With or without you: predictive coding and Bayesian inference in the brain

Some recent philosophical discussions consider whether the brain is best understood as an open or closed system. This issue has major epistemic consequences akin to the scepticism engendered by the famous Cartesian demon. Specifically, one and the same empirical theory of brain function, predictive coding, entailing a prediction model of brain, have been associated with contradictory views of the brain as either open (Clark, 2012, 2013) or closed (Hohwy, 2013, 2014). Based on recent empirical evidence, the present paper argues that contrary to appearances, these views of the brain are compatible with one another. I suggest that there are two main forms of neural activity in the brain, one of which can be characterized as open, and the other as closed. Stimulus-induced activity, because it relies on predictive coding is indeed closed to the world, which entails that in certain respects, the brain is an inferentially secluded and self-evidencing system. In contrast, the brain’s resting state or spontaneous activity is best taken as open because it is a world-evidencing system that allows for the brain’s neural activity to align with the statistically-based spatiotemporal structure of objects and events in the world. This model requires an important caveat, however. Due to its statistically-based nature, the resting state’s alignment to the world comes in degrees. In extreme cases, the degree of alignment can be extremely low, resulting in a resting state that is barely if at all aligned to the world. This is for instance the case in schizophrenia. Clinical symptoms such as delusions and hallucinations in schizophrenics are indicative of the fundamental delicateness of the alignment between the brain’s resting-state and the world’s phenomena. Nevertheless, I argue that so long as we are dealing with a well-functioning brain, the more dire epistemic implications of predictive coding can be forestalled. That the brain is in part a self-evidencing system does not yield any generalizable reason to worry that human cognition is out of step with the real world. Instead, the brain is aligned to the world accounting for “world-brain relation” that mitigates sceptistic worries.

Download Full-text

Backwards is the way forward: Feedback in the cortical hierarchy predicts the expected future

Behavioral and Brain Sciences ◽

10.1017/s0140525x12002361 ◽

2013 ◽

Vol 36 (3) ◽

pp. 221-221 ◽

Cited By ~ 5

Author(s):

Lars Muckli ◽

Lucy S. Petro ◽

Fraser W. Smith

Keyword(s):

Predictive Coding ◽

Internal Models ◽

Prediction Errors ◽

Conceptual Level ◽

Cortical Hierarchy ◽

The Brain ◽

Descriptive Level ◽

Perception Action ◽

The Way

AbstractClark offers a powerful description of the brain as a prediction machine, which offers progress on two distinct levels. First, on an abstract conceptual level, it provides a unifying framework for perception, action, and cognition (including subdivisions such as attention, expectation, and imagination). Second, hierarchical prediction offers progress on a concrete descriptive level for testing and constraining conceptual elements and mechanisms of predictive coding models (estimation of predictions, prediction errors, and internal models).

Download Full-text

Is predictability salient? A study of attentional capture by auditory patterns

Philosophical Transactions of the Royal Society B Biological Sciences ◽

10.1098/rstb.2016.0105 ◽

2017 ◽

Vol 372 (1714) ◽

pp. 20160105 ◽

Cited By ~ 50

Author(s):

Rosy Southwell ◽

Anna Baumann ◽

Cécile Gal ◽

Nicolas Barascud ◽

Karl Friston ◽

...

Keyword(s):

Attentional Capture ◽

Predictive Coding ◽

Visual Scene ◽

Scene Analysis ◽

Neural Responses ◽

Eeg Data ◽

Visual Domain ◽

The Brain ◽

Repeating Patterns

In this series of behavioural and electroencephalography (EEG) experiments, we investigate the extent to which repeating patterns of sounds capture attention. Work in the visual domain has revealed attentional capture by statistically predictable stimuli, consistent with predictive coding accounts which suggest that attention is drawn to sensory regularities. Here, stimuli comprised rapid sequences of tone pips, arranged in regular (REG) or random (RAND) patterns. EEG data demonstrate that the brain rapidly recognizes predictable patterns manifested as a rapid increase in responses to REG relative to RAND sequences. This increase is reminiscent of the increase in gain on neural responses to attended stimuli often seen in the neuroimaging literature, and thus consistent with the hypothesis that predictable sequences draw attention. To study potential attentional capture by auditory regularities, we used REG and RAND sequences in two different behavioural tasks designed to reveal effects of attentional capture by regularity. Overall, the pattern of results suggests that regularity does not capture attention. This article is part of the themed issue ‘Auditory and visual scene analysis’.

