stuff. But it’s all, I feel like this framing…

Paul Cisek 01:34:57 It’s actually not about vision, it’s about action.

Asia Shcherbakova 01:35:00 It’s about action in space. although there is also you’re writing about that, that it’s not really very correct to say that the brain just constructs a representation of the outside space and makes decisions based on that, but still the brain has to represent the peripersonal space, like the wider allocentric space, it has to model it somehow, right? And all that it seems to be happening in the posterior brain, right? And it’s not so much about vision.

Paul Cisek 01:35:31

• Yeah, no, I agree. And in fact, if you, but again, I think that’s mostly the primate people that are sort of vision obsessed. You know, the, people study other mammals are well aware of the fact that a lot of that activity is driven from somatosensation, you know? If you look at a mouse brain, it’s dominated by somatosensory input, you know?

Asia Shcherbakova 01:35:52

• But the framework of the dorsal stream, dorsal stream, it goes from V1.

Paul Cisek 01:35:57

• Yeah, no, but actually, but it shouldn’t though, right? It would really is, and yeah, so this is something that I’ve come to appreciate a lot more now. So what really is, is there are different parts of the forebrain that have developed over time, way before we had advanced visual systems, right? There’s a kind of a lateral, and it’s called a pallium, right? So if you, the sort of superset, if you think about the forebrain, it’s really pallium and sub-pallium. And sub-palium, we can think of basal ganglia, but also includes parts of the amygdala and septum and other structures. But it’s essentially kind of like the basal ganglia. The pallium is kind of like the thing sitting on top of the sub-palium, and includes things like cerebral cortex, but other things as well, right? Like the hippocampus, right? Hippocampus is a medial part of the pallium, and then there’s a lateral part, which includes piriform cortex and other things. Anyway, the point is that if you look way back at evolution, there was a very old separation between the stuff that you use to move around and the stuff that you use to select what to do and what not to do. So really like dorsal ventral streams in lamprey, right? But it’s not really about vision. It includes vision, but it’s everything, right? And then those things expanded dramatically over time, but they retained that kind of system that’s kind of medial and dorsal and very spatially topographic. So Barbara Finley has sort of made the argument that in all mammals there’s a sort of dorsal medial part of the cortex, which is all about space, right? It’s very topographic and it has this spatial gradients. And then there’s a lateral, ventral lateral part, which is all about non-topographic information. And that’s essentially the dorsal ventral streams, but it’s not about vision.

Asia Shcherbakova 01:37:59 Yes, so the dorsal and ventral kind of separation, we see those most in the frontal areas and in the parietal and all the other areas. So essentially, yeah, it does seem like there are two very broad domains.

Paul Cisek 01:38:14 Yeah, and it’s very interesting if you look at it anatomically, you know. The ventral parts are connected up to very different parts of the subcortical system, like hypothalamus, whereas the other parts are not, you know, much more connected with well, the hippocampus, but other things. Anyway, the point is that this idea of dorsal and ventral streams, I think, hit upon, discovered a principle which is more general than vision. Yes. Yeah. But, I mean, and I think it’s about action, obviously. And in the case of primates as highly vision-dominated, but in the case of other non-primate mammals that are still nocturnal, it’s not vision-dominated at all. It’s actually much more somatosensory, at least in the context of action control, where the ventral stream is more olfactory than anything else. In fact, the ventral part in mammals is almost exclusively, in most mammals, almost exclusively about alfaction. That’s how they figure out what to approach and not, you know? Anyway, we’re going a bit of a topic, but anyway, getting maybe back to your specific questions, I think, you know, I mean, I don’t know what I can, if I can shed any light. I mean, just having this discussion, I guess, is informative for both of us. And what I could do, if you want, is just give you more stuff to read than I found that I feel is possibly relevant.

Asia Shcherbakova 01:39:58

• Yeah, if you feel like I was talking about something I didn’t mention, and you feel it’s relevant, yes, I would agree with you.

