N400 evidence for musical facilitation of word boundary identification in second language exposure

Lexical acquisition requires the ability to identify word boundaries in a continuous auditory speech stream. This complex task is even more challenging when learning a new language in adulthood. Previous studies have shown that word boundary identification can be enhanced by pairing musical tones with native language phonemes. The objective of this dissertation study was to investigate whether musical tones also have this effect in a novel pseudo-language that uses non-native speech sounds. The N400, a brain event-related potential that has been linked with familiarity responses and detection of statistical regularities during exposure to pseudowords, provides an index of brain activation associated with semantico-lexical processing. In this study, language-like stimuli incorporating a French phoneme (a high, front, rounded vowel that is not part of the English phonetic inventory) were presented to typically developing English monolingual adults. Participants were presented to one of two types of exposure conditions for 7 minutes: monotone presentation of the concatenated language-like stimuli; or the same speech stream with a musical tone associated with each syllable. The exposure protocol was based on Schön, Boyer, Moreno et. al. (2008). Exposure was followed by a lexical decision task, requiring participants to distinguish “words” (heard during the exposure in a concatenated speech stream) from “part words” (end of one word and the beginning of another, crossing word boundaries). High-density EEG was recorded during the lexical decision and analyzed offline to determine N400 event-related responses to the stimuli in each condition. Although behavioral measures did not reveal any significant differences between groups or conditions, we found a N4 significantly different response to “partword” in the tone-exposed group, compared to the monotone. This difference only occurred in a frontal region with a right-hemisphere bias, and was not found to be significant over the left hemisphere. This difference suggests that participants in the tone group were supported in differentiating “words” from “partwords”, supporting the view that the inclusion of tonal information is beneficial in the early stages of L2 lexical learning.