 Hello and welcome, everyone. It is Acton Flab Livestream number 44.0. It's May 13th, 2022. Welcome to the Acton Flab. We are a participatory online lab that is communicating, learning and practicing applied active inference. You can find more information at the links on the slide. This is a recorded and an archived livestream, so please provide us with feedback so we can improve our work. All backgrounds and perspectives are welcome and will be following video etiquette for livestreams. If you want to learn more about the livestreams or other activities in the lab, head to activeinference.org. Today in 44.0, we are going to learn and discuss this awesome paper, Therapeutic Alliance as Active Inference. I'm Courtney from 2022 by Zoe McParlin, Francesco Saratelli, Carl Friston and Jorge Estves. And the video is just an introductory review slash scoping and contextualizing. We're going to have some discussions in the coming weeks on this paper with the authors and some non-authors. We're just going to be overviewing the paper, opening up some threads, getting excited about it, covering what they said so that we can jump into the .1 and .2 discussions. And let's just get right into it. I'm Daniel and I'm a researcher in California. I was very curious about touch and active inference, which is an area that hasn't been explored too much, but it's something that is so critical to our daily experience. So it sounds like a pretty important area to model or approach with active inference and I'll pass it to Ian. So thanks for joining and for all the contribution on the slides. We'd love to hear any context and welcome. Hi Daniel, thanks very much for having me in your lab. So yeah, I was drawn to this paper for several reasons. I first saw the paper by some of the authors that got me joining the dots between their work and your lab, which was that was called Osteopathy and Mental Health and then Body to Predictive Interceptive Framework. So I've got an interest in how we sense what's going on inside our body from a research point of view and a personal point of view. And then also, you know, my main, what I spend most of my time doing work wise is hands-on therapy, so manual therapy, which involves touch. When I saw the title of this paper, what we're talking about today, which is part of their series on inactivism and active inference in the therapeutic alliance. I, you know, what jumped out to me was the word synchrony in therapeutic alliance. So, you know, this is something in the therapy world, we often talk about synchrony. And I'm intrigued to see how that might fit in with re-energy minimization. Just one thought on what you said there about interception, which is something we've talked about. I know you've had many conversations and works on it is touch. It sometimes feels like it's out there, like I'm touching objects on my desk. It's how they are out there, just like vision is something out there, but then touch is actually inside the fingers. And so it's actually like it is an interception about the external world. So it's kind of like touch is something inside of us. But again, it's out, it's about external objects. And so how does the brain and body do that? And then what's the social dynamics of it? And then what are the clinical implications? I hope those are some questions we can approach with active inference and get some insight into. Yes. And the other thing that you were saying then about, and I think in the interception researchers, there is a little bit of debate over what types of touch might include as inter-receptive or inner felt. And there's low effective touch versus maybe different types of touch. But also when you were saying that, I was just thinking, when someone says I felt touched by that piece of music, the music didn't physically touch us apart from oscillating airwaves, but touching our eardrums. But again, it might be an inter-receptive experience if we were touched by some piece of music and it sent shivers up and down my spine. So yeah, and can voice maybe in therapeutic alliance vocal auditory input touch people? Cool. So we wrote out one view of a big question, which is kind of like a curiosity or an openness that might approach somebody to this paper, even without bringing active inference into it. And it says within a therapeutic alliance can touch be used to help a patient usefully update their beliefs. And then in the conclusion, the authors are where we're going to get to is touch appears to be critical in initiating the salience of synchronous relationships touch can be used to infer and predict other states of mind, which are crucial for developing dyadic and triadic relationships in general and in a clinical setting. So what do you think about that big question or what drew you to it? Yeah, on the surface of that question. So can touch be used to help a patient? So it seems in kind of the obvious answers is yes, it can. You know, you see if a child falls over and a parent comes up and it kind of soothes the child by stroking it or touching it or rubbing it better, then then we can kind of see an archetypal version of that of a caregiver helping someone who's distressed. So I'd say, you know, yeah, it looks quite obvious. But then, you know, this usefully updating someone's beliefs. I think that's where it gets maybe in dot one, dot two. We can explore that different more. And, you know, what is what is an adaptive update and what is a maladaptive update and who's who's setting the pollings and where does synchrony come in? So, yeah, loads to explore. And, you know, dyadic, they mentioned dyadic and triadic. And this therapeutic, we don't have to just think in terms of a professional therapeutic setting. So maybe in, you know, I can seek counsel from anyone. It doesn't have to be a professional practitioner in a business type setting or a clinical setting. It might be with a stranger on a bus. Yes. What are the top down social scaffolds that facilitate or enable or constrain certain kinds of touch and clinical relationships? So cool. Let's go to the aims and claims of the paper. So as we mentioned, the paper is McParlin at all in frontiers in psychology from 2022. And some of the aims, the way that the authors represented what they set out to accomplish in the paper. One of them was to present an empirical, integrative account of the bio behavioral mechanisms that underwrite therapeutic relationships through the lens of active inference. And then they're aiming to argue for the importance of therapeutic touch in establishing a therapeutic alliance and synchrony between the practitioner and the patient. And I think this will be really interesting to explore in the context of explicitly manual and hands on therapies and work, as well as areas that are perhaps not seen like at in the same way as physical therapy or osteopathy. But maybe the dentist just puts the hand somewhere in a way that is calming, for example. So even in not just the only manual areas. And then some of the claims that they make and some that will go into a lot more in the paper is they're going to argue that active inference provides insights into the why and the how of humans choosing to synchronize. They're going to claim that touch is used to establish and also to continually develop synchrony in our lived world as social beings. And those are some of the threads that they reference a lot of papers and kind of explore in the paper. So anything that you thought about just in the claims of the paper? Yeah, when I first read this, it created, you know, it seemed to remind me of what I hear in other sort of therapies. So the importance of safety, for example, in a therapeutic setting. So if a person feels safe, then they're going to be more receptive to therapeutic touch. And I'm interested to see on, you