Predictive coding accelerates word recognition and learning in the early stages of language development.

Area of Science:

• Cognitive Neuroscience
• Developmental Psychology
• Linguistics

Background:

• Predictive abilities are crucial for adaptive behavior across species.
• Inferring predictions is hypothesized to facilitate early speech processing and word learning.

Purpose of the Study:

• To investigate the role of predictive coding in infant speech processing and word learning.
• To examine electrophysiological brain responses to predictable and unpredictable word contexts in infants.

Main Methods:

• Electrophysiological brain responses (event-related potentials) were measured in 12- and 24-month-old infants.
• Infants were exposed to predictable and unpredictable word contexts.
• Correlation analysis was performed between prediction error responses and vocabulary scores.

Main Results:

• Infants' brain responses were faster and more robust to predictable word endings.
• Word-expectancy violations in novel words elicited a prediction error response.
• The strength of the prediction error response correlated significantly with 12-month-old infants' vocabulary scores.

Conclusions:

• Predictive coding accelerates word recognition and supports the learning of novel words and their meanings in infants.
• Prediction error may facilitate learning by modulating attention, linking prediction errors to word meaning acquisition.
• Principles of predictive coding observed in other domains apply to language development in the infant brain.