Noonan, N. & Archibald, L.
The University of Western Ontario
There is considerable evidence to support the role of statistical language learning in segmenting speech, but the cognitive mechanisms that support this process are not as well understood. In a previous study, we examined the impact on implicit learning of word boundaries when healthy adult participants were engaged in explicit working memory tasks. We found reduced learning of an artificial language learning after 28 minutes of exposure when participants were engaged in a same-domain (verbal) but not cross-domain (visual spatial) working memory task. In the present study, we examined whether this domain-specific interference effect was an artifact of the long exposure time employed in our previous study. Participants were exposed to an artificial language for seven minutes while concurrently engaged in one of four working memory task conditions: (1) verbal WM, high-load; (2) verbal, low-load; (3) visuospatial WM, high-load; (4) visuospatial, low load. Using a multiple regression correlation ANOVA approach, performance across all groups was compared. Results showed that those in either of the verbal conditions were less accurate at identifying words from the artificial language compared to the all other groups. This result strengthens previous findings, demonstrating an effect of explicit processing on implicit language learning.