Ahufinger, N. 1, 2 , Guerra, E. . 3 , Ferinu, L. 1 , Sanz-Torrent, M. 2 & Andreu, L. 1
1 Universitat Oberta de Catalunya
2 Universitat de Barcelona
3 Universidad de Chile
Cross-situational statistical learning (CSSL) is a fast-mapping mechanism that extends over multiple encounters where the ambiguity and exposure to different words and possible referents at the same time is presented (Smith & Yu, 2007). Past research has shown children with SLI have difficulties to solve tasks that require sequential implicit learning (Ullman & Pierpont, 2005). We investigated whether children with SLI can solve a non-sequential CSSL task and whether the moment-by-moment visual attention pattern reflected the performance in the task. 38 children with SLI (Mage=8:7 years), and 38 age/gender-matched TD children (Mage= 8;9 years) participated in the study. Behavioral data: results for SLI group were significantly poorer than the TD group. Eye-tracking data: TD children but not children with SLI showed a clear pattern of looks from the left-to-right through the exposure phase. Moreover, TD children showed a higher proportion of looks to the target and were faster when identifying it through the testing phase in comparison to SLI group. These results suggested differences between groups in the online cognitive process when learning novel words implicitly. Results reflect that the statistical learning difficulties in children with SLI were observed both in implicit behavioral measures as well as overt behavioral responses.