Paul Cisek 01:40:04

• Well, like you know, I thought the idea of the uncontrolled manifold and these ideas, I mean, I think the direct model captures that a little bit, but I’m not sure what paper I would suggest on that, though, because it’s kind of scattered. It’s an idea that’s been scattered and sort of rediscovered multiple times with different names. Like optimal feedback control is kind of a popular idea, but it captures that, but it also kind of carries along a lot of extra baggage, which I think is maybe not so good. I don’t know if there’s a good single or small set of things to read. You could maybe just look up uncontrolled manifold hypothesis. That’s what a particular group of people call it. But it’s really just the idea of how you deal with redundancy and how you make redundancy a good thing, supposedly.

Asia Shcherbakova 01:40:57 Yes, and actually the word redundancy, I sometimes use it in my slides because essentially what our system does it artificially adds so much redundancy and that redundancy as we redundantly add sensation and action that gives you much more information to choose from and that is so we reconfigure and this is how yeah but then you also reduce redundancy by telling them

Paul Cisek 01:41:24 don’t just do what’s natural do this a specific subset of the things that are yes yes you know Bernstein? Yes. Yeah, so I mean being Russian it must be Bernstein. Yes. Yeah, so I mean

Asia Shcherbakova 01:41:36 a lot of the ideas really come from him. Well, the most fundamental thing about the

Paul Cisek 01:41:44 more equivalent. Yeah, more equivalent, exactly. So a lot of that really comes from him, but it’s been sort of formalized by various groups in various ways and I think it’s very much

Asia Shcherbakova 01:41:54 related to what you’re doing. These theories they usually explain why people tend to choose the most effective way to do something, whereas we look at it from a completely different thing, we say, “Okay, look, this is actually, yeah, this is in most cases, this is the most efficient, but sometimes it creates problems, and also what if we want to change this status of how do we do that?” So I feel like the novelty is in the practical solutions how to change this.

Paul Cisek 01:42:22 Yeah, but I mean, some of them do address that. I mean, optimal feedback control is the idea that suggests that you can define the task in terms of the control variables. And then the system will find ways to use its available degrees of freedom to solve that task. And so that’s the idea of, you know, you could have your target could be anywhere here or you could even have a joint target.

Asia Shcherbakova 01:42:47 So basically it’s all about how you creatively define the task?

Paul Cisek 01:42:52 Yes, something like that. Now the specific optimal feedback control sort of way of thinking also adds to that a certain optimization procedure which I think is not what the brain does. So there’s parts of that theory I love and parts of the theory I disagree with. And it’s mostly because of the actual mathematics of the way they address the problem I think are not biologically plausible. But the way they define the problem, I think, is very informative and insightful. But again, Latash and the uncontrolled manifold theory. Anyway, that’s a useful one. The stuff about the action maps, I think. I mean, maybe I’m saying that because I really got into this idea. It’s like this idea that I didn’t agree with and then sort of it won me over and now I’m totally like a fan of it. So maybe I’m just kind of, you know, evangelizing it too much. But I think it might be useful because it’s this, it’s kind of a generalization of what I was talking about in the foreign competition, right? There I was just talking about, you know, how do you drive a hand through vision? But the idea that there is different tasks, spaces, different maps that are specific for different types of activity and then within them a competition happening.

Asia Shcherbakova 01:44:10 Yes, yes, I understand. And then in our language, we would say that to increase sense of resolution within the map. This is how it works.

Paul Cisek 01:44:19 Yes, exactly. But perhaps also force the different maps to kind of work together, right?

