The encoding of stochastic regularities is facilitated by action-effect predictions

Area of Science:

• Neuroscience
• Auditory Perception
• Cognitive Science

Background:

• The brain constantly forms predictive models using sensory input and actions.
• Auditory systems struggle to encode stochastic regularities with rare deviants, shown by a lack of mismatch negativity (MMN).
• Predictive coding theory suggests higher-order predictions influence sensory processing.

Purpose of the Study:

• To investigate if action intention predictions enhance auditory regularity encoding.
• To test if top-down predictions from action intention modulate sensory prediction error signals.
• To explore how action-effect predictions interact with stochastic regularity-based predictions.

Main Methods:

• Participants generated random high/low pitch sound sequences via button presses.
• Two conditions: 'specific' (one button per pitch) and 'unspecific' (both buttons random pitch).
• Measured mismatch negativity (MMN) and N2 event-related potentials to rare medium pitch deviants.

Main Results:

• Larger MMN and N2 responses to pitch deviants in the 'specific' condition versus the 'unspecific' condition.
• This occurred despite equal sound probabilities across conditions.
• Action intention predictions significantly modulated neural responses.

Conclusions:

• Action-effect predictions can enhance predictions based on stochastic regularities.
• Higher-order action intention signals engage deviance detection mechanisms at sensory levels.
• Findings extend the role of action predictions in sensory processing and predictive coding.