Impaired procedural learning in language impairment: Results from probabilistic categorization

Abstract

The Weather Prediction (WP) Task is a classical task of probabilistic category learning generally used for examining the dissociation of procedural and declarative memory. The current study focuses on performance of children with language impairment (LI) and compares their performance to that of typically developing (TD) children and adults with the aim of testing the procedural deficit hypothesis of LI (PDH; Ullman & Pierpont, 2005), which states that language impairment is not a specific linguistic phenomenon, but results from the dysfunction of a more general cognitive system: the procedural system. To test the generality of the procedural impairment, we needed a task that is dissimilar from language in that it does not build on sequential information. Children with language impairment show deficient learning on the Weather Prediction Task, which already appears at the early stages of the task. These results, in line with the PDH, point to the deficit of the procedural system in language impairment going beyond the language system. Whether this deficit is selective to the procedural system or is complemented by deficits in the declarative system is the subject of future studies.

Keywords:

• Implicit learning
• Procedural system
• Probabilistic categorization
• Language impairment
• Procedural deficit hypothesis

We are grateful to all adult participants and to the children and their teachers in the Dr. Nagy László Institute of Special Education in Kőszeg and in the primary school in Kerék Street in Budapest for their cooperation. Special thanks go to Attila Buknicz, István Kemény, and Bence Fadgyas for their help, to Dezső Németh and Szabolcs Kéri for their comments and suggestions, and to Karolina Janacsek and Gergely Éliás for their help in statistics. The research was funded by National Institutes of Health (NIH) R01DC00458 to Larry Leonard and the Országos Tudományos Kutatási Alapprogramok (Hungarian Scientific Research Fund, OTKA) TS 049840 to Csaba Pléh. During the research Ágnes Lukács was a grantee of the Bolyai János Research Scholarship of the Hungarian Academy of Science.

Notes

¹ Gluck et al. (2002) and studies that followed used the predictive values of 80%, 60%, 40%, and 20% in the WP task.