Asia Shcherbakova 01:44:27 Yes, yes. In ways that they’re not used to. Yes, integration of the idiosyncratic maps. Exactly. Yes. Yes. So for example, when I do lectures for like non-scientists, I’m talking about building richer action pairs and I usually show, so what I do is that I show that very often we think about perception, cognition and action like this, sensory areas and everything else is cognition, but it’s actually more like this, that we have the perception and action and cognition as the result of their interaction and what I find useful is that I’m trying to basically collapse everything into saying that look you have action maps you have possibilities of action and somewhere I had like something like this show and I’m explaining that they are kind of separate but we can teach them so I do show this picture I think. Oh yeah, there you go. You got it. Yes and then when I give concrete examples this is also another task that people struggle with so why do people struggle with this movement they don’t see the key features they don’t enough sensory map to sense the position and don’t have an action map pair to execute the movement. So this is how I’m trying to communicate this to people who

Paul Cisek 01:45:45 have no idea about the science. Yeah, that’s nice. That’s cool. Although you know I must say that for the non-scientists this is useful because it snaps them out of the thinking that comes out of the condition.

Asia Shcherbakova 01:46:00

• This slide which is how people typically think about it.

Paul Cisek 01:46:05

• Yes, exactly. But here what I would do, I would actually draw the borders differently. I actually think this thing like that is the, so including parts of prefrontal and the temporal lobe in the sort of ventral system.

Asia Shcherbakova 01:46:26

• So basically dorsal ventral.

Paul Cisek 01:46:28

• Yeah, dorsal ventral. Yeah, I think so. Because you know this border, this is recent, evolutionarily, motor cortex and somatosensory cortex are not even distinguishable on marsupials. So this is actually a relatively new distinction. If the dorsal and ventral, which I think is the much more dramatic, right?

Asia Shcherbakova 01:46:47

• I do talk about dorsal and ventral here, but—

Paul Cisek 01:46:51

• Yeah, there you go.

Asia Shcherbakova 01:46:51

• This is what class— - When deciding on one thing, you know what to highlight— - Exactly.

Paul Cisek 01:46:55 Yeah, yeah, no, no, I know, I know. What I’m saying though is that, yeah, So maybe something like this, but maybe not just about vision. So what I mean is that there’s an entire part of brain, which is more about selection, and another part, which is more about action control. At both abstract and concrete levels.

Asia Shcherbakova 01:47:15 Yes. Yeah, it’s a very good idea.

Paul Cisek 01:47:17 And then the action maps, which has become my favorite idea nowadays. Anyway, what I could do, I mean, I don’t know what I could contribute to it. I mean, I could send you a couple of papers like the Bufaki one, which I think is very nice. It’s a bit of a monster in terms of detail, that anatomy, more than you need, but it gives you a good resource. And I can give you, I mean, what I’ve been doing, the one with, Giovanni, I would have recommended it if you hadn’t already read it. The one about navigating the landscape. Yes, you did. You read a lot of your pages. Yes, you read a lot of that already. So maybe, I don’t know if the evolution ones would be useful.

Asia Shcherbakova 01:48:02 I also read the evolution ones. Those are just beautiful and it’s just, it’s really joy to read something like that because it’s just, it’s like a puzzle all coming together. That’s a really nice sense that somebody has spent so much time.

Paul Cisek 01:48:16 Well, I can tell you, it’s a joy reading all the stuff that I’m describing in that one. I mean, I’m making up a lot of stuff, a lot of it is just synthesizing a vast literature. Anyway, okay, well then you pretty much know where I stand in all these things. You pretty much understand my viewpoint.

Asia Shcherbakova 01:48:31 My question is that after listening for me for so long, does it sound like I have some blind spot or something that you completely disagree with? For me the problem is that I have very little interaction with people who actually understand this. Usually I can talk about this and I can just like lie and come make up.

Paul Cisek 01:48:53 Yeah, yeah, yeah, you have nobody that might help you steer away from the ideas.

Asia Shcherbakova 01:48:59 Yeah, so I’m absolutely alone with it. So basically that has been my like…

Paul Cisek 01:49:03 I can totally get that though because in that field people don’t really get into the neuroscience.

