Master of Science
Batterink, Laura J.
Statistical learning (SL) refers to our ability to extract patterns from the environment. Research has long acknowledged the importance of SL in language; however, the neural mechanisms underlying SL remain largely unknown. One potential mechanism is neural entrainment, which refers to the tendency of endogenous neural oscillations to align with an ongoing rhythmic stimulus. In the context of SL, neural entrainment may align neural excitability to the ongoing structure of the speech stream, increasing sensitivity to underlying patterns and supporting the learning of word boundaries. This thesis tested the hypothesis that neural entrainment plays a causal role in SL by directly manipulating entrainment at specific frequencies cross-modally using a visual stimulus during learning. We found that boosting neural entrainment to match the frequency of the most informative moments of continuous speech (word onsets) resulted in better performance in an implicit SL task. These results support that neural entrainment plays a causal role in SL, as opposed to simply reflecting downstream effects of the learning process.
Summary for Lay Audience
Every day humans are bombarded by a wealth of sensory input that needs to be successfully navigated to make sense of the world. Through continuous exposure to this sensory input, humans can extract statistical relationships about stimuli in the world, a skill often referred to as statistical learning (SL). Although this skill is found across domains, it is thought to be particularly important for language acquisition. More specifically, SL is thought to play a key role in an important initial step to acquiring language – the learning of word boundaries. Within a language, syllables within words occur together more frequently than syllables that span a word boundary. Becoming sensitive to these statistical relationships between syllables may allow learners to discover word boundaries, especially when other cues such as pauses may be unavailable or unreliable. Although research supports the relationship between SL and word segmentation, the neural mechanisms contributing to SL remain largely unknown. One potential mechanism is neural entrainment, which refers to the tendency of brain activity (neural oscillations) to align with an ongoing rhythmic stimulus. This thesis investigated the potential role of neural entrainment in SL by using a visual stimulus to induce entrainment at frequencies congruent or incongruent with word onsets (boundaries) during learning. We found that boosting neural entrainment to match the frequency of word onsets resulted in better performance in an implicit task, which supports the potential causal role of entrainment in SL. Theoretically, this work helped to advance our understanding of the neural mechanisms contributing to SL. Practically, future work in this area could reveal novel ways to boost SL and improve language acquisition for adult second language learners and children with atypical language development. Moreover, since SL is found across domains, techniques used and discovered may inform other areas beyond language, such as visual processing.
Mulgrew, Jordan-Jerrica, "The Role of Neural Entrainment in Statistical Learning" (2020). Electronic Thesis and Dissertation Repository. 7158.
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.