Related Experiment Video
Updated: Jul 9, 2026
13:51
Cross-Modal Multivariate Pattern Analysis
Published on: November 9, 2011
How Precise Is Predictive Coding for Things We Hear? Mismatch Negativity With Shepard Tones.
Urte Roeber1,2, Robert P O'Shea1,2
1Wilhelm Wundt Institute for Psychology, BioCog: Cognitive and Biological Psychology, Leipzig University, Leipzig, Germany.
The European Journal of Neuroscience
|July 8, 2026
Summary
The brain predicts auditory patterns. Mismatched tones, especially lower ones, trigger stronger neural responses, revealing the auditory system
Area of Science:
- Neuroscience
- Auditory Perception
- Predictive Coding
Background:
- Predictive coding theory posits hierarchical neural processing using top-down and bottom-up information to model and predict sensory input.
- Deviations from the model necessitate neural updates, reflecting the brain's continuous prediction and error-correction mechanisms.
Purpose of the Study:
- To investigate the precision of predictive coding models within the human auditory system.
- To determine the neural sensitivity to pitch variations in a discrete Shepard scale.
Main Methods:
- Utilized electroencephalography (EEG) with 128 scalp electrodes to record brain activity in 20 participants.
- Presented participants with a discrete Shepard scale, occasionally replacing expected tones with slightly altered pitches (±1/3 or ±2/3 semitone).
Main Results:
- Observed event-related potentials (ERPs) between 180-220 ms in response to unexpected tones.
- ERP negativity increased with the magnitude of the pitch deviation from the predicted tone.
- Found greater neural sensitivity to undershoots (tones lower than expected) compared to overshoots.
Conclusions:
- The auditory predictive model demonstrates high sensitivity, detecting pitch deviations as small as one-third of a semitone.
- The brain's predictive model exhibits asymmetric sensitivity, responding more strongly to pitch decreases than increases.