Asia Shcherbakova 01:49:09 They don’t sometimes even criticize me for dissecting experience too much because you’re supposed to just feel yeah

Paul Cisek 01:49:15 Yeah, well no, but I mean it’s that it doesn’t prevent you to feel Thinking about stuff. It doesn’t prevent you from feeling it I guess you’ve already told that Well, I mean I guess you know There are people who are sort of kind of like where you are in terms of wanting to do both. I mean, I can tell you, I have a student, Cesar, who is finishing his PhD with me. And what he wants to do is kind of… He wants to stay in academia, but he wants to do kind of like the science of expertise, the science of high performance, right? He himself is is sort of quite athletic and he comes from a kinesiology background. And what he’s interested in is something like understanding at a neurophysiological and a sort of biomechanical level what we need to do to be good at motor skills. So it’s kind of like what you’re doing.

Asia Shcherbakova 01:50:31 It’s very similar.

Paul Cisek 01:50:34 doing it from the from the practice of it towards the neuroscience he wants to go from the neuroscience towards the practice you know so I mean he’s he’s looking at labs to do that kind of work in not I mean I could I could find out from him a little bit what he’s found because you’re looking in collaborators

Asia Shcherbakova 01:51:00 in Montreal. I’m looking for some collaborations in just, I don’t know, just any type of exposure where I can find some application for this, but also… Well, I could certainly be a sounding

Paul Cisek 01:51:15 board for you, right? So if you just need bounce ideas off and to see if you’re missing something or not, I can help with that. But for like, sorry, I was just interrupting there

Asia Shcherbakova 01:51:27 briefly but go ahead with your say yes and I just wanted to overall to ask but do you think that I can go with these ideas as is to for example post them publicly somewhere and invite discussion because I’ve because I never have any feedback so I am never sure can’t even talk about this well have

Paul Cisek 01:51:50 Have you tried publishing it in academic journals that are sort of in between these fields?

Asia Shcherbakova 01:51:58 Well, what? I mean, there’s sports exercise. So, the thing is that when I think about how to publish what’s publishable, what’s the closest to publishable, there is an issue because I am, we’re a private company and I have an interest in the method, so I cannot really write papers about, you know, my method it’s so amazing, look at it. But there is a lot that comes from the research of the method, which I find, first of all, it’s just missing in the literature. And second, even though after spending a lot, many years reading about this, it all makes sense, it doesn’t exist anywhere unified from this perspective of what’s happening in an actual classroom environment. And also when you go to classroom environments, teachers are not aware of this. And because of that, the teaching is not effective. People suffer, they have pain. Millions, if not billions of people could benefit from our research getting public. So one idea that I had that is probably publishable is try to talk about, I usually refer to it as body awareness. I hate this term because it sounds very adjacent to all this cosmic stuff. But I was thinking about writing a paper about that. We have a framework of three types of body awareness because it’s applicable in classroom. We distinguish between spatial, proprioceptive/localized somatosensory, and interoceptive. Interoceptive awareness and interoceptive skills is more or less established right now in those therapy, adjacent circles. But there is actually some— there are some publications. But to distinguish between, for example, the proprioceptive awareness and proprioception. What’s conscious and what’s unconscious? You know, what’s known, what’s not known? My data about the muscular sensations, maybe I could get more people to answer this question.

Paul Cisek 01:53:56

• Where did you present the poster?

Asia Shcherbakova 01:53:58

• Well, so the first time I did it, it was actually at McGill University at the, but this is the second time. The second time was recently in, it’s in Kazakhstan. They just had their first conference.

Paul Cisek 01:54:12 Let me just look at something. I’m presenting at a conference that’s new to me. Maybe you know it. Scaps. Do you know this one? No, I don’t. It’s an Ottawa this fall. It’s the Canadian Society for Psychomotor Learning and Sport Psychology. That sounds like wet up your alley. So I’ve never gone to this conference. So this was last year it was in Charlotte’s, Charlotte Town, PEI. But anyway, I mean it’s, Aaron Kressman is the person who invited me. And this is, looks a lot. I mean, I don’t know that much about this field, right? So I guess I’ll learn about it. It’s in October in Ottawa. I mean, maybe this is something you should submit a poster to, you know, and meet all these people, right? I don’t think there’s anything that prevents being at a company from presenting at a conference, right? I mean, they presumably encourage you to do that.

