Mechanisms underlying naming impairments in post-stroke and progressive aphasia

Abstract

This thesis investigated mechanisms underlying impairments in naming, word repetition and nonword repetition in four different populations: patients recovering from post-stroke aphasia who received treatment (PSA), a control group of patients who did not receive treatment (CPSA), and patients with two variants of progressive aphasia (i.e., fluent FPA and nonfluent NPA). Predictions based on the interactive activation (IA) model of lexical retrieval (Foygel Dell, 2000; Hanley, Dell, Kay Baron, 2004) were tested at two time points in order to evaluate the viability of the model and its two mechanisms, the strength of semantic and phonological connections. The status of the putative non-lexical route in nonword repetition and its relationship to the model's other two mechanisms was also examined. Results of the naming study showed that, on average, patients in the PSA and CPSA groups presented with a combination of semantic and phonological impairments, while patients in the NPA and FPA groups presented with more uniform impairments to the semantic connections. Data from the word repetition study pointed to impairments mostly to the phonological connections in all groups. This combination of mechanisms appeared in the presence of what was usually impairment to the non-lexical route in the PSA and CPSA groups and a relatively spared non-lexical route in the NPA and FPA groups. The mechanisms underlying impairments were relatively stable over time but stability of lexical diagnosis was not seen across the tasks of naming and word repetition. Results also revealed that the non-lexical route, claimed by the model's proponents to be independent from the model's two other mechanisms, appeared in fact, to have a relationship with the phonological connections of the network. Participants with impaired phonological connections tended to have the non-lexical route impaired while those with impaired semantic connections (i.e., spared phonological connections) presented with a relatively spared non-lexical route. Findings from the three studies are discussed in the framework of the model's viability and its shortcomings. Limitations of this study are also discussed. Potential clinical and theoretical implications are presented along with directions for future research on lexical retrieval models.