know, to read and speak to the authors maybe about what things facilitate synchrony. And the what else have we got there? So yeah, how just how the active inference framework kind of can potentially unify some of these different languages that we've got in different therapies that kind of maybe are saying very similar things. And I think we spoke a few weeks ago, the beautiful simplicity of active inference maybe helps to unify those and help us understand them better. Awesome. All right. Onto the abstract, perhaps you could read the first four lines and I'll read the second four lines. Okay. So recognizing and aligning individuals unique adaptive beliefs or priors through cooperative communication is critical to establishing a therapeutic relationship and alliance. Next bullet points using active inference. We present an empirical integrative account of the bio behavioral mechanisms that write therapeutic relationship. Third bullet points a significant mode of establishing cooperative alliances and potentially synchrony relationships is through extensive cues generated by repetitive coupling during dynamic touch. And then established models speak to the unique role of affectionate touch in developing communication into personal interactions and a wide variety of therapeutic benefits for patients of all ages, both neuro physiologically and behaviorally. The purpose of this article is to argue for the importance of therapeutic touch in establishing a therapeutic alliance and ultimately synchrony between practitioner and patient. We briefly a therapeutic alliance in pro social and clinical interactions. We then discuss how cooperative communication and mental state alignment in intentional communication are accomplished using active inference. We argue that alignment through active inference facilitates synchrony and communication. The ensuing account is extended to include the role of C tactile afference in realizing the beneficial effects of therapeutic synchrony. We conclude by proposing a method for synchronizing the effects of touch using the concept of active inference. Active inference as a descriptive integrative account and then towards the end potentially some unique predictions or implications or usages of active inference. What would be different about the therapeutic alliance if the practitioner and or the patients were to be aware of synchrony or be able to integrate some of these important qualitative ideas like consensual affection at touch with potentially some of these ideas like priors and precision and synchrony. As you're saying that I'm thinking about trauma for example which is when a prior is so given so much waiting that it affects action and perception in a really strong way. And as you said if both the patients and the clients in that alliance they're aware of it in terms of what the updating of that of those beliefs in the whole process and aligned in the direction and the process that they're wanting to use to usefully update beliefs. Then then yeah really interesting thought being aware of what your the process you're going through. Cool. So that's where we're going to be going and the roadmap is how we're going to be traveling there. The paper has three sections introduction with a bunch of subsections. The second section is mechanisms increasing saliency within synchrony and the third section is a conclusion and within this very multi part introduction section. They talk about a range of topics range from physical touch and the clinical and therapeutic setting. There's a lot of discussion on the in utero initiation of synchrony and the way in which there's this transfer of the uterine environment to the postnatal scaffolding through the post postnatal where I guess most of us are most of the time out there in society. And the way that touch and communication and synchrony and all these other topics are sitting there in the therapeutic setting and striving for integration. And so they bring a lot onto the table with this paper. So it's going to be a great discussions and important lines of research. And in this dot zero, we're going to dive into mainly the keywords and give some background. It'll be awesome to hear about the perspective that you have being a lot closer to this work than I am. And we'll just talk about the keywords and then be ready in the dot one and dot two to take it anywhere else sound good. Excellent. Okay, the keywords are active inference, therapeutic alliance, affective touch, synchrony, predictive processing, allostasis, free energy principle and perinatal care. So let's just jump right into the keywords and see where it goes. So first, therapeutic alliance. What would you say about therapeutic alliance? Okay. So while reading the text here, therapeutic alliance is a collaborative working relationship between clinician and patients and is a critical component of the person centered care because it contributes to positive clinical outcomes outcomes across multiple health care disciplines. So what I like here is that they're sort of saying that it's a critical component. And then, you know, if we break down the term therapeutic alliance, then therapeutic meaning, you know, helpful, I suppose, beneficial to the person. And then alliance, as they said it, so it's, I think of, you know, being in agreement and working together, understanding each other, which is, you know, I think sometimes when people are going to a health care provider, it's maybe the practitioner or the clinician is fixing them or doing something to them. And maybe they can be passive in receiving that, whereas I think this therapeutic alliance is coming for the person receiving the health care. And it's then the, you know, the active inference comes in. And so it's active for both parties, not just a passive thing. And I see that. Yep. So there's another text there, which kind of says that similar things. So it's the therapeutic alliance is often modeled from a work, the working alliance described by Borden in 1979, which includes the collaborative agreement on goals. And that tasks and the development of successful relationships and cooperative communication. Yeah. Yeah. Totally true about the alliance and working together as a team. And they're also focused on the dyad. But I thought of like maybe a hospital setting where there's a lot of people of many different specialties working together like truly as a big team, the family. The patient, but also all these other people who come through that space. And then just in terms of the way that this paper approached it, they reference, I guess, a early critical citation board in 1979. And then also they reference two more recent citations. And so on the 2019 paper on the left, that paper was looking at physiotherapy and just highlighted like acknowledging the individual, the giving of self and using the body as a pivot point. So seeing that alliance as kind of like a constellation that's pivoting around the patient's body. And also there being sort of a broader scale movement, which I thought was really deep. And then the Ryu 21 citation, modeling therapeutic alliance in the age of telepsychiatry. They just begin with something that maybe we've all experienced or seen occur that things are online and remotely. And that includes psychotherapy happening online. And this paper is like a call for the investigation of the causes and consequences of successful psychotherapy. Augmented with computational tools and often carried out using computational means of communication. So a little bit of a classic take on the therapeutic alliance as well as bringing it into the early 2020s phase that we're in where we're having conversations with computers, but then where's touch and this disembodiment that could maybe occur if we don't have reminders to re