Asia Shcherbakova 01:55:12

• Well, yeah, yeah, well, they do, but then, you know, because they have no idea what I’m talking about, the reaction that I get is people say that it’s interesting, but they can never give me any critical input, and also they usually lack the practical understanding, like for example, that why these movements are difficult, They have never experienced teaching people movement because if you just, when you start teaching people and you see how much they fail and these things, it just shifts something in your brain, just how you see more the science in general. Because there’s a very big gap that doesn’t address it at all and clinicians are not aware of it, scientists are not aware of it. Basically, nobody’s aware of it. Those who are don’t know anything about neuroscience and they don’t understand what’s novel, what’s important. So I’m in this gap.

Paul Cisek 01:56:07

• Yeah, no, yeah. No, it’s like Cesar, it’s like Cesar. He wants to become one of the people that filled that gap. I mean, it’s not exactly the same, but it’s kind of also in between these fields. Well, you know, we actually, with Giovanni, we actually published recently in a sports and exercise like special issue. So Giovanni Pazzullo is fairly well aware of that literature, much more than I am. Let me see, where would that be? Where would I have put that one? Miss Projects probably. I’m gonna have a hard time finding it. But anyway, we published something, I guess it’s on my CV somewhere. I’ll tell you what the journal name was. It was very, very theoretical neuroscience paper, right? And it was in this, I think it’s a special issue of a journal called, let’s see, Psychology of Sports and Exercise, there you go. And our paper is essentially about this kind of stuff, right? It’s like the affluence competition, hierarchical decision making, et cetera. So Psychology of Sports and Exercise. I mean, they’re interested in this kind of thing. This conference, again, this GAPS conference, might also be a place for you to certainly meet people who will understand exactly what you’re talking about. Scaps. Yeah, let me see again what they’re called. Society, Canadian Society for Psychomotor Learning and Sports Psychology. And just like scaps.org. Anyway, it’s going to be in Ottawa in October, first to third. So the fields of sports psychology, exercise psychology, and motor control and learning. I think this is where you’ll find kindred spirits here. And they’re going to understand what you’re talking about. So yeah, why don’t you come to that?

Asia Shcherbakova 01:58:25 Okay, thank you. I’ll check it out.

Paul Cisek 01:58:29 Again, I don’t know most of the people. I think, although I’ll probably realize that I do know some of them once I go and see who’s there. Because motor control people, I actually think there’s a lot of, for us, motor control theory people to learn from sports and exercise people. Because especially now that we’re interested more and more in natural behavior, so we’ve We’ve been studying monkeys doing this, but now we want to study more natural behavior, and we don’t even know how to describe it, right? But the sports people have been doing that for a long time, and maybe not in a very quantified way, but at least in some basic principles. So, it’s good for our field also, too.

Asia Shcherbakova 01:59:14 You know, another field where actually I think it’s very relevant is, we did one project in Japan with Panasonic who had, they developed a technology to track movement from camera input. Yeah, right, exactly. And we had to help them design basically demonstration for a conference. And honestly, our task was just to find a task that their system could solve, because it was so poorly calibrated for movement. So we had to test it and then find something that would actually be interesting for people who come to the conference. And well, so we managed to do that. But my feelings after all that project were that obviously looking at this technology, so much resources went to develop it, but obviously there was no person around who actually understands movement. And they could have used that technology for so many different things, but they didn’t. And they just lack experience.

