Predicting discourse ability in TBI: Cognitive and communicative factors

Le, Karen

Abstract

The goal of the proposed research is to advance our knowledge of cognitive-communicative disorders following traumatic brain injury (TBI) by identifying the cognitive and communicative processes underlying narrative discourse ability. Using structural equation modeling, the proposed research will examine the role of working memory (WM) and inferencing in narrative discourse and test key assumptions posited by the Structure Building Framework (SBF), a cognitive model of normative discourse comprehension. Aspects of executive functions, such as WM, have emerged as potential substrates of discourse. A related factor is inferencing, an often diminished ability in TBI. Discourse production deficits are well-documented in the TBI literature, but data on discourse comprehension in TBI is scarce. The extent to which discourse comprehension predicts discourse production remains unclear. The lack of a discourse model to explain cognitive-communicative breakdowns in neurologic populations hinders the development of theoretically-based discourse interventions. The SBF is a potential candidate, but its assumptions need to be tested, in particular the positing of deployed cognitive processes as domain-general. First, we examine the contributions of WM and inferencing to narrative discourse ability. Second, we test the hypothesis that discourse relies on domain- general processes by using verbal and nonverbal WM tasks. Third, we investigate the extent to which narrative discourse comprehension accounts for narrative discourse production. Analysis 1 uses verbal and nonverbal WM updating and inferencing tasks to predict performance on two narrative discourse comprehension tasks. Analysis 2 uses the same predictors as Analysis 1 to predict performance on two discourse production tasks, story completeness and story grammar. The literature suggests that WM underpins both discourse modes. If WM is domain-general in discourse, we predict that verbal WM will not account for discourse comprehension beyond that of nonverbal WM and vice versa. This hypothesis will hold if performance on verbal and nonverbal WM is highly correlated and the amount of variance in discourse tasks explained by the verbal and nonverbal WM is equivalent. If inferencing contributes to discourse ability after accounting for WM, this suggests that inferencing also marshals other cognitive processes. In Analysis 2, we predict that WM will carry more of the variance in the content measure, story completeness, than inferencing whereas inferencing will carry more of the variance in the organizational measure, story grammar. Analysis 3 examines the prediction of discourse production from discourse comprehension. The magnitude of the contribution of discourse comprehension to discourse production reflects the dependence of production on comprehension. A greater magnitude of the relationship implicates that discourse production and comprehension likely operate in the same system while a smaller one would suggest that comprehension and production operate as parallel systems that draw upon the same cognitive processes but to differing degrees.

Public Health Relevance

The proposed research will contribute to our knowledge of cognitive-communicative disorders following traumatic brain injury by identifying the cognitive and communicative processes underlying discourse ability. The findings from this study will be instrumental in refitting current models of discourse or in developing new ones to better explain cognitive-communicative functioning in neurologic populations. A working model of discourse ability will provide a solid foundation for the development of meaningful and targeted discourse interventions grounded in theory.