You are viewing the SNL 2017 Archive Website. For the latest information, see the Current Website.

Poster A25, Wednesday, November 8, 10:30 – 11:45 am, Harborview and Loch Raven Ballrooms

Semantic control does not relate to domain-general components of executive function.

Curtiss Chapman1, Randi Martin1;1Rice University

Theories of semantic memory sometimes claim a distinction between stored semantic representations and mechanisms for accessing those representations. For individuals with aphasia, researchers have claimed multi-modal (i.e., both verbal and nonverbal) semantic deficits (i.e., semantic aphasia, SA) arise from damage to the semantic access mechanism, called ‘semantic control’ (Jefferies & Lambon Ralph, 2006). Supporting this claim are findings showing correlations between semantic and executive function (EF) performance. However, these correlations rely on small samples and use of global executive tasks (e.g., Raven’s Progressive Matrices; WCST: Wisconsin Card Sort Task) that prevents determination of which aspect of EF is related to semantic deficits. Our goals were to 1) validate the correlations between semantics and EF by combining data from the published SA literature, 2) to explore which component of EF (updating, shifting, inhibition) is related to semantic access. We identified 20 unique patients from the semantic control literature and examined correlations between semantic and executive measures. Sample sizes varied depending on which patients completed which tasks. To address specific components of EF, we tested a new group of 9 SA patients on semantic tasks and tasks tapping specific aspects of EF. Semantic tasks included: picture and word versions of the Camel and Cactus Test of associative relations; environmental sound–picture matching; auditory word-picture matching with within-category distracters; and picture naming. Executive tasks included: 1) updating (phonological working memory; pWM): digit and word span, N-back, 2) shifting: cued task-set shifting, a trail-making task, 3) inhibition: verbal and non-verbal Stroop, picture-word interference, 4) global: WCST and Raven’s. For SA patients from the published literature, the semantic composite correlated significantly (p < .05) with Raven’s (r(18)=.44) and an executive composite (Raven’s & WCST; r(10)=.73) and marginally with a pWM composite (digit span forward & backward; r(18)=.40, p=.08). When partialling out the influence of pWM, the executive composite remained predictive of semantics (r(9)=.68, p=.02), whereas Raven’s did not (r(17)=.38, p=.11). For our 9 SA patients, the semantic composite did not correlate significantly (ps > .58) with the shifting composite (r(7)=-.17), Raven’s (r(5)=.15), WCST (r(6)=.15), or an EF composite containing all EF measures (r(5)=.10). (An inhibition composite was not calculated, as inhibition measures were uncorrelated). There was a moderate correlation with the phonological WM composite, (r(6)=.54), which was non-significant (p=.17) with this sample size. Using the published data, we confirmed significant correlations between semantics and performance on global EF tasks. For our new group, we failed to replicate the correlations with global EF and did not find correlations with specific aspects of EF. Some prior case study results suggest that semantic control depends on EF specific to semantics, rather than on domain-general executive abilities (Hoffman et al., 2013). If so, the lack of convergence between the findings from the published literature and ours may result from a greater overlap of domain-general EF deficit with the critical semantic EF deficit in the published cases than in our sample. Future work will address whether a specific semantic EF deficit can be demonstrated in our patients.

Topic Area: Language Disorders

Back to Poster Schedule