Free Energy and the Self: An Ecological–Enactive Interpretation

Kiverstein, Julian

doi:10.1007/s11245-018-9561-5

Cited by 32 publications

(39 citation statements)

References 51 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the future this 652 can be unpacked further by examining the divergence between prior and posterior beliefs 653 about these policies (e.g., through inferred epistemic value). Through Landauer's principle, 654 this divergence may be equated with the associated metabolic costs of computation and the 655 conceptual notion of interoceptive self-modelling (Kiverstein, 2018;Limanowski and 656 Blankenburg, 2013; Seth and Tsakiris, 2018). 657…”

Section: Synthetic Heart-rate Variability (Hrv) and Embodied Computatmentioning

confidence: 99%

In the Body’s Eye: The Computational Anatomy of Interoceptive Inference

Allen

Levy

Parr

et al. 2019

Preprint

114

View full text Add to dashboard Cite

10 11 A growing body of evidence highlights the intricate linkage of exteroceptive perception to 12 the rhythmic activity of the visceral body. In parallel, interoceptive inference theories of 13 emotion and self-consciousness are on the rise in cognitive science. However, thus far no 14 formal theory has emerged to integrate these twin domains; instead most extant work is 15 conceptual in nature. Here, we introduce a formal model of cardiac active inference, which 16 explains how ascending cardiac signals entrain exteroceptive sensory perception and 17 confidence. Through simulated psychophysics, we reproduce the defensive startle reflex and 18 commonly reported effects linking the cardiac cycle to fear perception. We further show that 19 simulated 'interoceptive lesions' blunt fear expectations, induce psychosomatic 20 hallucinations, and exacerbate metacognitive biases. Through synthetic heart-rate 21 variability analyses, we illustrate how the balance of arousal-priors and visceral prediction 22 errors produces idiosyncratic patterns of physiological reactivity. Our model thus offers the 23 possibility to computationally phenotype disordered brain-body interaction. 24 25 26 Our focus on the multimodal integration of interoceptive and exteroceptive domains 60 is driven by the overwhelming evidence for interoception as a key modality in hedonics, 61 arousal, emotion and selfhood Apps and Tsakiris, 2014; Gallagher 62 and Allen, 2018;Seth, 2013;Seth and Friston, 2016). This is generally treated under the 63 rubric of interoceptive inference; namely, active inference in the interoceptive domain. 64There are several compelling formulations of interoceptive inference from the perspective 65 of neurophysiology, neuroanatomy and, indeed, issues of consciousness in terms of minimal 66 selfhood. However, much of this treatment rests upon a purely conceptual analysis -67 underpinned by some notion of active (Bayesian) inference about states of the world 68 (including the body). In this work, we offer a more formal (mathematical) analysis that we 69 hope will be a point of reference for both theoretical and empirical investigations. 70In brief, we constructed a (minimal) active inference architecture to simulate 71 embodied perception and concomitant arousal. Here, we focused on simulating interactions 72 between the cardiac cycle and exteroceptive perception. In principle however, our 73 simulation provides a computational proof-of-principle that can be expanded to understand 74 brain-body coupling at any physiological or behavioral timescale. Using a Markov decision 75 process formulation, we created a synthetic subject who exhibited physiological (cardio-76 acceleration) responses to arousing stimuli. Our agenda was twofold: first, to provide a 77 sufficiency proof that -in at least one example -the interaction between interoception and 78 exteroception emerges from the normative (formal) principles of active inference. 79 Furthermore, having an in silico subject at hand, means that we can simulate the effects of...

show abstract

Section: Synthetic Heart-rate Variability (Hrv) and Embodied Computatmentioning

confidence: 99%

In the Body’s Eye: The Computational Anatomy of Interoceptive Inference

Allen

Levy

Parr

et al. 2019

Preprint

114

View full text Add to dashboard Cite

show abstract

“…Active inference also provides a suitable framework for investigating the emergence of action-oriented models. Previous work has highlighted the fact that active inference is consistent with, and necessarily prescribes, frugal and parsimonious generative models, thus providing a potential bridge between 'representation-hungry' approaches to cognition espoused by classical cognitivism and the 'representation-free' approaches advocated by embodied and enactive approaches [6,12,13,64,[67][68][69][70][71][72][73][74][75].…”