Paul Cisek 02:00:16 Yeah, no, that’s true. There’s a lot of that right now. We’re going to set that up as well. So we’re going to have monkeys just running around and doing things in an open environment and track them with cameras. It’s a tough thing to do, but there’s a lot of technology

Asia Shcherbakova 02:00:30 now being developed for it. But again, because they are tracking natural movement, for example, very often the equations will be based on all this reverse kinematics because it’s much easier. You have less data and then you just do some computation. Whereas work with actually. For example, if that system would not be able to understand what’s going on in our system, because it’s not created to interpret all the degrees of freedom that are possible for human movement, because it just assumes that the movement will be natural, but sometimes to unlearn, to learn, you need to do something that is unnatural, which follows different laws and these systems

Paul Cisek 02:01:10 don’t always account for this. Well, we’re gonna be looking at some of that. I be looking at some of that because some of our, I mean I’m not saying my lab, but some of my collaborators and colleagues are doing things like learning, like how animals learn skills that are close to their repertoire but not quite, like manipulating something to open and get a nut out of it, you know, or completely novel, you know, like, force environments. So, I mean, I don’t know. I mean, it’ll be far from the richness of natural movement that other people study, but still we’ll be looking at it. We’ll be looking at it in terms of neurons, you know. Anyway, you know, I realize that it’s already noon. I gotta get going. Yes. But it was really a lot of fun talking. Thank you so much. And I guess, I mean, what I can offer you is, you know, I can be a sounding board of sorts. I mean, I don’t, I think you’re, you know, when you first wrote to me, I didn’t see the connections. I mean, I saw maybe some potential for connections, but I didn’t see the connections. Now, now I do, but I also see connections with some of these other studies. But again, you’re already, I mean, you already read Graziano, so you already know some stuff.

Asia Shcherbakova 02:02:27 Well, yeah, I read a lot of stuff trying to find connections.

Paul Cisek 02:02:31 I would say poke around in the uncontrolled manifold and optimal feedback control. So maybe look at Steve Scott. He’s in Kingston, Ontario. He’s a former supervisor of mine, actually. So he’s done a pretty nice version of optimal feedback control. He’s got a few reviews of it. He’s not the person who pioneered it, but he’s synthesized it best with actual neurobiology. I think. So it’s Steven H. Scott. And then for the uncontrolled matter fold, I really don’t know who. I really like— what’s the guy’s name? But I shouldn’t block— Gregor Schoner. S-C-H-O with an umlat, N-E-R. Gregor Schoner is really good. Very computational theoretical. And it’s very much in the way I think about things as well. So he was one of the people that you’ll see me cite in my papers, because he came up with similar ideas earlier. And Marc Latash, who’s kind of maybe a little bit too particular direction, but he’s outlined those ideas fairly well. I think Schoner would probably be a better place to see. Anyway, you can look at some of those authors. Of course, there’s a million papers you’ll discover. but just so that you’re aware of those ideas, ‘cause I think you’re already kind of thinking this way. And again, it all comes from Bernstein, so you already have that background.

Asia Shcherbakova 02:04:10

• Well, I think I’m at the same wavelength in terms of the problem, but I don’t always get the mathematics of it. And sometimes, maybe if you understand the mathematics, it can give you some intuition about how to—

Paul Cisek 02:04:26 So all of those authors are pretty good at getting the gist of the idea across. Because it’s not that complicated. It doesn’t have to be so complicated. The gist of the idea is not so hard. And I think they’ve done a pretty good job. Anyway, I mean, I can send you more stuff. You can send me an email.

Asia Shcherbakova 02:04:42 I can send you more things to read. Thank you very much. And also just the last question. Does the phrase perceptual skills make sense to you in this context? Do you think it’s the best phrase I could use? or does it sound like something else that already exists?

Paul Cisek 02:04:58

• Well, some people will feel that perception will skills, implies a lot of sort of cognitive reconstruction and conscious thinking, which you might not necessarily wanna.

Asia Shcherbakova 02:05:11

• Usually when I explain, I say that we’re talking about somatosensory discrimination skills.