answer our body during these relationships. Yes. Yeah. Re-entering our body. The other thing that I was thinking of then when you were sort of reading the early reference of the citation, the acknowledging of the individual giving of the self and using body as a pivot, thinking about that. And something that's coming up later in the definitions about this expert sort of synchronizing to an expert. And maybe there's, is there any tension there between the alliance and agreeing on goals and, you know, giving oneself to an expert. So, yeah, interested to explore that later. Great. So onto synchrony. What is synchrony and how is it being used here? So the definition in the paper. Synchrony is defined as the interpersonal coordination of behavioral and neuro physiological rhythms that can aid in sensory processing, learning emotion and arousal regulation, self regulation and the establishment of communicative relationship. So another extract from the paper. Therefore, considering the therapeutic alliance principles of agreement on goals, tasks and developing strong relationships, a degree of bio behavioral synchrony will occur within a clinical settings. They seem to be saying that, you know, if it involves agreeing on where we're heading together, then synchrony has to occur. And additionally, they're sort of saying that innate, it is human nature to want to share physiological and emotional states with others through the structured pattern of communication between mother and infant. The interaction synchronize and individuals behavior, hidden states of mind and biological rhythms, resulting in an intertwined unit akin to a sophisticated waltz. And I see you've put some diagrams there from some studies involving drumming. Yeah, I was just looking to see how people were studying synchrony in the modern time where they were studying various aspects of coordinated drumming. And I also really like how the authors mentioned the sophisticated waltz or insert your own genre of dance or movement or coordination, because synchrony is not just lockstep metronome, everyone doing the same thing. Synchrony can refer to turn taking or synchrony can refer to sort of joint improvisation. And that's this notion of generalized synchrony that comes up again and again in active inference. So what does it mean for there to be synchrony in the therapeutic alliance with people who have different expectations, preferences, priors, bodies, minds, lives. So what is being synchronized that can be measured the heart rate or some other biometric features? And then what is that reflecting? And are we looking for lockstep? But then if we're not, what is this synchrony with difference that we're going to be seeking and how is this related to positive clinical outcomes? Yeah. Yes, so much. I could go down there and yeah, maybe in the dot one dot two and evolution as well. This the regimented waltz versus the more playful jamming even. Cool. Yes, we will go there. All right. So one of the ways in which synchrony is going to be engendered. And one of the ways that is highlighted in this paper is affective touch. So what is affective touch? Quoting from the paper effective touch via tactile stimulation can be viewed as an example of a particular embodied social behavior that maintains homeostasis and influences. The perception of or inference about the mental states of self and other. Moreover, it has been speculated that touch may cause a rapid reduction in the activity of prefrontal cortex before instantiating a sustained period of preemptive homeostatic regulation. There's also been suggested that the use of social touch decreases the level of effort needed to overcome stress. So here, you know, from reading that it's sort of saying that effective touch is nice, I guess, compared to maybe a punch which might have a different effect on our regulation, autonomic regulation. And so the next quote is an effective method for establishing this cooperative alliance and potential synchrony is through ostensive cues generated by repetitive coupling during physical touch. Affective touch is unique in its ability to foster communication, interaction and various therapeutic benefits. For example, in the context of perinatal care, it contributes to an infant's development, both neurologically and behaviorally. So again, you know, there's this assumption that the touch is going to have a helpful outcome for the party that's receiving it, or both maybe, as opposed to other types of touch which might be, you know, update someone's beliefs in a way that isn't a doubt. The part that really stood out to me is like social touch decreasing the level of effort needed to overcome stress. Like a pat on the back, which is just one encultured touch that maybe in a different culture or different setting wouldn't be supportive, but it's like there's effort needed to overcome some stress. And then there's a pat on the back, literally or figuratively, and then it's like, oh, it's less stressful to bring the effort to do this. And then that is not a punch, like you said, that would like to have touch. So what is happening when touch occurs in the body and in the mind? What is decreasing the level of effort and just what is touch doing in that setting? And then how is that related to synchrony? What is the belief that's being reinforced through touch? So is it, you know, something like, I'm doing okay. I'm able to cope here. I'm, you know, I am an organized being, not a, my free energy is minimizing, not max, not increasing. And then at the social level, like, I'm there for you, regardless of the outcome or something, it could be many different things, of course. And it may not even be possible to translate it into language because that's, this is its own language of touch. So what is the grammar of that language of touch? And how do we make that affective touch, which sounds so endearing and intimate, make sure that it is that way for everybody who's involved in that situation. So that it's not like, for one person, like their intention is positive in giving or receiving. But then for the other person, it's like nails on a chalkboard, you know? Cool. Okay. Well, the setting that they mention as critical for affective touch is like the in utero and also the perinatal, like post birth care. So what's going on with perinatal care? Yeah, I thought it's really interesting that so much because, you know, when I first saw this paper, I thought it was about, as you said, post post post post post natal people, adults going for some kind of touch therapy. But, you know, when you dive into the paper, there's a lot of time spent on looking at young, how young people learn. So quoting from the paper, affective touch is unique in its ability to foster communication, interaction and various therapeutic benefits. For example, in the context of perinatal care, it contributes to an infant's development, both neuro physiologically and behaviorally. So it's contributing to their development. And, you know, I care of a potter molding a piece of clay and, you know, thinking that infants are, if we think of a baby as being like a wild animal that's born and the caregivers are really molding it into being something that fits in with culture And society, sort of, and neuro physiologically and behaviorally, then how does touch contribute to that molding? Second quote, interpersonal synchrony between parent and infant is well documented through nonverbal behaviors such as touch, gaze, voice, particularly in mothers. So there they're acknowledging it's not necessarily all about touch. It's, you know, we've got these other cues, face facial expressions, tone of voice that may facilitate or be involved in synchrony between the parent and infants in shaping their development. The clay metaphor is really