Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Profiles of narrative discourse impairment in acute right hemisphere stroke
Poster Session A, Wednesday, September 30, 11:00 am - 1:00 pm, Wangari Maathai
Emilia J Cichocki1, Lokesha Pugalenthi1, Tatiana T Schnur2; 1Rice University, 2UTHealth Houston
Right hemisphere damage (RHD) from stroke can cause impairments in narrative discourse that significantly affect communicative competency. In chronic RHD (>6 months after damage), patients commonly present with macrolinguistic deficits characterized by reduced organization, relevance, or informative content, while performing similarly to neurologically healthy controls on microlinguistic (lexical and syntactic) measures. However, chronic performance may reflect functional reorganization and compensatory strategies that emerge over time. Studying language in the acute stage of RHD (<1 week after damage) therefore provides an opportunity to determine the immediate contribution of the right hemisphere to discourse before substantial recovery-related adaptation occurs. Yet, few studies examine narrative discourse during the acute stage of RHD, leaving the communication impairments from acute RHD underexplored. Here, we asked whether acute RHD disrupts connected speech primarily at the macrolinguistic level, as suggested by chronic RHD evidence, or whether deficits extend to microlinguistic aspects of discourse production. We elicited narrative discourse using a Cinderella story retell task in 81 participants tested after acute RHD undifferentiated by stroke location or severity, with a median testing time of 3 days post-stroke. We derived 16 features reflecting information content, coherence, speech rate, lexical properties, and syntactic structure from transcribed responses using an automated linguistic analysis pipeline and computerized Quantitative Production Analysis (Fromm et al., 2021). To construct independent feature combinations capturing discourse variability, we performed principal component analysis on the derived feature set. We used hierarchical clustering of resulting participant component scores to determine unique discourse profiles, then compared mean component scores in each profile with the predicted scores of 27 neurologically healthy controls to identify impairments within clusters. Principal component analysis returned six independent components with eigenvalues >1, explaining 78.5% of variance. Components represented quantity of speech (productivity, 23.8% of variance), type and function of words produced (lexical selection, 13.0%), complexity of grammatical structures (syntactic complexity, 11.6%), global and local consistency of the narrative (coherence, 10.8%), proportion of repairs and repetitions (fluency, 10.6%), and amount of included propositions (information density, 8.7%). We identified five language profiles through hierarchical clustering: reduced noun production (impaired lexical selection) with low productivity (32% of participants); no impairment (30%); reduced noun production with low productivity, information density, and coherence (26%); low productivity and information density (8%); and low productivity and coherence (4%) (t’s < -3.5, p’s < 0.001). In sum, acute RHD produces heterogeneous connected speech deficits captured by both microlinguistic and macrolinguistic components of narrative discourse. Consistent with chronic RHD findings, many patients showed macrolinguistic impairments in coherence and information density. Others showed disruptions involving speech quantity and lexical selection, suggesting that microlinguistic impairments from acute RHD are not syntactic in nature. Thus, by examining patients before substantial functional reorganization and compensatory strategies develop after stroke, we provide evidence that the right hemisphere contributes not only to macrolinguistic discourse characteristics, but also to select word-level aspects of narrative production.
Topic Areas: Disorders: Acquired, Meaning: Discourse and Pragmatics