The Influence of Phonological Components Analysis Treatment on Lexical Access in Individuals with Aphasia

• ¹ University of Ottawa, Audiology and Speech-Language Pathology Program, Canada
• ² University of Toronto, Department of Speech-Language Pathology, Canada
• ³ University of Iowa, Communication Sciences and Disorders, United States
• ⁴ Heart and Stroke Foundation Canadian Partnership for Stroke Recovery, Canada
• ⁵ UHN Toronto Rehabilitation Institute, Canada

Background: The Phonological Components Analysis (PCA) treatment for naming deficits in aphasia has been shown by us and others to be efficacious (Leonard et al., 2015; Leonard et al., 2008; van Hees et al, 2013), with associated neural changes (Rochon et al., 2010). This treatment protocol was modeled after the Semantic Features Analysis approach (Boyle & Coelho, 1995) and both capitalize on the presumed bidirectional connections between semantics and phonology during lexical access (Foygel & Dell, 2000). Based on this two-step interactive activation model, the purpose of the present study was to examine the influence of PCA therapy on the strength of both the semantic and phonological connections during lexical access in individuals with aphasia. Methods: The data reported here are part of a larger study investigating the influence of intensity of PCA treatment on naming performance (Rochon et al., in preparation). In that study participants with aphasia were randomly assigned to one of two treatment conditions (intense or standard). They were asked to name pictured items and naming accuracy of treated (n = 30) and untreated (n = 30) words was collected pre- and post-therapy. Two independent raters, blind to treatment condition and pre- post-treatment status coded the responses by error type (e.g., semantic, formal, unrelated, etc.) as per Dell et al. (1997). Point-to-point agreement was 93%. The scored responses, per patient, were entered into a web-based computational model of lexical access (Walker & Hickok, 2015) in order to determine the strength of the phonological and semantic connections pre- and post-therapy. Results: The outcome measure for this analysis, connection weight, was submitted to a 2 x 2 x 2 x 2 mixed ANOVA with 3 repeated measures (treatment status [treated vs untreated]; session [pre vs post]; parameter [semantic vs phonological]; and one between-subjects factor (intensity [intense vs standard]). There were significant main effects of treatment status (treated > untreated), F (1, 12) =12.30, p = .004; session (post > pre), F (1, 12) = 24.92, p < .001; and parameter (phonological > semantic), F (1, 12) =9.78, p = .009. A significant 3-way interaction, treatment status x session x parameter, F (1, 12) =5.35, p = .039, was also found. Post-hoc tests showed that the change in the connection weight pre- and post-treatment was specific to the semantic weights for the treated items only (p < .001; see Figure 1). Discussion: It was found that PCA treatment, a phonologically-based therapy, strengthened the semantic connections during lexical access in individuals with aphasia. This finding is consistent with our previous work (Leonard et al., 2015; Rochon et al., 2010) where it was found, using fMRI, that changes in cortical activation post PCA therapy were associated with a semantic task. These results are consistent with Foygel and Dell’s (2000) interactive activation model of lexical access, which posits a strong interaction between the semantic and phonological levels during naming. Figure 1: Semantic and Phonological Connection Weights Pre- and Post-therapy