interesting and it makes me think about how the baby, the larva, is born physically softer. Like a lot of the bones are not converted to their final hardness and the skull has some openings and some flexures that decrease throughout life but never go away. And so it's like, that also has analogy with like a Bayesian learning perspective where in the early phases, there's like more plasticity and learning, more softness and openness of that statistical distribution. And then there's kind of a hardening or a sharpening of certain things to oversimplify that situation. And then also one relevant paper that they cite is from Sia Unica et al. And that's called the first prior from co-embodyment to co-homostasis in early life. And so that was like a very, a really rich contribution to thinking about co-embodyments. And that is also moving the discussion of the clinical setting and bridging it with this ecological and evolutionary and physiological setting. So this is a very interesting paper and something we'll talk about. So on to part two of Perinatal, what else can we share about this? Okay, so some other quotes from the paper. Often the first communication and connection with one's mother occurs immediately after birth through skin-to-skin contact. This skin-to-skin contact is recommended because touch has been shown to help reset neural oscillators and align them with their parent's oscillation patterns in the neocortex, particularly in pre-genual AAC, which I guess is the anterior cingulate cortex, all while reducing overall stress. Additionally, a caregiver's touch frequently induces a state of calm alertness, which frequently signals to the infant that the caregiver is attempting to communicate when newborn infants coordinate their limb movements to the rhythm of the adult's speech, this intention to communicate is reciprocated. A strong bond between parent and child will amplify the parasympathetic response to affectionate touch. Yeah, what I'm noticing there is that there's chosen different examples of what might be oscillating or synchronizing. So the neural oscillations seem to be something back to your question about rigid patterns. I think of neural oscillations of having relatively fixed tempos, whereas something like limb movements is not necessarily at a certain amount of hurts. So it may be more playful, but unstructured. So yeah, that's interesting. Yeah, and one interesting part here is a caregiver's touch is signaling to the infant that the caregiver is attempting to communicate. And so from our adult perspective, we might be more familiar or remember more recently, like a child tugging on our shirt with the child trying to signal that they would like to communicate. That's like an ostent of cue from an active inference perspective. It's like a signal that is directing attention, but then this is actually looking at it the other way, just the touch from the caregiver signaling to the infant who may be in a pre linguistic state that they want to communicate. And then also just this is a really transdisciplinary area to think about like the perinatal care in terms of touch, which they're looking at here, but also other mechanisms that they brought up like gaze body language. But then there's other features that aren't even explored as much in this paper's regime of attention, like the microbiome and the enculturation and so. And then it's just funny again to think about like the postnatal care does not end. It just kind of trails and develops. Yeah, you've included two really kind of big topics there. The microbiome and the appreciation now that passing on or the microbiome from mother to infant is really, really useful. And then bioelectricity is that more sort of like the Michael Levin stuff. And what I'm thinking there is that from that change from co embodiment to what was the other phrase co embodiment and something else. Yeah, what's the co homeostasis. Yeah. Yeah. Yeah. Yeah. By by electricity there and can it return during touching a therapeutic setting. Yes. Okay. On to allostasis. So allostasis, the authors define in the paper, allostasis is a process by which the brain predictively regulates the body metabolic and energy needs using an internal model of that body in the world. And so the citation there is to bear it 2017 with the work theory of constructed emotion in active inference account of interception and categorization. So that's really cited and impactful paper that defined allostasis as the regulation of the internal milieu by anticipating physiological needs and preparing to meet them before they arise. So people might be more familiar with the more popular cousin of allostasis homeostasis, which is sometimes seen as like the return to a target range. When there are deviations occurring away from that range, like when the blood sugar or the temperature gets too low, a process kicks in that brings it back up. Allostasis is highlighting the way that sometimes that kind of physiological regulation cannot wait until something has already left a controlled range. And in fact, sometimes changes need to be taken in an anticipatory way to help meet expected changes that are upcoming. And then also a kind of corollary of that is homeostasis would suggest that we would just be converging and staying within a target range. But allostatic behavior can include, for example, taking deviations out of the range again in service of staying in that racket before one goes outside into the colds. That's going to transiently be raising the body temperature in a way that wouldn't make sense unless the person was preparing to go outside. But then in the sort of integral of action through time, putting on the jacket before going out into the colds is going to keep that physiological system happy and healthy and meeting its expectations in a way that's tractable. Whereas being inside and not needing a jacket at that moment and then waiting until one was too cold and then putting it on then, that sounds like some childlike behavior perhaps in the way in which different physiological regulators work. And then just one place where we've gone into a bit more on the technical side of allostasis was in live stream number 38. Where we looked at certain cognitive and computational architectures and the way that they might evolve and the way that simpler homeostatic system could elaborate over evolutionary time to include some allostatic features. Yeah. And what I'm thinking here is you helped me when you came on the Innocence channel to think about allostasis in a kind of use this and colonies and policy choices of what the colony might do. And I was thinking back to the temperature example. Not only do we think about putting coats on and building houses and control controlling our immediate environment but you're saying collectively we have weather forecasters and weather stations to help our whole world predict into the future about how to take take action on our temperature regulation. Another thing then I'm just thinking what came to mind is we're going off the therapy. The film Don't Look Up came to mind of this impending pressures on humanity, maybe climate change or whatever and what adaptive measures are we doing collectively for the future? Yeah, like a global or a bio regional allostasis instead of waiting for some toxic chemical to exceed a threshold. What is the anticipatory action that's going to avert that from happening? How can we recognize and take effective allostatic action, not just homeostatic for all the reasons why bodies do allostasis in addition to homeostasis? And that concept of allostasis, it does rely on this generative model that the entity is casting out into the future. And very much in line with that is the framework of predictive processing and predictive coding. We explored this