Predictive inference comprehension in adults with traumatic brain injury (TBI): The effects of salience and working memory

Abstract

Objective: The purpose of this study was to investigate the impact of salience on the comprehension of predictive inferences in adults with traumatic brain injury (TBI), by increasing the visual salience of the predictive sentence. This study also investigated whether a relationship existed between performance on a predictive inferencing comprehension task and working memory for this population. Increasing the salience of a crucial sentence in the predictive inferencing task may lead to better memory for the inferred information within the focused portion of the text (Gernsbacher & Jescheniak, 1995; Parkhurst, Law, & Niebur, 2002). Method: Six participants with TBI and six non-brain injured peers (NBI) took part in the study. Each participant was administered an inference comprehension task which consisted of a series of 55 stories. Each story incorporated one of five conditions: 1) a Recent salient condition (inferred information immediately preceded the comprehension question and was visually salient); 2) a Recent non-salient condition (inferred information immediately preceded the comprehension question but was not visually salient); 3) a Distant salient condition (inferred information occurred early in the story and was visually salient); 4) a Distant non-salient condition (inferred information occurred early in the story and was not visually salient); and 5) a Control condition (no inferred information in the story). In addition there were 20 filler stories. The predictive sentence was bolded in half the stories in order to increase the visual salience of the stimuli. In addition, a measure of working memory span (Lehman-Blake & Tompkins, 2001) was administered. Results: A significant main effect was found for Group, F(1,11) = 7.6, p= 0.019, with adults with TBI performing more poorly than matched controls. A significant main effect was also found for Condition, F(3,33) = 3.159, p = 0.038, with all participants performing more poorly in the Distant non-salient condition. No statistically significant interaction between Group x Condition was observed, F(3,33) = 0.469, p =0.706. Post-hoc comparisons revealed that all participants performed more poorly in the non-salient condition when the storage load was high (distant non-salient condition). Significant correlations were found for working memory span and the Distant salient condition (r =0.677, p < 0.05) and Distant non-salient condition (r = 0.646, p < 0.05). Conclusion: The results have both theoretical and clinical implications. Theoretically, the role of attention in working memory is of interest in language comprehension (e.g. Montgomery, Evans, & Gillam, 2009). This study may further contribute to studies of allocation of attention using increased salience to enhance comprehension. Clinically, the use of enhancing the salience of key information is a practical strategy that can be employed.