Arousal coherence, uncertainty, and well-being: an active inference account

Hello and welcome. This is Active Inference Guest Stream 81.1. It is April 23rd or 24th, depending on where you are. And we're here with Hannah Biddell discussing Feeling Our Way Through Uncertainty. Thank you for joining Hannah. We'll have a presentation followed by a discussion section. So looking forward to anyone's questions in the live chat and thanks again for joining and looking forward to it. Thanks so much, Daniel. I'm really looking forward to it too. Like I said before, I've been a big fan of the YouTube channel for a long time. So it's really nice to be here. So today I'm going to talk about some of my work in my PhD, which is about arousal coherence, which will make more sense in a moment. But the basic topic is how we might feel our way through uncertainty through the link between affect and interoception. So gives me one minute. All right. So I'm going to be talking about this paper that I wrote with Mark Sohn's, Aline Slagda and Ruben Laucanen, both in Neuroscience of Consciousness recently. And we'll talk about it kind of step by step. I will note at the get-go that there's no formal modeling yet for this concept. So this talk and the paper is basically a series of claims, which I kind of back up with other people's experiments, models and simulations. So I'm just piecing together some prior ideas, basically, and looking at where those might lead. In particular, interceptive self-inference, deep affective inference, and, excuse me, affect in terms of precision optimization. And if those aren't clear yet, they might be soon, hopefully. So these are the basic claims that I'm going to make. And it's going to be quite a quick talk because there's only five. So firstly, that your autonomic arousal states track with changes and uncertainty, and that means multiple things. And I'll break down some of those in the next slide. That accurately inferring autonomic arousal is important for the way that our brains estimate uncertainty. And that affective arousal might optimize the way that how confident we are in autonomic prediction errors. So it's going to affective arousal, and that affective arousal will also impact potentially the accuracy of our interoceptive autonomic inferences. Which will lead me to the kind of point of the paper, which is arousal coherence. So that is, you can kind of see here, this is a very simple depiction, oversimplified depiction of what arousal coherence would look like. So we have changes in autonomic arousal across time, increases and decreases. Coherent affective arousal would be changing with the autonomic arousal, and discoherent affective arousal would be dissociated from changes in autonomic arousal. Like, it is not this simple, obviously, um, those layers of depth to change. So the simple time axis is an oversimplification, but this is a basic idea of arousal coherence. And I think and have argued in this paper, that it impacts and reflects our capacity to adapt to changes in uncertainty, and of multiple different kinds. And the final point is meta awareness of arousal. So knowing your arousal, or having a higher order representation of your arousal, might enhance arousal coherence. And I'll talk about some of the benefits associated with higher arousal coherence too. Just quickly to mention the way that I am conceiving affective arousal is a lower order prediction. So it's not the same as emotion, like stress, anxiety, excitement, but a high arousal is common across those kinds of emotions. Whereas a relaxing state, for instance, would be a low arousal. And if we're talking about autonomic arousal, um, I'd like to clarify that I'm not being very specific here about whether we're talking about sympathetic arousal, sympathetic activation, or parasympathetic, or the balance between those. Just broadly, increases of autonomic arousal or decreases, um, conceived through things like heart rate increase or changes in pupil dilation. Um, okay. So, um, I have not included as yet a, uh, a…