more quite recently in live stream number 43. And without going into all the details predictive processing brings together a few big ideas. First, it is going to hinge around the idea that what the brain or another cognitive system is doing is not simply recognition of, for example, incoming stimuli, but rather what's occurring is like a compliment down predictions. The top is just a spatial metaphor here, but the top down predictions that are being met with the bottom up sensory input. That's a predictive coding or predictive processing architecture. And that allows noisy or sparse or incomplete sensory data to kind of meet in the middle with rich generative models of the world so that action can be done effectively. Again, even when there's only partial data, like a dark room. And then also, especially when we think about action and how to include predictions about not just future stimuli, but predictions of one's future actions and the causes of one's actions and then the outcomes of one's actions in the world. This is going to return us to a core idea from active inference. And this is a quote from that paper in 43 where the minimization of a prediction error or surprise or free energy with some slight differences but more similar than not the minimization of those terms can be achieved through multiple ways. The first two are through immediate inference about the hidden states of the world, which is related to perception, as well as updating one's world model to make better predictions, which is learning. This is pointing to the relationship and the similarity between perception and learning in the Bayesian brain and active inference and related theories. And so the way in which you can reduce your free energy through learning and inference and perception and also reducing your free energy or your prediction error through action to sample sensory data in a different way as well as to potentially like change the world and modify the niche. So predictive processing brings together the anticipatory angle that we were just exploring with allostasis and connects it to specific computational and cognitive architectures that might actually be implementing that kind of an algorithm. Yeah, and where does touch fit into that? Yep. And then just one more slide on predictive processing. People can pause to read more or again check out 43. But one thing that Maria brought up and we've continued to like think about is where do we use predictive coding and where are we thinking about predictive processing. Maybe another way to think about the difference between these two for the formalisms and the implementations and predictive processing for a philosophical understanding that prediction is the basis of signal interpretation, as opposed to description and recognition being the essence of signal interpretation or or semiosis. So I think it's referring to taking the active stance, not just with respect to action but also with respect to our own sensory experience, which can feel or one can be in culture to feel like it is a receptive process. I mean, don't the photons hit the retina and don't the pebbles hit our skin and so on. Yet even that, how can we see it in a generative and in a way potentially even with agency. And so here's some books that are in the last several years that are reflecting the philosophical and the neurobiological work that's being done in this predictive area. Yeah. And yeah, I really like that difference there between coding and processing. If I'm thinking about touch and your comment earlier about one person might feel like a path on the back or is is a path on the back but for some one person, the processing might be supportive for another person, it might be like fingers down a chalkboard. And yeah, with, you know, any type of some person might, you know, most people wouldn't like a punch but there might be some people who get pleasure from from a punch. So the, yeah, the coding and the processing. Yeah. Sensing and interpreting. Great point. Okay. Finally to active inference. So the authors in this paper wrote active inference, which they use the AI acronym for active inference is an empirical integrative model that has been proposed to explain the dynamics and bio behavioral mechanisms of cooperative communication. So they cite the vassal at all paper on 2020. It's called a world unto itself human communication as active inference. And this was actually discussed way back when in live stream number three. This paper has many, many interesting figures and formalisms. The top figure is these two brains that are communicating. And it's fun because it's kind of like an equation free version of active inference. And an idea on the left side through behavior is invoking something in the communicative partner. And that is updating their ideas, their cognitive model, which is reflected through their behavior. And that is like a feedback. So one can imagine this is like a conversation between two brains and bodies. So it does bring in embodiment because there is the integration of behavior in the interface between these two communicating brains. And it also opens up the door to thinking about like, well, how do we model ideas and cognition and things that we can't directly observe? What would be different if it wasn't just behavior in the interface? What if there was biofeedback or what if there was other kinds of interfaces? So that's hinting at the flexibility of active inference and also the way in which it can hopefully be insightful qualitatively or formally. And then we won't go into it right now, but in figure one is where we start to see what that formal layer looks like. All states could be an entity and the niche, like an entity in the world, or it could be two participants in a conversation. And then what are some of the formalisms that help us model that situation and add in empirical data and ask what if questions in a really structured way? Yeah. And as I was looking as you were talking through those two brains coming together, I was thinking back to a therapeutic alliance and then thought about other types of nervous systems that might come together therapeutically. So domesticated animals, if you've got a pet dog, then we might not be able to fully communicate like you and I can with words and actions as deep in such a deep understand walkies and come to the door wagging its tail. And there's almost a therapeutic alliance if you know I'm stroking my dog and my parasympathetic nervous system is there's an agreement there between me and the dog that there's an alliance that we're going to sue each other. But we're not fully, you know, compared to a wild beast that might attack me. There's a different, you know, very different coupling coupling going on. Yep. The flexibility you mentioned. Yep. A little bit more on active inference, which will hopefully unpack more in the dot one and two. But the authors write about active inference referring to the inversion of Bayesian generative models of the sensed world. So we can talk more about the sort of recognition model and the generative model. And then they also explore in their unpacking of active inference, the ways in which perception and action are both in the common game of reducing surprise, reducing expected free energy. So lots to talk about. But the big questions are pretty much as always, what is active inference and how is active inference being applied in this paper. And then this is the realism instrumentalism question. Are the bodies doing active inference? Do they have a choice to do it or not? Or is this an approach that we're taking to model these bodies? So is active inference the territory of the clinical relationship? And or is active inference a map of the clinical relationship that's constructed about that territory? And