Using discourse to model language use after right hemisphere stroke - ABSTRACT Almost 80% of right hemisphere stroke survivors suffer brain damage (RHD) that results in disabling language use impairments. These impairments are not recognized, evaluated, or treated because intelligibility, grammar, and sentence structure are preserved. However, the language used often lacks context-specific content needed for communication. Survivors of RHD produce language that is tangential, inappropriate, and self-centered, and contains inappropriate comments, humor, and few questions. The impairments are most apparent during discourse and negatively affect social interactions, employment, and quality of life for RHD survivors and their families. Despite the significant impact, RHD language impairment characteristics are relatively undefined, with little consideration of the impacts of race and sex; and there are no dedicated discourse-level diagnostic tools for this neglected population. The long-term goal of this research is to improve patient outcomes by establishing a foundation for the development of an evidence-based assessment and treatment approaches for RHD language use impairments. Our preliminary work demonstrated that disruptions in language use are fundamental to RHD communication impairments and may be mediated by sex and race. The project objective is to define RHD patients’ discourse-level language use and use this information to develop a predictive model of discourse production with considerations of race and sex, thus advancing the field. Using RHDBank, an internationally shared, multimedia database for the study of RHD language impairments using discourse, and RHDBank GrandRounds, an online educational complement, we will comprehensively characterize, quantify, and compare the discourse production of RHD survivors to controls. The central hypothesis is that RHD language use impairments can be differentiated from controls using discourse for clinical purposes. The proposed work will leverage the PI’s previous work using the RHDBank Protocol to reveal the types, degrees, and patterns of RHD language use in discourse. To achieve our goals, we propose the following aims: Aim 1) Quantify the effects of RHD on language use with RHDBank Protocol discourse tasks (First-Encounter Conversation Task, Question Production Task, Cat Rescue Task, & Cookie Theft Task) that are important to daily communicative interactions. Aim 2) Determine the effect(s) of race and/or sex on RHD language use behaviors using demographic identifiers from Duke Health electronic medical records. We will use study data to augment the RHDBank Grand Rounds stroke education samples and encourage community participation in future studies. Aim 3) Develop a predictive model of language use in discourse tasks. Using logistic regression analysis, we will create a prototype for a discourse-level diagnostic tool that can provide a disordered versus non-disordered designation among RHD survivors. The proposed study will provide objective methods to identify RHD language use impairments; guide RHD treatment; shift clinical delivery processes; reduce disparate care regarding race, sex, and stroke hemisphere; and improve health outcomes and quality of life after RHD for survivors and their families.
