Influence of Lexical Versus Procedural Knowledge in Sentence Production in Response to Thematic Role Assignment Treatment



Journal Title

Journal ISSN

Volume Title



We tested whether one source of variability in treatment response and generalization for sentence production (Mitchum, Greenwald, & Berndt, 2000; Rochon, Laird, & Scofield, 2000) in individuals with non-fluent agrammatic aphasia is a deficit in the ability to assign thematic roles (procedural mapping). We propose that agrammatic individuals with better lexical verb retrieval abilities have a procedural mapping deficit (Schwartz, Saffran, Fink, Myers, & Martin, 1994; Marshall, 1995), and, therefore, demonstrate better response and generalization to treatment which focuses on the procedural assignment of thematic roles. In contrast, we hypothesized agrammatic individuals with poorer lexical verb retrieval abilities have a lexical mapping deficit and exhibit worse treatment outcomes and generalization of treatment. Fifteen right-handed, monolingual English-speaking adult males and females who had unilateral left hemisphere CVAs with subsequent chronic non-fluent aphasia and no co-morbid neurologic impairments participated in this study. All participants had agrammatism as confirmed by their poor performance during production of passive reversible verb sentences at baseline (≤30% accuracy). Verb retrieval, as measured by the Action Naming Test (Obler & Albert, 1986), and active sentence comprehension, as measured by the active sentences on the Circles and Squares Syntactic Comprehension Test, (adopted from Schwartz, Saffran, & Martin, 1980) served as indicators of lexical and procedural mapping abilities, respectively. The dependent variables were treatment response, as determined by sentence production to picture stimuli using correct thematic role assignment, and treatment generalization, as measured by the non-treated sentence production using correct thematic role assignment. In order to examine the relationship between verb retrieval and active sentence comprehension correlation analysis was used. Linear regression was used to determine the strength of the relationship between the independent variables (severity of the type of mapping deficit) and dependent variables (treatment outcomes), while considering the potential impacts of overall aphasia severity and age. The analyses revealed no significant correlation between verb retrieval and active reversible sentence comprehension (r=0.417, p =.122) suggesting these two measures test separate stages of sentence processing. We found overall aphasia severity as indicated by the Western Aphasia Battery A.Q. (Kertesz, 1982) correlated with verb retrieval (r=0.781, p=.001) and active sentence comprehension (r=0.653, p=.008), suggesting that overall aphasia severity considers linguistic impairments captured by both verb retrieval and active sentence comprehension measures. Overall aphasia severity was the main predictor of treatment response (R²=.382), whereas age was not correlated with either treatment response (r=-0.18, p=.53) or generalization (r=-0.15, p=.60), which is consistent with the literature (Pederson, Vinter, & Olsen, 2004; Plowman, Hentz, & Ellis, 2011). Verb retrieval, however, was the only statistically significant predictor of treatment generalization (R²=.351). Our results demonstrate that better mapping treatment outcomes are predicted by relatively preserved verb retrieval, where individuals with poor verb retrieval demonstrate treatment specific gains but no treatment generalization. However, individuals who can access the lexical representation of the verb for further processing demonstrate not only a positive treatment response, but also a generalized response suggesting procedural re-learning of the process of thematic role assignment.



Aphasia, Mapping treatment, Mapping deficit hypothesis, Agrammatism