Beta- and gamma-band activity reflect predictive coding in the processing of causal events

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Neuroscience

Background:

• Complex events are perceived causally, implying predictive brain processes.
• Predictive coding theory links beta-band activity to top-down predictions and gamma-band to bottom-up prediction errors.
• Neurocognitive evidence for predictive coding in higher-order cognitive areas is limited.

Purpose of the Study:

• Investigate neural activity during probability-dependent action perception.
• Examine predictive coding mechanisms in brain regions crucial for causal inference.
• Determine the roles of beta and gamma oscillations in action perception.

Main Methods:

• Magnetoencephalography (MEG) was used to record neural activity.
• Participants observed bowling action animations with varying predictability.
• Granger-causal connectivity analysis was performed on oscillatory activity.

Main Results:

• Beta-band Granger-causal connectivity showed strongest backward, top-down connections.
• Gamma-band Granger-causal connectivity showed strongest feed-forward, bottom-up connections.
• Temporoparietal junction (TPJ) beta-band power correlated with action predictability; TPJ and medial prefrontal cortex (MPFC) gamma-band power correlated with prediction error.

Conclusions:

• The brain employs predictive coding-like computations for higher-order cognitive functions, including causal event perception.
• Specific oscillatory bands (beta and gamma) reflect distinct predictive coding processes in key brain regions.
• Findings support predictive coding as a fundamental mechanism for integrating sensory information and generating predictions in the brain.