The Influence of Musical Training on Statistical Learning in Speech Segmentation: Combining Offline Methods With Online Neural Tracking (EEG)
Summary
Background: While Statistical Learning (SL), which is crucial for speech segmentation, is well-researched, the exact underpinnings of individual differences in SL remain unclear. Research has found that Musical Training (MT) positively affects various linguistic abilities,
but the influence of MT on SL in speech segmentation has yet to be studied.
Aim: The present study investigated the influence of MT on SL in speech segmentation, and examined whether MT is related to musical, specifically rhythmic, abilities.
Method: We used the data of 29 neurotypical Dutch monolingual adults who participated in the study by van der Wulp and colleagues (2023). With EEG, participants’ neural entrainment to two artificial languages, namely a structured stream (consisting of trisyllabic non-words with a TP of 1.0 between syllables within words and 0.33 across words) and a random stream (with a TP of 0.09 between all syllables), was measured. Afterwards, they completed a rating task and several musicality tasks (the Gold-MSI, CA-BAT, and PROMS).
Results: In the rating task data, years of MT seemed to positively influence SL. In the EEG data, SL was positively affected by the CA-BAT (i.e., one of the musicality tasks that assesses rhythmic ability) instead of MT. The MT measures (i.e., years of MT and the Gold-MSI’s MT subscale) positively correlated with each other and with the CA-BAT. The PROMS did not significantly correlate with any of the musicality measures.
Discussion and Conclusion: Musical ability involves aspects beyond just MT and seems positively related to SL, but it remains unclear which aspects of musicality affect SL. Future research is needed to establish the exact relationship between musicality and SL, preferably by exploring multiple aspects of musicality.