Neural Correlates of Native-language Speech Perception and Non-native Speech Sound Learning

Many studies of non-native speech sound learning report a great deal of individual variability; some learners master the sounds of a second language with ease, while others struggle to perceive and produce sounds, even after years of learning the language. Although some contributions of phonological, auditory, or cognitive skills have been found to predict non-native speech sound learning ability as measured by laboratory tasks, the field lacks a comprehensive understanding of where these differences originate from. Recent findings, however, suggest that individual differences in sleep duration may predict learning after a period of offline consolidation, though these findings are mixed. Another issue is that the large amount of individual variability seen in studies of non-native learning makes it difficult to obtain precise estimates of effect sizes. Therefore, the first aim of this dissertation was to replicate and extend recent behavioral and neuroimaging findings in non-native speech sound learning with a larger sample size than is typical. The second goal was to test a new question, namely, that how consistently and categorically listeners perceive native-language sounds will predict success on non-native speech sound learning tasks. Finally, we sought to establish whether measures of brain structure can predict how categorically listeners perceive sounds in the native language and how consistently they respond to those sounds. We did not replicate recent findings showing behavioral improvement after sleep on non-native speech sound learning tasks, nor did we replicate the finding that sleep duration predicts overnight improvement. However, gyrification of the bilateral transverse temporal gyri and hippocampal volume predicted an individual’s overnight improvement, suggesting a role for memory consolidation, even though we did not see overnight improvement at the group level. We additionally did not find that individual differences in categorical perception predicted non-native speech sound learning, which presents a challenge for some predominant theories of non-native speech sound learning, which future research will have to address. Overall, learners with reduced surface area and volume in frontal regions showed more graded and consistent perception of native-language speech sounds, supporting the notion that these regions underlie categorical perception