so that's something that's always very rich to explore. Again, is it something that the bodies are doing that the physical therapist and the clients are doing? Or is this how we're modeling what they are doing? Then what are they doing? Yes. So much to possibly say about that. Yep, won't go there quite yet. Perfect. And then what would you like to add about the free energy principle and about the Bolan et al paper? Yeah, so I took a quote back from some of the authors on this paper were on the other paper in the series on inactivism and active inference. So they said the free energy principle here in free energy is defined as the difference between systems predicted states and their actual states. Thus minimizing free energy means avoiding surprise to keep within physiological bounds and the entropy of the system low. And, you know, a prerequisite is that for this notion is that different states are separated by Markov blankets, which define the boundaries of a system statistically by separating the internal states from the external states. However, active and sensory states, active states are governed by internal states, but affect external states, whereas sensory states are governed by external states, but affect internal states. Free energy is minimized either by perception, as you said on the earlier slides, by updating the prediction based on the sensation or action changing the sensation through action to match prediction. So when I when I read these last bits in thinking about the Markov blanket and touch, and you know you sort of said is the touch internal or external, and then I'm thinking about this co embodied baby inside the womb. Where's the Markov blanket there and where does the touch begin and end and, you know, touching someone physically in a therapy session or touching them with your words. And when someone becomes synchronized and starts to minimize free energy is a new Markov blanket formed between the coupled agent lots I'd like to have answers for awesome questions will go into it. So that covers many of the keywords and background concepts. Let's jump to figure one. And so the paper has just very clear and delightful figures. And this is figure one. The caption is an overview of the effects of touch and therapeutic alliance on the different networks of the brain. So here in the center is a representation of a brain. And then with the labeling of some different brain regions and networks and systems, they're referencing different empirical work over the previous decade that has been connecting different brain region function and activation to all of these functions of bodies and brains and minds. So what's one that's like interesting to you or what do you see or how do we think about this sort of representation of brain regional function while also respecting the whole system and the integrity of like the brain as a unit and the brain and the body. Yeah, I'm going to sort of pick out there the the person on the massage couch being touched and you know you've got the CT fibers, the C tactile fibers which maybe are involved in this slow affectionate touch and the word insular there. So we talked before about that picture of fascia on your your wall and you know is the connection between the C tactile fibers and the insular through the connective tissue. That's you know joining the brain a region of the brain up to the person's fingers. That's that stood out to me. Cool. Yeah, it'll be awesome like to hear from authors and others just what kind of work did draw these connections and what are the next steps for developing that connection and understanding how there's synchrony within and among brain regions and within and among brains. Anything else to add on figure one. Yeah, just that the hypothalamus words just jumped out to me there so if we're talking about sort of minimizing free energy and the autonomic nervous system. So one question that sort of came up when I first read the paper was the, what are we on what level are we agreeing where the set points are that we're synchronizing towards or staying within so the 37 degrees C is the archetypal temperature set point but not all members of, you know, a duck egg has a slightly different preferred temperature than the chicken egg as it's being incubated and at what point in our evolution did we those those birds cooperatively agree within their therapeutic alliance to read to adjust their set points and maintain that through whatever active inference. Yeah, those sorts of questions. Cool. So on this slide I picked out just a few words that were really jumping out like touch as a core term, and then the authors talked a lot about uncertainty, repetition, safety, empathy, synchrony, insalience. And I found the paper of Mason 2019 and this safety and uncertainty framework, which is kind of interesting because an active inference we often take a very formal approach to precision and certainty and confidence approaching those terms. Basically, how they're used in Bayesian statistics, like confidence is a hyper parameter, or precision is a parameter that reflects the variance of a distribution. And here's like a two by two matrix that is just saying, where are we from certainty to uncertainty and safety to unsafety. And that can be the physical and the social and the cognitive. And so how do we find ourselves in safer situations while respecting the variation among situations and where do we want to be. And something about seeing this matrix gave me some certainty and safety almost like actually it is okay to be uncertain. And if it's harmful, then unsafe is not where one might want to be most of the time. But we also cannot only stay in the pure safety area 100% of the time. So how is the therapeutic alliance and our own lives outside of the doctor's office. How are we in these quadrants and in this way. And how does this qualitative and kind of personal way of relating to safety and certainty related to perhaps more formal representations that we would bring an active inference. Brilliant. Yeah, and that, you know, those axes remind me of the valence and affect axes. So uncertainty could be like, uncertainty, certainly could be like arousal and calmness and safety and unsafe could be pleasant and unpleasant. And then, you know, it's my, as you said, it might be fine to be safe and uncertain, it might be fine to be aroused and pleasant, which is excited. But I don't want to spend too much time excited because it is energy, you know, it's from a free energy minimization, it's quite maybe costly. So yeah, and where does the therapeutic alliance and co-regulation come into all of that? What we're agreeing on is a good balance. And safety and certainty are going to be very situational and related to that experience of the patient in this setting. So how can we foster a setting where people with very different backgrounds and priors and expectations and preferences can feel and be where they want to be in that quadrant. Maybe they even have different preferences for safety or there might be different preferences for certainty itself. Okay. So adaptation to experts based on hierarchical structure and you have brought up this creation of Adam painting by Michelangelo. So what made you share this art and what did you see in the section? So when I think about touch, I started to think and I saw I really liked this section in the paper about the role of experts and I've not thought about it in what in that way in the therapeutic alliance. So they used an example which was really helpful to me of a child again going back to child and the parent is the expert and it's through touch. The child is learning how to maybe regulate itself or soothe itself. So that's one example. And then I started to think, okay, as we grow older or in different environments, who's