How much about our interaction with – and experience of – our world can be deduced from basic principles? This talk reviews recent attempts to understand the self-organised behaviour of embodied agents, like ourselves, as satisfying basic imperatives for sustained exchanges with the environment. In brief, one simple driving force appears to explain many aspects of action and perception. This driving force is the minimisation of surprise or prediction error. In the context of perception, this corresponds to Bayes-optimal predictive coding that suppresses exteroceptive prediction errors. In the context of action, motor reflexes can be seen as suppressing proprioceptive prediction errors. We will look at some of the phenomena that emerge from this scheme, such as hierarchical message passing in the brain and the ensuing conscious inference.. Free energy and consciousness Karl Friston, University College London

Overview The statistics of life Markov blankets and ergodic systems The anatomy of inference graphical models and predictive coding Action and perception inference and consciousness

“How can the events in space and time which take place within the spatial boundary of a living organism be accounted for by physics and chemistry?” (Erwin Schrödinger 1943) The Markov blanket as a statistical boundary (parents, children and parents of children) Internal states External states Sensory states Active states

The Markov blanket in biotic systems Active states External states Internal states Sensory states

The Fokker-Planck equation And its solution in terms of curl-free and divergence-free components lemma : any (ergodic random) dynamical system ( m ) that possesses a Markov blanket will appear to actively maintain its structural and dynamical integrity

But what about the Markov blanket? Reinforcement learning, optimal control and expected utility theory Information theory and minimum redundancy Self-organisation, cybernetics and homoeostasis Bayesian brain, active inference and predictive coding Value Surprise Entropy Model evidence Pavlov Ashby Helmholtz Barlow

res extensa (extensive flow) res cogitans (beliefs) Belief productionFree energy functional “I am [ergodic] therefore I think”

Overview The statistics of life Markov blankets and ergodic systems The anatomy of inference graphical models and predictive coding Action and perception inference and consciousness

“Objects are always imagined as being present in the field of vision as would have to be there in order to produce the same impression on the nervous mechanism” - von Helmholtz Richard Gregory Hermann von Helmholtz Impressions on the Markov blanket… Plato: The Republic (514a-520a)

Bayesian filtering and predictive coding prediction update prediction error

Making our own sensations Changing sensations sensations – predictions Prediction error Changing predictions Action Perception

Descending predictions Descending predictions Ascending prediction errors A simple hierarchy whatwhere Sensory fluctuations Hierarchical generative models

aminergic cholinergic occipital cortex geniculate visual cortex retinal input pons Oculomotor signals Prediction error (superficial pyramidal cells) Conditional predictions (deep pyramidal cells) Conditional precision (neuromodulatory cells) Hierarchical message passing in the brain Top-down or backward predictions Bottom-up or forward prediction error reflex arc

Overview The statistics of life Markov blankets and ergodic systems The anatomy of inference graphical models and predictive coding Action and perception inference and consciousness

res extensa (extensive flow) res cogitans (beliefs) Belief productionFree energy functional

Sensorium Hidden causes Posterior expectations Inference (prediction errors) Predictions Interoception Exteroception Proprioception Visual Auditory Motor Articulation Bayesian belief propagation Hierarchical (active) inference Sensorimotor (conceptual) expectations (Access consciousness) Sensory (perceptual) expectations (Phenomenal consciousness) – no direct access to active expression

Biological agents minimize their average surprise (entropy) They minimize surprise by suppressing prediction error Prediction error can be reduced by changing predictions (perception) Prediction error can be reduced by changing sensations (action) Perception entails recurrent message passing to optimize predictions Action makes predictions come true (and minimizes surprise)

Thank you And thanks to collaborators: Rick Adams Andre Bastos Sven Bestmann Harriet Brown Jean Daunizeau Mark Edwards Xiaosi Gu Lee Harrison Stefan Kiebel James Kilner Jérémie Mattout Rosalyn Moran Will Penny Lisa Quattrocki Knight Klaas Stephan And colleagues: Andy Clark Peter Dayan Jörn Diedrichsen Paul Fletcher Pascal Fries Geoffrey Hinton James Hopkins Jakob Hohwy Henry Kennedy Paul Verschure Florentin Wörgötter And many others

19 Quantities encoded by a generative model of sensory input Switching variables that determine the presence of a connection or causal relation among states of the world: Encoded by the presence or absence of connections in the brain that are optimised by synaptic homoeostasis Parameters of functions that model dependencies among states: Encoded by the synaptic efficacy of connections in the brain that are optimised by associative or experience dependent plasticity The precision or inverse amplitude of random fluctuations on states: Encoded by the synaptic gain of connections in the brain that are controlled by neuromodulatory systems States of the world generating sensory inputs: Encoded by the neuronal activity that is optimised during perceptual synthesis The brain stem controls the access to the world and predictions about its own activity (see Figures 1 and 2) Sensory consequences of changing states in the world: that are sampled by sensory receptors

sleep wake Action and behaviour Perceptual inference Attentional modulation Synaptic plasticity Synaptic regression Dreaming Attentional modulation Synaptic plasticity Synaptic regression Rise in (aminergic) sensory precision Fall in (aminergic) sensory precision Sensory surprise is enabled leading to perception and action Sensory gating takes sensation off-line and suppresses sensory surprise The sleep wake cycle _ + _ +

21 aminergic cholinergic occipital cortex geniculate pons visual cortex Retinal input aminergic cholinergic occipital cortex geniculate pons visual cortex Retinal input Saccadic predictions Proprioceptive consequences Visual consequences Prediction error Changes in retinal inputOculomotor commands Corollary dischargePredictive control Wake – perception with visual inputREM sleep – perception without vision Oculomotor commands reafference Saccadic predictions Proprioceptive consequences Visual consequences Prediction error Oculomotor commands Corollary dischargePredictive control reafference