CoChat
Sign Up
Cited by 724Visual perception and processing mechanismsNeural dynamics and brain functionNeural Networks and Applications

Learning and inference in the brain

Friston, Karl·2003·Neural Networks
DOI 10.1016/j.neunet.2003.06.005
Open this paper in CoChat — free account

Abstract

This article is about how the brain data mines its sensory inputs. There are several architectural principles of functional brain anatomy that have emerged from careful anatomic and physiologic studies over the past century. These principles are considered in the light of representational learning to see if they could have been predicted a priori on the basis of purely theoretical considerations. We first review the organisation of hierarchical sensory cortices, paying special attention to the distinction between forward and backward connections. We then review various approaches to representational learning as special cases of generative models, starting with supervised learning and ending with learning based upon empirical Bayes. The latter predicts many features, such as a hierarchical cortical system, prevalent top-down backward influences and functional asymmetries between forward and backward connections that are seen in the real brain. The key points made in article are: (i). hierarchical generative models enable the learning of empirical priors and eschew prior assumptions about the causes of sensory input that are inherent in non-hierarchical models. These assumptions are necessary for learning schemes based on information theory and efficient or sparse coding, but are not necessary in a hierarchical context. Critically, the anatomical infrastructure that may implement generative models in the brain is hierarchical. Furthermore, learning based on empirical Bayes can proceed in a biologically plausible way. (ii). The second point is that backward connections are essential if the processes generating inputs cannot be inverted, or the inversion cannot be parameterised. Because these processes involve many-to-one mappings, are non-linear and dynamic in nature, they are generally non-invertible. This enforces an explicit parameterisation of generative models (i.e. backward connections) to afford recognition and suggests that forward architectures, on their own, are not sufficient for perception. (iii). Finally, non-linearities in generative models, mediated by backward connections, require these connections to be modulatory, so that representations in higher cortical levels can interact to predict responses in lower levels. This is important in relation to functional asymmetries in forward and backward connections that have been demonstrated empirically.

Integrity signals

Registry

Crossref · 10.1016/j.neunet.2003.06.005

Cites

91 of 91 references verified0 citation errors · 0 not checkable

Statistics

Recomputing p-values needs the paper’s full text.

Claims the paper makes

This paper’s full text isn’t openly available, so its claims and self-check weren’t run. The registry and citation checks still apply.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • This procedure combines top-down and bottom-up influences in an especially delicate and potent fashion, and leads to neurons with selectivity that is not intrinsic to the area but depends on interactions across levels of a processing hierarchy.

    “This procedure combines ' top-down ' and ' bottom -up ' in fl uences in an especially delicate and potent fashion, and it leads to the development of neurons that exhibit a ' selectivity that is not intrinsic to the area but depends on interactions across levels of a processing hierarchy ' (Friston 2003, p. 1349).”

    From Whatever next? Predictive brains, situated agents, and the future of cognitive science (2013) · cites this paper as Friston 2003

  • Grush's account of perception as 'environmental emulation' looks highly congruent with the depiction of perception as reconstructing the hidden causes structuring the sensory signal.

    “In addition, Grush ' s account of perception as ' environmental emulation ' (see section 5.2 of Grush 2004) looks highly congruent with the depiction (Friston 2003 and elsewhere) of perception as reconstructing the hidden causes structuring the sensory signal.”

    From Whatever next? Predictive brains, situated agents, and the future of cognitive science (2013) · cites this paper as Friston 2003

  • Numerous neurophysiological studies have provided evidence for signaling of prediction errors along forward connections and predictions along backward connections.

    “Numerous neurophysiological studies have provided evidence for signaling of prediction errors along forward connections and predictions along backward connections [(63, 72) for reviews see (126, 127)].”

    From Can Bayesian Theories of Autism Spectrum Disorder Help Improve Clinical Practice? (2016) · cites this paper as [126]

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Authors
Friston, K.
Affiliations
UCL Queen Square Institute of Neurology (GB)
Corresponding author
Friston, Karl
ORCID
Friston, Karl
Venue
Neural Networks
Publisher
Elsevier BV
Published
2003-11
Volume
16
Issue
9
Pages
1325-1352
ISSN
0893-6080 (ISSN-L)
DOI
10.1016/j.neunet.2003.06.005
OpenAlex
W2131352719
Cited by
724
References
91
Citations by year
2026: 26 · 2025: 51 · 2024: 34 · 2023: 52 · 2022: 51 · 2021: 60 · 2020: 83 · 2019: 33 · 2018: 38 · 2017: 36 · 2016: 25 · 2015: 31 · 2014: 24 · 2013: 26 · 2012: 23
FWCI
4.56
Type
journal article
Language
en
Open access
closed
Licence
www.elsevier.com/tdm/userlicense/1.0/ (tdm)
Locations
  • Neural Networks (publishedVersion)
  • PubMed (publishedVersion)
  • UCL Discovery (University College London) (submittedVersion)
Topics
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Visual perception and processing mechanisms
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Neural dynamics and brain function
  • Physical Sciences › Computer Science › Artificial Intelligence › Neural Networks and Applications
Keywords
representation learning, generative models, top-down connections, empirical Bayes, empirical priors, functional asymmetry, sparse coding, efficient coding, information theory, perception, brain architecture
MeSH terms
Bayes Theorem, Brain, Learning, Neural Networks, Computer

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review