An event-related potentials account of brain predictive coding.

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Neuroscience

Background:

• Predictive coding theory explains brain processing as anticipatory, updating neural networks based on prediction accuracy.
• Electroencephalography (EEG) offers continuous, near-instantaneous brain activity recording.
• Event-related potentials (ERPs) are crucial for understanding neural responses to stimuli.

Purpose of the Study:

• To review the neural validity of predictive processing using ERPs.
• To investigate how the brain predicts future events, assesses predictions, and updates probabilities.
• To link specific ERP components to distinct stages of predictive coding.

Main Methods:

• Utilized predictive tone sequences and central cue Posner paradigms.
• Employed Bayesian modeling to analyze experimental data.
• Reviewed existing literature on ERPs and predictive coding.

Main Results:

• Contingent Negative Variation (CNV) is associated with prior expectation.
• Mismatch negativity (MMN) and P300 are linked to Bayesian surprise or prediction error.
• Post Imperative Negative Variation (PINV) may reflect outcome assessment in uncertain situations.

Conclusions:

• ERPs provide neural evidence supporting predictive coding mechanisms.
• Specific ERPs correlate with different aspects of prediction, error detection, and outcome evaluation.
• Predictive coding offers a viable framework for understanding brain operations indexed by ERPs.