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

Poster A31, Wednesday, November 8, 10:30 – 11:45 am, Harborview and Loch Raven Ballrooms

The role of executive functions in anaphora resolution in non-fluent variant Primary Progressive Aphasia

Eleni Peristeri1, Ianthi-Maria Tsimpli2, Kyrana Tsapkini3;1Department of English Language and Liguistics, Aristotle University of Thessaloniki, Thessaloniki, Greece, 2Department of Theoretical and applied linguistics, University of Cambridge, UK, 3Department of Neurology, Johns Hopkins University, Baltimore, USA

Syntactic processing theories suggest that executive functions (EF) play a critical role in sentence parsing and comprehension (Novick, etal. 2005). For individuals with Primary Progressive Aphasia (PPA), verbal EF deficits have been claimed to reflect the patients’ language deficits. However, findings supporting this claim are still inconclusive (Wicklund etal., 2004). One possible reason for the controversial results is that we may be missing some important EFs that may affect language processing, such as the attentional demands of verbal and non-verbal tasks. The present study examines for the first time the global-local attentional dimension in nfv-PPA patients’ performance in sentence parsing. We employed an anaphora choice task that allowed us to isolate effects of global language context as well as local parameters of referential processing. The performance of the same patients in the referential processing task was correlated to their performance on a visual attention task tapping the ability to shift attention from the global to the local level as well as in other EF tasks (e.g., digit span). Data were collected from 8 Greek-speaking patients with nfv-PPA (Mean age: 62;9) along with 8 age-matched healthy controls. The anaphora resolution task was a self-paced picture-selection task during sentence listening, in which participants had to pick an antecedent for an ambiguous subject pronoun that was either a null or an overt pronoun. The candidate antecedents could be either the syntactic subject or the object of the sentence, which was closer to the critical pronoun than the subject. Co-indexing null pronouns to syntactic subjects in null subject languages has been claimed to be subject to global discourse constraints, while linking the overt pronoun to the object requires integration of local discourse cues. Executive abilities were assessed with the global-local task (Navon, 1977). Participants also completed a Digits Backward task to measure their working memory capacity. Given the sample size, correlation analyses were used to examine the relationship between these two EF measures and performance in anaphora resolution. One-way ANOVA analyses showed that for controls, object preference rates were significantly higher when the pronoun was overt (p=.04), while subject preference was stronger in null pronoun trials (p=.05). This suggests that controls could integrate both global and local discourse constraints during anaphora resolution. On the other hand, PPA patients systematically picked the object across all trials irrespective of whether the pronoun was null or overt (p<.001). This finding reflects PPA patients’ preference to resolve the pronoun to the closest antecedent, rather than an interpretive preference guided by the form of the pronoun. In controls, subject and object referent rates were correlated with the cost (in RTs) they have experienced when switching attention from the global to the local level and vice versa, in the global-local task (r=.75), while PPA patients’ object referent rates were negatively correlated with their performance in the Digits Backward task (r=-.90). Taken together, PPA patients’ preference for the object referent may be attributed to working memory limitations, rather than to global and local attention parameters that regulated controls’ referential preferences.

Topic Area: Language Disorders

Back to Poster Schedule