Temporal and spatial localization of prediction-error signals in the visual brain

Area of Science:

• Neuroscience
• Cognitive Science
• Visual Perception

Background:

• The brain is thought to predict environmental changes, but direct electrophysiological evidence of visual prediction violations is limited.
• Understanding how the brain processes context and generates predictions is crucial for cognitive neuroscience.

Purpose of the Study:

• To provide real-time electrophysiological evidence of prediction violations in visual perception.
• To investigate the neural correlates of expectancy violations in response to visual stimuli.
• To explore the role of early visual evoked potentials in signaling prediction errors.

Main Methods:

• Participants viewed image sequences with predictable or violated implied sequences.
• Electroencephalography (EEG) and Magnetoencephalography (MEG) recorded neural responses.
• Stimulus design ensured neural differences were due to context, not image content.

Main Results:

• Early (N170) and mid-latency (N300) visual evoked potentials were modulated by violated sequences.
• Expectancy violation signals (N/M170) were localized to motion perception areas.
• Modulation occurred across various image changes and object categories.

Conclusions:

• The N/M170 brain signal indexes mismatches between predicted and actual visual input.
• This suggests the visual system predicts trajectories based on context.
• The N/M170 may represent a broader class of signals resolving top-down and sensory data conflicts.