Work

Active Dependency Formation in Agrammatic Aphasia: Online Structural Prediction and Working Memory Processes

Public

Downloadable Content

Download PDF

Sentence comprehension requires both syntactic prediction and working memory processes. Evidence suggests that both are impaired in agrammatic aphasia, and can independently contribute to comprehension breakdowns, but the relationship between them, especially during real-time sentence comprehension remains unclear. Two EEG studies investigated on-line processing of sentences with non-local long-distance dependencies in nine individuals with agrammatic aphasia and groups of neurologically unimpaired controls. Study 1 examined syntactic prediction and working memory storage using an animacy-mismatched paradigm, with ERP measures corresponding to each process; respectively N400 for prediction, and Sustained Anterior Negativity (SAN) for online working memory storage. This study addressed the questions (1) whether agrammatic individuals actively predict an upcoming licensing element during long-distance dependency processing, and (2) whether participants show real-time evidence of storing the unresolved dependency in working memory. Results showed evidence of a timely SAN but delayed N400, indicating that agrammatic individuals show normal-like working memory processing but delayed structural prediction. Study 2 further investigated online working memory storage by manipulating processing complexity, and addressed whether increased processing demands affect online working memory processing. Participants performed a word-by-word sentence reading task with short- vs. long- dependency conditions, again using SAN as the online working memory measure. Results revealed that while the agrammatic group showed SAN in both short- and long- dependency conditions, the effects in the long-dependency condition were weaker, with evidence of memory decay, suggesting that working memory processing in agrammatic aphasia is more vulnerable when processing demands increase. These findings indicate a top-down parsing strategy in agrammatic aphasia where working memory resources are used to facilitate active linguistic prediction during dependency formation and provide support for the deficient lexical-semantic integration account of asyntactic comprehension in aphasia. This research advances our understanding of the linguistic and cognitive mechanisms underlying agrammatic sentence processing, and provides directions for further studies.

Creator
DOI
Subject
Language
Alternate Identifier
Keyword
Date created
Resource type
Rights statement

Relationships

Items