Canonical microcircuits for predictive coding.

By André Moraes Bastos, W. Martin Usrey, Rick A Adams, George R. Mangun, Pascal Fries, Karl J. Friston. 2012

2,289 citations recorded by the source corpus.

Summary This review considers the influential notion of a canonical (cortical) microcircuit in light of recent theories about neuronal processing. Specifically, we conciliate quantitative studies of microcircuitry and the functional logic of neuronal computations. We revisit the established idea that message passing among hierarchical cortical areas implements a form of Bayesian inference – paying careful attention to the implications for intrinsic connections among neuronal populations. By deriving canonical forms for these computations, one can associate specific neuronal populations with specific computational roles. This analysis discloses a remarkable correspondence between the microcircuitry of the cortical column and the connectivity implied by predictive coding. Furthermore, it provides some intuitive insights into the functional asymmetries between feedforward and feedback connections and the characteristic frequencies over which they operate.

Subfields: Cellular and Molecular Neuroscience, Cognitive Neuroscience, Optimization & Training, Probabilistic & Bayesian Methods

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Cellular and Molecular Neuroscience

In Cognitive Neuroscience

In Optimization & Training

In Probabilistic & Bayesian Methods

Citations in the corpus

What can the brain teach us about building artificial intelligence? (2016, background): “There is growing evidence for the use of MP-like inference in cortical areas [7, 8] and they could offer a happy medium where inference is fast as in neural networks, while retaining MCMC’s capability for answering arbitrary queries on the model.”
Deep Predictive Coding Networks for Video Prediction and Unsupervised Learning (2016, methodology): “Our work here builds on previous work in next-frame video prediction [26, 28, 29, 38, 48], but we take particular inspiration from the concept of “predictive coding” from the neuroscience literature [2, 5, 7, 9, 11, 21, 25, 32, 39, 40, 47, 49].”
The free energy principle for action and perception: A mathematical review (2017, background): “Here, the simple idea is that top-down cortical signalling conveys predictions while bottom-up activity returns prediction errors (Bastos et al., 2012).”
The neural architecture of language: Integrative modeling converges on predictive processing (2020, positive): “…Clark, 2013; Elman, 1990, 1991, 1993; McRae et al., 1998; Rohde & Plaut, 1999; Spivey & Tanenhaus, 1998; Tenenbaum et al., 2011) and neuroscience (Bastos et al., 2012; Keller & Mrsic-Flogel, 2018; Mumford, 1992; Rao & Ballard, 1999; Srinivasan et al., 1982), including in the domain of language…”
A free energy principle for a particular physics (2019, background): “r over future autonomous states; i.e., beliefs about autonomous behaviour. The models usually entertained in this setting consider continuous state-spaces – leading to schemes like predictive coding (Bastos et al., 2012; Rao and Ballard, 1999) – or discrete state-space models that are more apt for m”
Affective Value in the Predictive Mind (2017, background): “One of the most influential views on curiosity and exploration is Berlyne’s optimal level account (Berlyne 1970).”
Computing with a Canonical Neural Circuits Model with Pool Normalization and Modulating Feedback (2014, result): “This is in line with recent experimental observations that FB signals realize predictive coding principles (Bastos et al., 2012; Rauss, Schwartz, & Pourtoise, 2011); a reduction in response amplitude indicates that the predictions generated by FB signals match the driving FF signals (Markov &…”
Artificial Neural Networks for Neuroscientists: A Primer (2020, background): “Predictive coding Another canonical computation proposed for the brain is to compute predictions [Rao and Ballard, 1999, Bastos et al., 2012, Heilbron and Chait, 2018].”
Alpha-beta and gamma rhythms subserve feedback and feedforward influences among human visual cortical areas (2016, neutral): “…in essence that higher areas constantly derive predictions based on incoming evidence and prior experience, and feed those predictions back to lower areas, where they are subtracted from sensory evidence, such that only prediction errors need to be forwarded to higher areas (Bastos et al., 2012).”
Neural and Computational Mechanisms of Action Processing: Interaction between Visual and Motor Representations. (2015, background): “Other aspects are more speciﬁc, and some predictions about the signal ﬂow across cortical layers seem to match observations in real neurons (Bastos et al., 2012).”
Anatomy of hierarchy: Feedforward and feedback pathways in macaque visual cortex (2013, background): “…construct large-scale computational models, incorporating hierarchical integration of interareal connections into the local circuit of the cortex (Bastos et al., 2012), and thereby link up the concepts of cortical hierarchy, Bayesian inference, and the canonical circuit (Douglas and Martin,…”
Predictive Processing: A Canonical Cortical Computation (2018, positive): “It is possible that internal representation neurons are intermixed with prediction error neurons in different cortical layers, or that they are enriched in deep layers of cortex (Bastos et al., 2012), which are the main source of topdown signals (Felleman and Van Essen, 1991; Markov et al.”