Download Full-text

Changepoint detection versus reinforcement learning: Separable neural substrates approximate different forms of Bayesian inference

10.1101/591818 ◽

2019 ◽

Author(s):

Wolfgang M. Pauli ◽

Matt Jones

Keyword(s):

Bayesian Inference ◽

Neural Substrates ◽

Statistical Structure ◽

Statistical Process ◽

Future Events ◽

Reward Contingencies ◽

Foraging Task ◽

The Brain ◽

Adaptive Foraging ◽

Previous Learning

AbstractAdaptive behavior in even the simplest decision-making tasks requires predicting future events in an environment that is generally nonstationary. As an inductive problem, this prediction requires a commitment to the statistical process underlying environmental change. This challenge can be formalized in a Bayesian framework as a question of choosing a generative model for the task dynamics. Previous learning models assume, implicitly or explicitly, that nonstationarity follows either a continuous diffusion process or a discrete changepoint process. Each approach is slow to adapt when its assumptions are violated. A new mixture of Bayesian experts framework proposes separable brain systems approximating inference under different assumptions regarding the statistical structure of the environment. This model explains data from a laboratory foraging task, in which rats experienced a change in reward contingencies after pharmacological disruption of dorsolateral (DLS) or dorsomedial striatum (DMS). The data and model suggest DLS learns under a diffusion prior whereas DMS learns under a changepoint prior. The combination of these two systems offers a new explanation for how the brain handles inference in an uncertain environment.One Sentence SummaryAdaptive foraging behavior can be explained by separable brain systems approximating Bayesian inference under different assumptions about dynamics of the environment.

Download Full-text

The cutaneous borders of interoception: Active and social inference of pain and pleasure on the skin

10.1093/oso/9780198811930.003.0006 ◽

2018 ◽

Cited By ~ 2

Author(s):

Mariana von Mohr ◽

Aikaterini Fotopoulou

Keyword(s):

Social Factors ◽

Predictive Coding ◽

Active Inference ◽

Top Down ◽

Bottom Up ◽

Social Inference ◽

Brain Levels ◽

The Brain

Pain and pleasant touch have been recently classified as interoceptive modalities. This reclassification lies at the heart of long-standing debates questioning whether these modalities should be defined as sensations on their basis of neurophysiological specificity at the periphery or as homeostatic emotions on the basis of top-down convergence and modulation at the spinal and brain levels. Here, we outline the literature on the peripheral and central neurophysiology of pain and pleasant touch. We next recast this literature within a recent Bayesian predictive coding framework, namely active inference. This recasting puts forward a unifying model of bottom-up and top-down determinants of pain and pleasant touch and the role of social factors in modulating the salience of peripheral signals reaching the brain.

Download Full-text

Information theoretic evidence for predictive coding in the face processing system

10.1101/089300 ◽

2016 ◽

Author(s):

Alla Brodski-Guerniero ◽

Georg-Friedrich Paasch ◽

Patricia Wollstadt ◽

Ipek Özdemir ◽

Joseph T. Lizier ◽

...

Keyword(s):

Prior Knowledge ◽

Human Subjects ◽

Brain Area ◽

Predictive Coding ◽

Sensory Information ◽

Low Frequency ◽

Processing System ◽

Information Storage ◽

Beta Band ◽

The Brain

AbstractPredictive coding suggests that the brain infers the causes of its sensations by combining sensory evidence with internal predictions based on available prior knowledge. However, the neurophysiological correlates of (pre-)activated prior knowledge serving these predictions are still unknown. Based on the idea that such pre-activated prior knowledge must be maintained until needed we measured the amount of maintained information in neural signals via the active information storage (AIS) measure. AIS was calculated on whole-brain beamformer-reconstructed source time-courses from magnetoencephalography (MEG) recordings of 52 human subjects during the baseline of a Mooney face/house detection task. Pre-activation of prior knowledge for faces showed as alpha- and beta-band related AIS increases in content specific areas; these AIS increases were behaviourally relevant in brain area FFA. Further, AIS allowed decoding of the cued category on a trial-by-trial basis. Moreover, top-down transfer of predictions estimated by transfer entropy was associated with beta frequencies. Our results support accounts that activated prior knowledge and the corresponding predictions are signalled in low-frequency activity (<30 Hz).Significance statementOur perception is not only determined by the information our eyes/retina and other sensory organs receive from the outside world, but strongly depends also on information already present in our brains like prior knowledge about specific situations or objects. A currently popular theory in neuroscience, predictive coding theory, suggests that this prior knowledge is used by the brain to form internal predictions about upcoming sensory information. However, neurophysiological evidence for this hypothesis is rare – mostly because this kind of evidence requires making strong a-priori assumptions about the specific predictions the brain makes and the brain areas involved. Using a novel, assumption-free approach we find that face-related prior knowledge and the derived predictions are represented and transferred in low-frequency brain activity.

