Frontiers in Psychology (Apr 2014)
The (null) impact of pseudohomophones in lexical decision: No phonological activation in semantic dementia?
Abstract
Individuals suffering from the semantic variant of primary progressive aphasia (svPPA) show semantic degradation and impaired lexical decision (LD) (Rogers, Lambon Ralph, Hodges, & Patterson, 2004). The PDP-triangle model states that written word recognition depends on the interaction between perceptual and conceptual processing, i.e. orthography-to-semantics (O-S) mappings. Thus semantic impairment should affect normal recognition. Additionally, the pseudohomophone (PsH) effect in LD is traduced in longer latencies and more errors for pseudowords that sound like a real word (roZe), as compared to pseudowords that do not (rone). This effect is considered a marker of phonological activation in LD (Braun, Hutzler, Ziegler, Dambacher, & Jacobs, 2009) since a PsH will activate an existing phonological representation, creating a conflict between orthography-phonology (O-P). The aim of this study was to examine the performance of patients with svPPA in a LD task with PsHs. We predicted that patients should be impaired because of a deficit in O-S mappings. However, O-P mappings are assumed to be intact in svPPA and a PsH effect should be found for patients, as described for normals. Method Participants. Eight French-speaking patients that fulfilled clinical and imaging-supported criteria for svPPA (Gorno-Tempini et al., 2011) were matched by age, gender and education to 16 normal controls. Stimuli. LD task with (1) irregular- and (2) regular-words, (3) PsH and (4) pseudowords (n=60 per condition). All conditions were matched by length (in letters and phonemes), bigram frequency and N-size. Conditions 1-3 were additionally matched by frequency and imageability. Procedure. Stimuli were randomly presented on a laptop using the DMDX software. Each participant was asked to decide whether stimuli were real words. Data analysis. ANOVAs were carried out with RTs and errors as dependent variables and type of stimuli as repeated measures and participant group as between subject factor using SPSS-19. Simple effects were studied with paired t-tests. Results For both RTs and errors: (a) svPPA patients were significantly slower and more error-prone than controls; (b) the effects of lexicality and PsHs were found for controls only. Discussion The results in controls reproduce those found in literature (Braun et al., 2009). The general impairment and the lack of lexicality effect are also in line with current literature on svPPA and are likely explained by semantic impairment (Rogers et al., 2004). Interestingly, the novel result here is the absence of PsH effect in svPPA. This suggests that phonology is not correctly activated and that O-P mappings are also impaired in svPPA.
Keywords