the expert and who's regulating us. And you could say it's, you know, here's it could be a boss at work is the expert or in the practitioner patient example. It's assumed to be the clinician or the practitioner. But then I thought, okay, above that we've got governments and politicians and then I thought above that we've got these symbolic maybe or from religion. We've got the God as the expert and that's when I sort of Google this picture. And interestingly, on Wikipedia, it says that got, you know, if you look at that picture, the God touching Adam is the shape around him is looked like a brain. And it's been supposed that that's, you know, suppose the internal model of the the world's held in the brain and, you know, which which model which held with insiders. And how is that turning out to active influence. Wow. Here's all the all the brain regions in figure one. All of these other spirits and entities and then crossing the Markov blanket tantalizingly close to touch but with that interface between them. But is there a space between with the shared gaze and the shared regime of attention? So it's very powerful interpretation. We can do some more art and meme interpretation because it's just really awesome how you shared that here. There's a section in the paper on adaption of priors to a and that's going to be something we'll go more into detail in but in this paper. They wrote the canonical loop of coupled action perception cycles is the process by which diads generate and modify cycles until they achieve the end goal of aligning their mental states and communicative exchanges. This really highlights that we're not again talking about metronomes in lockstep. Synchrony can include a conversation with two people with different perspectives. So synchrony is something a bit more general or looser than just the metronome and the paper that they cite first in 2015 shows some of the interesting modeling that's been done around bird song. And so they model a bird song conversation where there's a shared narrative or they're singing from the same hymn sheet. And without going into any of the technical details, there's an expectation of how the song should sound like I know how some songs sound. And then a bird either finds itself engaging in the action of singing that part of the song. And then stops singing after its turn duration has occurred and there's some reasons why it stops singing. And then the other bird expecting the song to continue can reduce its uncertainty about the continuity of the song through two ways perception slash learning and action. And so as long as your partner is singing, you're reducing your uncertainty about the song or the conversation by listening actively. And then when they stop, you might continue to reduce your uncertainty about the conversation or the song by engaging in action. And so that's exactly what happens with this generalized synchrony and the synchrony manifold that's happening within one bird among brain regions. And then model in this paper of Friston and Frith with between two birds engaging in this bird song conversation. Yeah, and that's reminded me of something else that stood out to me in the paper about this. They mentioned a couple of times that the basic one of the basic things that is being assumed in a therapeutic alliance or hopefully is an end point is that we are all identical or we're all alike. That was mentioned a couple of times in the paper. And as you're sort of talking about the birds and their expectation that that song is something that they hear and that they're going to sing themselves. You can see maybe why music and language is helping us self evidently perpetuate our identity. But then, you know, I think I heard somewhere might not be true that bird song is soothing for humans as well. So is it helping us remember that we're part of this community ecological community that's not not identical in the species sense, but some other sort of a likeness. Or expectation. The interspecies alliance. One other section that they have in the paper is increasing saliency within synchrony. So some important questions that will discuss more. What is salience? How are salience and synchrony associated? How is salience similar or different to observation or attention? Regime of attention. And then they focus on a few different categories of mechanisms, including physiological mechanisms, social mechanisms that promote synchrony via salience and clinical mechanisms. So if it's about synchrony, again, not in the lockstep way in this generalized synchrony way, like a conversation, therapeutic alliance as an improvised controlled novelty productive conversation. How will synchrony occur? And what is the relationship of salience up to synchrony? Is it a precursor? Is it an after effect? Are they one in the same? And then what are some of the mechanisms that could be fostered to increase salience and synchrony? Or what are some of the rate limiting steps that one might look at if they wanted to promote like or heal synchrony and salience? Healing synchrony and salience, yeah. And since the free energy principle and a lot of the active inference work uses exploits mathematics, Stephen Strogatz's work on the maths of synchrony. And I don't, you know, I don't profess to understand it. But it'd be nice interesting to see if that mathematics could be integrated with the FEP maths to help understand how, like you say, synchrony in these, in something like social mechanisms might occur mathematically. So in figure two, we have pulled back from the single brain figure one, we were, it was inside out. We were in the one brain context and figure two takes us into the two brain communicating brains and bodies context and the caption is therapeutic alliance as active inference. And there's a similar aesthetic genre where now instead of the one brain with the brain regions being highlighted in different functions. Now we're seeing at the dyad with this sort of handshake or touch occurring between them. We're seeing different functions and facets of the therapeutic alliance that are labeled with a through J. So what were one of these like functions that stood out to you or you think is interesting? Yeah, I'm looking at the, you know, so the, the, it looks like someone on some crutches and then there's caregiver helping the person and then thinking back to the, you know, the alliance bits. I'm thinking what the shared goals might be there. So, you know, it's, it's the, the, the caregiver, you know, agreeing with the person on what they're going to do rehab wise. And then how might that when it comes to touch, help the person update their beliefs about how soon they can, they can heal so that it becomes a smooth the process is possible. Yeah. And what I'm really liking about this, it's bringing in some aspects of the therapeutic alliance and process that at least from what I've seen are not discussed in such a holistic way. Like F, what have I done? Like just the confusion and shock when something surprising is happening in the body and it's like, am I going to be able to use this finger again? And how could I feel selfish about having injured my finger in this way when others are in this situation? And some of these like metacognitive states like H anxiety over the injury, the bark can be worse than the bite. And then J previous priors. So that could be consciously or experienced priors or just in a broader sense, implicit statistical priors surrounding the injury, including injury beliefs, social expectations and family and injury history. And so that kind of takes the predisposition family history, epigenetic, perpetuating cycles, discussion and puts it into this prior expectation preference framework. Like what does one expect health to look like given the types of