Download Full-text

Attention Promotes the Neural Encoding of Prediction Errors

10.1101/522185 ◽

2019 ◽

Cited By ~ 1

Author(s):

Cooper A. Smout ◽

Matthew F. Tang ◽

Marta I. Garrido ◽

Jason B. Mattingley

Keyword(s):

Information Processing ◽

Coding Theory ◽

Predictive Coding ◽

Sensory Information ◽

Neural Response ◽

Prediction Errors ◽

Feature Information ◽

The Difference ◽

Stimulus Features ◽

The Brain

AbstractThe human brain is thought to optimise the encoding of incoming sensory information through two principal mechanisms: prediction uses stored information to guide the interpretation of forthcoming sensory events, and attention prioritizes these events according to their behavioural relevance. Despite the ubiquitous contributions of attention and prediction to various aspects of perception and cognition, it remains unknown how they interact to modulate information processing in the brain. A recent extension of predictive coding theory suggests that attention optimises the expected precision of predictions by modulating the synaptic gain of prediction error units. Since prediction errors code for the difference between predictions and sensory signals, this model would suggest that attention increases the selectivity for mismatch information in the neural response to a surprising stimulus. Alternative predictive coding models proposes that attention increases the activity of prediction (or ‘representation’) neurons, and would therefore suggest that attention and prediction synergistically modulate selectivity for feature information in the brain. Here we applied multivariate forward encoding techniques to neural activity recorded via electroencephalography (EEG) as human observers performed a simple visual task, to test for the effect of attention on both mismatch and feature information in the neural response to surprising stimuli. Participants attended or ignored a periodic stream of gratings, the orientations of which could be either predictable, surprising, or unpredictable. We found that surprising stimuli evoked neural responses that were encoded according to the difference between predicted and observed stimulus features, and that attention facilitated the encoding of this type of information in the brain. These findings advance our understanding of how attention and prediction modulate information processing in the brain, and support the theory that attention optimises precision expectations during hierarchical inference by increasing the gain of prediction errors.

Download Full-text

Predictive coding as a unifying principle for explaining a broad range of brightness phenomena

10.1101/2020.04.23.057620 ◽

2020 ◽

Author(s):

Alejandro Lerer ◽

Hans Supèr ◽

Matthias S.Keil

Keyword(s):

Visual System ◽

Computational Model ◽

Predictive Coding ◽

Gain Control ◽

Physical World ◽

Input Image ◽

Visual Illusions ◽

Contrast Effects ◽

Dynamic Filtering ◽

The Brain

AbstractThe visual system is highly sensitive to spatial context for encoding luminance patterns. Context sensitivity inspired the proposal of many neural mechanisms for explaining the perception of luminance (brightness). Here we propose a novel computational model for estimating the brightness of many visual illusions. We hypothesize that many aspects of brightness can be explained by a predictive coding mechanism, which reduces the redundancy in edge representations on the one hand, while non-redundant activity is enhanced on the other (response equalization). Response equalization is implemented with a dynamic filtering process, which (dynamically) adapts to each input image. Dynamic filtering is applied to the responses of complex cells in order to build a gain control map. The gain control map then acts on simple cell responses before they are used to create a brightness map via activity propagation. Our approach is successful in predicting many challenging visual illusions, including contrast effects, assimilation, and reverse contrast.Author summaryWe hardly notice that what we see is often different from the physical world “outside” of the brain. This means that the visual experience that the brain actively constructs may be different from the actual physical properties of objects in the world. In this work, we propose a hypothesis about how the visual system of the brain may construct a representation for achromatic images. Since this process is not unambiguous, sometimes we notice “errors” in our perception, which cause visual illusions. The challenge for theorists, therefore, is to propose computational principles that recreate a large number of visual illusions and to explain why they occur. Notably, our proposed mechanism explains a broader set of visual illusions than any previously published proposal. We achieved this by trying to suppress predictable information. For example, if an image contained repetitive structures, then these structures are predictable and would be suppressed. In this way, non-predictable structures stand out. Predictive coding mechanisms act as early as in the retina (which enhances luminance changes but suppresses uniform regions of luminance), and our computational model holds that this principle also acts at the next stage in the visual system, where representations of perceived luminance (brightness) are created.

