Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Deep temporal models and active inference.

Karl J Friston1, Richard Rosch1, Thomas Parr1

  • 1Wellcome Trust Centre for Neuroimaging, Institute of Neurology, University College London, WC1N 3BG, UK.

Neuroscience and Biobehavioral Reviews
|April 19, 2017
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Fragile memories for fleeting percepts.

Consciousness and cognition·2026
Same author

Supporting recruitment into complex trials: An embedded qualitative process evaluation in the RADICAL trial of radiofrequency denervation for chronic low back pain.

British journal of pain·2026
Same author

Breakthrough percepts of familiar faces.

Brain informatics·2026
Same author

Corrigendum to 'Intradiscal linezolid (PP353) treatment for chronic low back pain associated with Modic change type 1: an international, first-in-human, randomised, sham procedure-controlled, double-blind, phase 1b clinical trial' eClinicalMedicine 2026; 92: 103764.

EClinicalMedicine·2026
Same author

Intradiscal linezolid (PP353) treatment for chronic low back pain associated with Modic change type 1: an international, first-in-human, randomised, sham procedure-controlled, double-blind, phase 1b clinical trial.

EClinicalMedicine·2026
Same author

Reframing the Expected Free Energy: Four Formulations and a Unification.

Neural computation·2026
Same journal

Peripheral Nervous System Involvement in Parkinson's Disease: Peripheral Neuropathy, Neuromuscular Junction Dysfunction, and Clinical Implications.

Neuroscience and biobehavioral reviews·2026
Same journal

Violence on the self: Exploring the intersection of trauma and identity among sexual and gender minorities from a neurobiological and developmental perspective to advance our understanding of PTSD, moral injury, and dissociation symptoms.

Neuroscience and biobehavioral reviews·2026
Same journal

Meaning in life and biological functioning: A multisystem synthesis and agenda for future research.

Neuroscience and biobehavioral reviews·2026
Same journal

Beyond Diagnosis: Why and How Virtual Reality Should be Used in Research on Neurodevelopmental Conditions?

Neuroscience and biobehavioral reviews·2026
Same journal

What eye-movements tell us about Disorders of Consciousness?

Neuroscience and biobehavioral reviews·2026
Same journal

Systematic Review of Tactile-Based Interventions Combined with Multisensory Stimulation Protocols in the Rehabilitation of Patients with Disorders of Consciousness.

Neuroscience and biobehavioral reviews·2026
See all related articles

This review explores active inference using deep temporal models to understand how we navigate complex information. These models simulate brain activity during tasks like reading, explaining responses to both local and global context changes.

Area of Science:

  • Computational Neuroscience
  • Cognitive Science
  • Machine Learning

Background:

  • Understanding how the brain processes complex, structured information is a key challenge.
  • Previous work in active inference provides a framework for understanding decision-making and perception.
  • Hierarchical generative models offer a way to represent nested temporal dependencies.

Purpose of the Study:

  • To extend active inference with deep temporal models for simulating brain function.
  • To model sequential inference and evidence accumulation over nested time scales.
  • To explain behavioral and electrophysiological responses during information processing, like reading.

Main Methods:

  • Simulated behavioral and electrophysiological responses using hierarchical generative models.
Keywords:
Active inferenceBayesianFree energyHierarchicalMMNP300ReadingViolation

Related Experiment Videos

  • Employed active inference principles with deep temporal models incorporating nested time scales.
  • Utilized Bayesian belief updating and neuronal process theories for simulations.

    • Main Results:

    • Simulations successfully reproduced empirical findings, including perisaccadic delay period activity and local field potentials during reading.
    • The models explained epistemic foraging behavior observed in visual search tasks.
    • Simulated responses to local (e.g., font) and global (e.g., semantic) context violations matched empirically observed mismatch negativity and P300 potentials.

    Conclusions:

    • Deep temporal models within the active inference framework can successfully simulate complex cognitive processes.
    • These models provide a unified account of sequential inference, temporal scene understanding, and context-dependent neural responses.
    • The findings offer insights into how the brain navigates structured environments and processes information across multiple time scales.