Aphasic comprehension: conflict resolution and short-term memory

Veterans Health Administration, Decatur PA

Investigators

Abstract

DESCRIPTION (provided by applicant): Sentence comprehension deficits are ubiquitous in persons with aphasia although they can be mild or severe and variable over short or long periods of time. It is difficult to explain the variability and differences in severity within the construct of impaired linguistic-specific representations, rules or algorithms. Among other cognitive mechanisms, short-term memory (STM) and conflict resolution (CR) are prominent among those proposed to account for these difficulties. STM is defined as the process underlying the maintenance of activated linguistic information over a short period or delay. CR is defined as an executive attention process underlying the inhibition of prepotent or interfering stimuli or response selections. It is well established that persons with aphasia exhibit deficits in STM and demonstrate impairments of executive attention functions, most prominently, CR. However, research in aphasia has produced conflicting results as to whether these cognitive functions cause, contribute to, co-occur or are unrelated to sentence-level language comprehension processing in healthy persons or to their deficits in persons with aphasia (PWA). Most studies have relied on small subject samples and relatively simple correlational analyses to support or refute the relationships among the cognitive and linguistic abilities in and impairments in PWA. In order to understand the mechanisms underlying sentence-level listening and reading comprehension and determine whether relationships among these cognitive and linguistic exist, it is essential to study large participant samples that support the use of more sophisticated statistical procedures. This study will investigate the relationships among the two well- defined cognitive constructs of STM and CR, and the three linguistic domains of phonology, semantics and syntax that are involved in sentence comprehension. It will address a fundamental question relating to the cause of language disorders in aphasia (are these disorders caused by deficits in linguistic-specific processes, by other cognitive functions such as STM and CR or are they simply coexisting disorders?). This study will test the extent to which STM and CR contribute, beyond basic language processes (LP), to online and off-line measures of sentence processing of structures that are assumed to engage these cognitive functions. One hundred and twenty PWA who meet the strict criteria for enrollment will participate in this multi-site study. Each participant will complete a battery of two to three tasks for each language domain, designed to engage primarily phonological, semantic, or syntactic LP. Participants will also receive two to three independent measures of STM in each of these domains, and three measures designed to engage CR. One of the key challenges to disentangling the unique and shared contributions of LP, STM and CR is that any task intended to measure a specific process, be it a language process or a memory or executive process, draws to some extend on the other domains through stimulus material and task demands - the task impurity problem in cognitive science. Because these functions are not directly observable and are latent variables, the study will use a structural equation modeling (SEM) approach to the analysis. SEM is used to make explicit and test hypotheses about inter-relationships between latent variables as well as their relationships to the observed measures used to assess them. Outcome measures reflect accuracy and reading times at critical regions in sentences hypothesized to draw on STM and CR, at the final word of the sentence (wrap-up), and during response planning and execution. The latent variables will be used to estimate the relationship among their underlying constructs and the outcome measures. Specific latent variable models are designed for individual research questions.