Validation of linguistic analyses in narrative discourse after traumatic brain injury

Abstract

Narrative discourse tasks are a common feature of assessment and research after traumatic brain injury (TBI) and other types of brain damage. Although stimulus materials and analysis methods have been developed from a variety of theoretical perspectives, many do not challenge cognitive-linguistic skills sufficiently to highlight individual difficulties in assessment after TBI. This study employed a complex story recall task and it aimed to develop analysis methods that were sensitive to differences in recalled narratives and which could be validated against the perceptions of external raters. Twenty TBI adults and 20 matched control subjects were tested and their narratives analysed in terms of six measures (T-units, Noun:Pronoun Ratio, Scenes, Errors, Comments and Tentative Statements). Two groups of raters (one professionally trained, the other not) judged the transcribed narratives on ratings of Content and Clarity. Multiple regression analyses established that two of the linguistic measures in combination predicted ratings of Content, while a further three in combination predicted ratings of Clarity. Cut-off scores were established in order to categorize the TBI and control subjects' performance. The validated measures provide the foundation for analyses of complex narrative as a means of assessment after TBI.