Section: Active Inferencementioning

confidence: 99%

Learning action-oriented models through active inference

2020

View full text Add to dashboard Cite

Converging theories suggest that organisms learn and exploit probabilistic models of their environment. However, it remains unclear how such models can be learned in practice. The open-ended complexity of natural environments means that it is generally infeasible for organisms to model their environment comprehensively. Alternatively, action-oriented models attempt to encode a parsimonious representation of adaptive agent-environment interactions. One approach to learning action-oriented models is to learn online in the presence of goal-directed behaviours. This constrains an agent to behaviourally relevant trajectories, reducing the diversity of the data a model need account for. Unfortunately, this approach can cause models to prematurely converge to sub-optimal solutions, through a process we refer to as a bad-bootstrap. Here, we exploit the normative framework of active inference to show that efficient action-oriented models can be learned by balancing goal-oriented and epistemic (information-seeking) behaviours in a principled manner. We illustrate our approach using a simple agent-based model of bacterial chemotaxis. We first demonstrate that learning via goal-directed behaviour indeed constrains models to behaviorally relevant aspects of the environment, but that this approach is prone to sub-optimal convergence. We then demonstrate that epistemic behaviours facilitate the construction of accurate and comprehensive models, but that these models are not tailored to any specific behavioural niche and are therefore less efficient in their use of data. Finally, we show that active inference agents learn models that are parsimonious, tailored to action, and which avoid bad bootstraps and sub-optimal convergence. Critically, our results indicate that models learned through active inference can support adaptive behaviour in spite of, and indeed because of, their departure from veridical representations of the environment. Our approach provides a principled method for learning adaptive models from limited interactions with an environment, highlighting a route to sample efficient learning algorithms.

show abstract

“…This provides an interesting perspective on the connections between frontal and posterior cortices, as the former houses representations of controllable variables, while the latter receives data about their sensory consequences (Shulman et al, 2009 ; Szczepanski et al, 2013 ; Limanowski and Blankenburg, 2018 ). Descending connections from frontal to parietal areas can then be thought of as predictions about the sensory input expected contingent upon a given action (Zimmermann and Lappe, 2016 ), endorsing an enactive perspective (Bruineberg et al, 2016 ; Kiverstein, 2018 ) on perceptual inference. In the context of the visual system, this implies visual space might be represented in terms of saccadic sensorimotor contingencies [i.e., “what I would see if I were to look there” (Parr and Friston, 2017a )].…”

Section: Perceptual Inferencementioning

confidence: 99%

The Anatomy of Inference: Generative Models and Brain Structure

Parr

Friston

2018

Front. Comput. Neurosci.

159

142

View full text Add to dashboard Cite

To infer the causes of its sensations, the brain must call on a generative (predictive) model. This necessitates passing local messages between populations of neurons to update beliefs about hidden variables in the world beyond its sensory samples. It also entails inferences about how we will act. Active inference is a principled framework that frames perception and action as approximate Bayesian inference. This has been successful in accounting for a wide range of physiological and behavioral phenomena. Recently, a process theory has emerged that attempts to relate inferences to their neurobiological substrates. In this paper, we review and develop the anatomical aspects of this process theory. We argue that the form of the generative models required for inference constrains the way in which brain regions connect to one another. Specifically, neuronal populations representing beliefs about a variable must receive input from populations representing the Markov blanket of that variable. We illustrate this idea in four different domains: perception, planning, attention, and movement. In doing so, we attempt to show how appealing to generative models enables us to account for anatomical brain architectures. Ultimately, committing to an anatomical theory of inference ensures we can form empirical hypotheses that can be tested using neuroimaging, neuropsychological, and electrophysiological experiments.

show abstract

Free Energy and the Self: An Ecological–Enactive Interpretation

Cited by 32 publications

References 51 publications

In the Body’s Eye: The Computational Anatomy of Interoceptive Inference

In the Body’s Eye: The Computational Anatomy of Interoceptive Inference

Learning action-oriented models through active inference

The Anatomy of Inference: Generative Models and Brain Structure

Contact Info

Product

Resources

About