# Establishing neural representations for new word forms in 12-month-old infants

## TL;DR

This study explores how 12-month-old infants process newly learned words in their brain, revealing insights into early language development.

## Contribution

The study reveals new neural patterns in infants processing newly learned words, challenging predictive coding assumptions.

## Key findings

- Newly learned words elicit mismatch responses similar to known words in infants.
- Prediction errors from acoustic novelty show distinct positive polarity responses.
- Results suggest electric brain activity isn't fully explained by predictive coding.

## Abstract

During the first year of life, infants start to learn the lexicon of their native language. Word learning includes the establishment of longer-term representations for the phonological form and the meaning of the word in the brain, as well as the link between them. However, it is not known how the brain processes word forms immediately after they have been learned. We familiarized 12-month-old infants (N = 52) with two pseudowords and studied their neural signatures. Specifically, we determined whether a newly learned word form elicits neural signatures similar to those observed when a known word is recognized (i.e., when a well-established word representation is activated, eliciting enhanced mismatch responses) or whether the processing of a newly learned word form shows the suppression of the neural response along with the principles of predictive coding of a learned rule (i.e., the order of the syllables of the new word form). The pattern of results obtained in the current study suggests that recognized word forms elicit a mismatch response of negative polarity, similar to newly learned and previously known words with an established representation in long-term memory. In contrast, prediction errors caused by acoustic novelty or deviation from the expected order in a sequence of (pseudo)words elicit responses of positive polarity. This suggests that electric brain activity is not fully explained by the predictive coding framework.