Cracking the language code: Neural basis of word segmentation throughout development

Research has shown that a fundamental aspect of language acquisition involves the identification of word boundaries in continuous speech. We know that this process, called word segmentation, occurs as the brain computes the statistical regularities and prosodic cues available in the input; however, little was previously understood about the neural mechanisms by which word segmentation takes place throughout development. A first functional magnetic resonance imaging (fMRI) study provided a neural signature for online implicit word segmentation in adults, as indicated by significant signal increases over time in left-lateralized temporal regions as a function of exposure to continuous speech streams with strong statistical regularities between syllables and not during exposure to a speech stream with weak statistical cues. Two further fMRI studies investigated developmental changes in the neural mechanism of word segmentation. While both children and adults displayed increases in activity in temporal regions while listening to the streams containing strong statistical cues (which we take to reflect successful implicit parsing of the speech stream), adults engaged a more focal neural network comprised of canonical language cortices and children recruited a wider neural network including regions involved in attention and working memory. In addition, there was a significant shift in the laterality of the signal increases in temporal cortex from the right to the left hemisphere with age. Interestingly, children appear to be more sensitive than adults to subtle statistical cues that guide word segmentation, as evidenced by signal increases over time in temporal cortices for the speech stream that contained weak statistical cues in six- and ten-year-old children but not thirteen-year-olds or adults. Importantly, after the brief exposure to the speech streams, listening to the isolated words that comprised an artificial language resulted in greater activation in the left inferior frontal gyrus than listening to combinations of syllables that did not form words within that artificial language for both children and adults. Finally, an investigation of the impact of exposure to a second language and pubertal development on the learning-related signal increases in ten-year-old children revealed that the neural circuitry for language learning is influenced by both experiential and maturational factors. Taken together, these findings begin to elucidate the complex relationship between language learning, maturation, and linguistic experience and lend insight into why children are better language-learners than adults.