health that they saw in their family development? Yeah, in their family and maybe culturally we've got these expectations about obesity and then what we're being, what types of products are available to us, what types of foods are in our niche. And then when you get into epigenetics and all of that, then it soon becomes very interesting. Great. Well, I think the figures are very evocative and they'll be excellent for the discussants to just reflect on. One other section, just to leave a little footnote is they wrote about allostasis coordination, attention and the stag hunt. So they brought together all these very cool ideas and brought it to the setting of the stag hunt, which is a classical game theory and philosophy setting. And this is a matrix that can be read as like a payoff matrix. And there's two participants who can either choose to hunt for rabbit hair, which has a smaller possible payoff, but one can hunt it alone. However, if they both engage in the stag hunt, like the big prize, they can have a higher payoff. So if they both collaborate and work on the stag hunt, they could both get, you know, half of a stag. Whereas if they both go for the hair, then perhaps they could both be confident or sure or not sure about getting a hair, but like it's like a smaller possible reward, but it doesn't require coordination. And this is explored some of the philosophical and social implications with the citation that they provide to Skirms 2001. And this is just a very rich area for thinking about collaboration, coordination, public goods. What kind of pre-play helps people go for the stag? Again, this isn't just about meat hunting. This is like a model of generalized coordination. Like we could co-author an amazing paper, or we could have two lesser papers that might take longer. But of course, we could also engage in the collaboration and get bogged down and then we could have no paper. So how do we manage the risks and the reward of coordination and collaboration? How is that related to shared attention? And how is that related to allostasis? Where's touch? So they're really just bringing in so many rich threads. And I was not expecting the stag hunt to appear, but then applying the stag hunt to the therapeutic alliance is quite interesting. Yeah, and they're the kind of alliance that they're working together for the benefits of the alliance versus maybe giving someone some exercises to do on their own. Or pursuing self-care on their own versus care in an alliance way. I've not thought about it in that way before in terms of the group benefit of group work versus self-care on one's own. Yeah, like a regional emergency room. What does it look like for us to be cooperating as best as we can with the health needs of this area? Or are we going to be going for the stag together rather than succeeding or failing to get the rabbit alone? And then, oh yeah, go for it. No, I think it's slightly off topic. I was just thinking about the more related to the stag, how we approach agriculture at the moment and the benefit of being able to get lots of food for cheap and in abundance on our health seems immediately to be a benefit. But then for the overall health of cheap, abundant food, long-term might not be that good as opposed to going out and moving your body and foraging on your own, which takes more energy, but it's better for your health, arguably. Cool. And then just one other random note or reference to bring in. This is a paper that I worked on in 2018 with my partner Alexandra, and we were highlighting this communication and co-creation, not from the clinical therapeutic alliance setting, but from the joint drawing and joint improvisation setting. But we touched on a lot of very similar topics about consensual and positive communication and controlled novelty and the narrative trajectory. So I wondered, where's art and aesthetics and beauty and some of these other ineffables or semi-effables in the therapeutic relationship? You have a body on your wall behind you. I have a body on the wall behind me. The body is beautiful. And where does art and culture and touch and healing come together? Well, yeah, the art is the expressed, outward kind of output of some imagery or symbols that the artist has found within them. And therapeutically, you know, the use of visualization along with therapeutic touch to update someone's belief. So, you know, I use my clients sometimes ask them to depict how their injured shoulder, what image, what vision do they get when they describe their injured shoulder compared to their healthy shoulder? And, you know, one of them said that the bad shoulder feels like a rigid piece of guttering on a house that's gone braille in the sun, whereas the good shoulder feels like gooey custard. It's nice and warm. And then, you know, how can you transfer that belief about the injured shoulder into the gooey one? So, you know, you've got some potential use of art therapy or even just visualization and symbols to in combination with touch and movement and conversation. So, yeah, very awesome, awesome. And then we have the open slide for 44.1 and 0.2. We're going to continue to add questions in the coming week and people can submit questions or participate live if they'd like. I'm really interested to do some discussion of like, what are the active inference entities that we're going to model? Is it the person? Okay, then what are the sense states, the action states? What is the generative model that we're going to actually put on paper? And so, where does active inference come into play? And how are we going to take the next step? Because so many threads have been brought together in this paper, evidenced by the keywords and the bibliography and so on. And then what is the next step? Who are the next audiences to update their cognitive model to think and act differently? How does the clinical setting look different in the active hospital or in the active exercise area or whatever it happens to be? Like, what does the therapeutic alliance look like in an active inference model? And what are the next steps for those who want to be engaged in that work? And yeah, we'll just continue to explore. What are some of your any closing-ish thoughts? Yeah, wonderful paper. I'm really interested in how other... So this is written by a group of people who are interested in osteopathy. The way I see it, it could be applied to talking therapies, other hands-on therapies. So is this a useful common language to understand lots of different already successful therapeutic approaches and interventions? That's one thing. The other thing is, you know, how do we begin to use some of the jargon that we might have been using now? How do we find a way to take the active inference language to people who've never heard of active inference before? And then, you know, finally, something that came up was I was reading. It's kind of an alternative hypothesis about this, you know, assumption that the synchrony and the therapeutic alliance will always result in minimizing free energy. And, you know, I just had this thought, what if within the therapeutic alliance, a caregiver and the patient agreed that they synchronized their quantum reference frame or whatever to say, OK, we're going to choose medically assisted dying here. Is that still minimizing free energy or not? And is it helpful for the person? Is it soothing them, you know? So is it always minimizing free energy? Wow, very deep with the birth and life control and the vital nature of what the therapeutic alliance truly is. So also thanks for just this awesome discussion and all the dot zero slide making that we got to engage in. So looking forward to the coming weeks of discussion. Talk to you later. Thank you very much, Daniel. Excellent. See you later. See you next week. Bye.