Abstract:We introduce dynamic predictive coding, a hierarchical model of spatiotemporal prediction and sequence learning in the neocortex. The model assumes that higher cortical levels modulate the temporal dynamics of lower levels, correcting their predictions of dynamics using prediction errors. As a result, lower levels form representations that encode sequences at shorter timescales (e.g., a single step) while higher levels form representations that encode sequences at longer timescales (e.g., an entire sequence). We tested this model using a two-level neural network, where the top-down modulation creates low-dimensional combinations of a set of learned temporal dynamics to explain input sequences. When trained on natural videos, the lower-level model neurons developed space-time receptive fields similar to those of simple cells in the primary visual cortex while the higher-level responses spanned longer timescales, mimicking temporal response hierarchies in the cortex. Additionally, the network's hierarchical sequence representation exhibited both predictive and postdictive effects resembling those observed in visual motion processing in humans (e.g., in the flash-lag illusion). When coupled with an associative memory emulating the role of the hippocampus, the model allowed episodic memories to be stored and retrieved, supporting cue-triggered recall of an input sequence similar to activity recall in the visual cortex. When extended to three hierarchical levels, the model learned progressively more abstract temporal representations along the hierarchy. Taken together, our results suggest that cortical processing and learning of sequences can be interpreted as dynamic predictive coding based on a hierarchical spatiotemporal generative model of the visual world. The brain is adept at predicting stimuli and events at multiple timescales. How do the neuronal networks in the brain achieve this remarkable capability? We propose that the neocortex employs dynamic predictive coding to learn hierarchical spatiotemporal representations. Using computer simulations, we show that when exposed to natural videos, a hierarchical neural network that minimizes prediction errors develops stable and longer timescale responses at the higher level; lower-level neurons learn space-time receptive fields similar to the receptive fields of primary visual cortical cells. The same network also exhibits several effects in visual motion processing and supports cue-triggered activity recall. Our results provide a new framework for understanding the genesis of temporal response hierarchies and activity recall in the neocortex.

Posteromedial cortical networks encode visuomotor prediction errors.

Cooperative thalamocortical circuit mechanism for sensory prediction errors

Stimulus history, not expectation, drives sensory prediction errors in mammalian cortex

Cortical networks with multiple interneuron types generate oscillatory patterns during predictive coding

Hierarchy of prediction errors shapes the learning of context-dependent sensory representations

Learning prediction error neurons in a canonical interneuron circuit

Computational components of visual predictive coding circuitry

High-level visual prediction errors in early visual cortex

Neural Elements for Predictive Coding

Predictive Coding of Reward in the Hippocampus

Crossmodal hierarchical predictive coding for audiovisual sequences in the human brain

Expectation violations produce error signals in mouse V1

Learned response dynamics reflect stimulus timing and encode temporal expectation violations in superficial layers of mouse V1

Stimulus-dependent adjustment of reward prediction error in the midbrain

Spatiotemporal dynamics across visual cortical laminae support a predictive coding framework for interpreting mismatch responses

Temporal Prediction Errors Modulate Cingulate–insular Coupling

Top-down modulation of sensory processing and mismatch in the mouse posterior parietal cortex

Learning within a sensory-motor circuit links action to expected outcome

Typical Neural Representations of Predictive Coding in Visual and Auditory Sensory Processing

Striatal hub of dynamic and stabilized prediction coding in forebrain networks for olfactory reinforcement learning

Dynamic predictive coding: A model of hierarchical sequence learning and prediction in the neocortex