Download Full-text

Strengthened beliefs under psychedelics (SEBUS)? A Commentary on “REBUS and the Anarchic Brain: Toward a Unified Model of the Brain Action of Psychedelics”

10.31234/osf.io/zqh4b ◽

2020 ◽

Author(s):

Adam Safron

Keyword(s):

Free Energy ◽

Conscious Experience ◽

Predictive Coding ◽

Alternative Account ◽

Levels Of Abstraction ◽

Energy Principle ◽

Free Energy Principle ◽

Free Energy Landscapes ◽

The Mind ◽

The Brain

In introducing a model of “relaxed beliefs under psychedelics” (REBUS), Carhart-Harris and Friston (2019) have presented a compelling account of the effects of psychedelics on brain and mind. This model is contextualized within the Free Energy Principle (Friston et al., 2006; Friston, 2010), which may represent the first unified paradigm in the mind and life sciences. By this view, mental systems adaptively regulate their actions and interactions with the world via predictive models, whose dynamics are governed by a singular objective of minimizing prediction-error, or “free energy.” According to REBUS, psychedelics flatten the depth of free energy landscapes, or the differential attracting forces associated with various (Bayesian) beliefs, so promoting flexibility in inference and learning. Here, I would like to propose an alternative account of the effects of psychedelics that is in many ways compatible with REBUS, albeit with some important differences. Based on considerations of the distributions of 5-HT2a receptors within cortical laminae and canonical microcircuits for predictive coding, I propose that 5-HT2a agonism may also involve a strengthening of beliefs, particularly at intermediate levels of abstraction associated with conscious experience (Safron, 2020).

Download Full-text

Computational Approaches to Spatial Orientation: From Transfer Functions to Dynamic Bayesian Inference

Journal of Neurophysiology ◽

10.1152/jn.90677.2008 ◽

2008 ◽

Vol 100 (6) ◽

pp. 2981-2996 ◽

Cited By ~ 65

Author(s):

Paul R. MacNeilage ◽

Narayan Ganesan ◽

Dora E. Angelaki

Keyword(s):

Bayesian Inference ◽

Spatial Orientation ◽

Particle Filtering ◽

Transfer Functions ◽

Linear Acceleration ◽

Frequency Domain Analysis ◽

Internal Models ◽

Velocity Storage ◽

Computational Approaches ◽

The Brain

Spatial orientation is the sense of body orientation and self-motion relative to the stationary environment, fundamental to normal waking behavior and control of everyday motor actions including eye movements, postural control, and locomotion. The brain achieves spatial orientation by integrating visual, vestibular, and somatosensory signals. Over the past years, considerable progress has been made toward understanding how these signals are processed by the brain using multiple computational approaches that include frequency domain analysis, the concept of internal models, observer theory, Bayesian theory, and Kalman filtering. Here we put these approaches in context by examining the specific questions that can be addressed by each technique and some of the scientific insights that have resulted. We conclude with a recent application of particle filtering, a probabilistic simulation technique that aims to generate the most likely state estimates by incorporating internal models of sensor dynamics and physical laws and noise associated with sensory processing as well as prior knowledge or experience. In this framework, priors for low angular velocity and linear acceleration can explain the phenomena of velocity storage and frequency segregation, both of which have been modeled previously using arbitrary low-pass filtering. How Kalman and particle filters may be implemented by the brain is an emerging field. Unlike past neurophysiological research that has aimed to characterize mean responses of single neurons, investigations of dynamic Bayesian inference should attempt to characterize population activities that constitute probabilistic representations of sensory and prior information.

Download Full-text