Predictive encoding of pure tones and FM-sweeps in the human auditory cortex

Area of Science:

• Neuroscience
• Auditory Neuroscience
• Cognitive Neuroscience

Background:

• Expectations significantly shape sensory perception, yet the precise neural underpinnings remain unclear.
• A leading hypothesis suggests sensory neurons signal prediction errors relative to anticipated sensory input.
• Previous studies in the auditory cortex (AC) demonstrated prediction error encoding, but often used stimulus repetition, risking conflation with neural habituation.

Purpose of the Study:

• To investigate prediction error encoding in specific human auditory cortical fields (Te1.0, Te1.1, Te1.2, Te3).
• To examine prediction error for both pure tones and frequency-modulated (FM) sweeps, independent of stimulus repetition.
• To clarify the role of predictive coding in the AC.

Main Methods:

• Functional magnetic resonance imaging (fMRI) was employed to study subjective expectations.
• Participants' brain activity was measured while they received pure tones and FM sweeps.
• Analysis focused on specific auditory cortical fields to assess differential responses.

Main Results:

• Prediction error signals were observed in response to subjective expectations, irrespective of stimulus repetition.
• These prediction error signals were consistently found across different auditory cortical fields.
• Both pure tones and FM sweeps elicited prediction error encoding within the AC, despite distinct processing strategies.

Conclusions:

• The auditory cortex (AC) employs predictive coding as a fundamental mechanism for processing sensory information.
• Prediction error is a general encoding principle in the AC, not confounded by stimulus repetition or specific stimulus types.
• This study provides robust evidence for predictive coding across multiple auditory cortical fields and stimulus modalities.