Paul Cisek 02:05:16

• Yeah, yeah, yeah. So I think…

Asia Shcherbakova 02:05:20

• It’s more about, because I talk about perceptual skills

Paul Cisek 02:05:23 and movement indication. But it’s the difference between perceptual versus sensation. Perception versus sensation. Some people will say sensation is being sensitive to stuff and being able to discriminate and detect and things like that. When you say the word perceptual, it’s all that plus conscious perception.

Asia Shcherbakova 02:05:42 Yes.

Paul Cisek 02:05:43 And so it’s a question of whether you want to include the conscious perception aspect of it. Now I myself don’t care so much. I myself will be perfectly fine with both terms. I’m not gonna give you a hard time because you’re applying perception. But some people might say, might immediately think you mean very much like a reconstruction. You know, something like you’re reconstructing a world inside your head based on a few stimulus cues, right? Which you might not mean, right? So that’s the one thing I would caution. but it’s not a bad term, perceptual skill.

Asia Shcherbakova 02:06:21

• Okay.

Paul Cisek 02:06:22

• Well, maybe you, yeah. Well, I can answer discrimination people don’t like ‘cause they think you’re like racist or something. But that’s not what you mean. That’s not what you mean. You could probably just tell them, no, I’m not talking about that. And then they should be okay with it. Yeah, no, I don’t think there’s a problem with it. I think that’s a term that I’ve heard people say that.

Asia Shcherbakova 02:06:43

• I’ve seen that before, but usually again, it’s used in digital domain?

Paul Cisek 02:06:48

• Yeah, but I think it’s implied that whatever modality you’re talking about is gonna be that modality. It’s just that yeah, usually it’s gonna be. But like people talk about it in the auditory a lot. You’re ready to discriminate the location of a sound, let’s say.

Asia Shcherbakova 02:07:08

• Or also high-construts. - Inside the mood. - For example, for musicians, if somebody is completely to Indian music and I play some melody and I say, “Oh, can you hear that minus third how it works?” And what I have no idea what minus thirds are.

Paul Cisek 02:07:22 Miner 30, yeah. Yeah, exactly. Yeah. Yeah. That’s a skill, right?

Asia Shcherbakova 02:07:28 Yes. That is a skill that you need to build to… But it’s not really discrimination, it’s about learning to identify a pattern reliably. Yeah, exactly. Yeah, exactly. Yeah.

Paul Cisek 02:07:39 and sometimes categorize.

Asia Shcherbakova 02:07:41

• Yeah, so the word category I use a lot and I usually make analogies with music

Paul Cisek 02:07:47 and with language, more than exercise

Asia Shcherbakova 02:07:49 because music and language give much better material to construct an analogy, which is—

Paul Cisek 02:07:55

• But you could also discuss in terms of sort of dorsal ventral distinction. My ventral stuff tends to be more categorical.

Asia Shcherbakova 02:08:02

• Yes, well, I mean, this is what my pictures are about by talking to non-scientists of dorsal and ventral. - Well, no, no, yeah, of course not.

Paul Cisek 02:08:09 Of course not. You could tell them though, you could talk to them about the fact that some things are continuous like your position, and some things might be categorical like what type of thing you’re trying to accomplish. And you can tell them, there’s different parts of the brain that seem to do this.

Asia Shcherbakova 02:08:23 They don’t need to know which part. - But it’s true that dorsal and ventral stuff, because pedagogically speaking, there are certain things we need to be aware of ventrally and dorsally, and there are different things that we need to fix in both these domains. - Yeah. Okay, that’s useful. Thank you very much, yes. - All right, well, it was nice talking to you. - It was very nice.

Paul Cisek 02:08:46

• Anyway, so let me know if you, I mean, if you have some questions just sent to me. I’m often very slow in responding, but I do some day respond. It’s just that, you know, I usually, I just get a little bit overwhelmed. - Okay. - But anyway, but yeah, I mean, I mean, think about those, think about those.