Changes in Discourse Production Following Left Hemisphere Stroke



Journal Title

Journal ISSN

Volume Title



Discourse is any unit of connected speech longer than a sentence, organized sequentially and logically, to effectively communicate a group of ideas to a listener (Kong, 2016). Discourse can become impaired following injury to the brain, resulting in either acute impairments that get better over time, or residual chronic communication impairments (e.g., aphasia). This study used a modified multi-level analysis developed by Marini and colleagues (2011, 2012, 2014, 2015) comprised of microlinguistic and macrolinguistic variables to examine whether acute left hemisphere stroke patients experience deficits in discourse and how those deficits change over time as reorganization and recovery takes place. A picture naming task and a narrative retell task were administered to 16 patients at bedside acutely (2-7 days post stroke), then again sub-acutely (1-3 months post stroke), and the results of the discourse measures were compared with the data of 14 control subjects. Results indicated that 69% of our population had some type of discourse deficit acutely, either microlinguistic or macrolinguistic, compared with the data from controls. For our population at the acute stage, 69% of subjects showed deficits on macrolinguistic aspects of discourse, whereas 50% of subjects showed deficits on microlinguistic aspects of discourse, indicating that macrolinguistic deficits were more prevalent for our cohort. A subset of subjects demonstrated significant macrolinguistic deficits without the presence of microlinguistic deficits. These results support previous aphasia research stating that aphasia batteries may not be sensitive enough to detect subtle discourse deficits in this population. Results also indicated a relationship between lexical retrieval (picture naming) and cohesion for the acute population, but correlations did not hold at the sub-acute time point due to subjects performing at ceiling on the naming measure. These data are supportive of the need for multilevel analysis to examine changes in discourse and show that discourse deficits may be missed using traditional aphasia batteries which detect primarily microlinguistic variables.



Discourse, Stroke