Background: Studies examining the discourse abilities of adults who have suffered traumatic brain injuries (TBI) have revealed that although these individuals display relatively normal language skills on traditional aphasia tests, they demonstrate varying levels of impairment in the coherence, cohesion, and informational content of their extended verbal output. Discourse studies with the TBI population have varied relative to the type of discourse task as well as the level or approach of analysis. One approach to discourse analysis has involved examination of the cognitive functions distinguishing macrostructural and microstructural discourse processing. Macrostructural analyses are global analyses focused at the level of the entire text, dealing with large supra-sentential discourse units. Microstructural analyses are concerned with smaller structural elements in a text and the relations between these elements; the focus is at the local level of the word or sentence. Aims: In the current investigation, three analyses were undertaken to examine descriptive discourse production of five high-functioning TBI adults (three males, two females) who had suffered brain injury as the result of motor vehicle accidents. The first analysis, thematic coherence, is a macrostructural analysis, whereas the other two analyses, cohesion and lexical errors, are microstructural analyses. Methods & Procedures: Two topics, family description and work experience, were used to elicit discourse; performance was compared between the language samples. Each participant's performance was reviewed individually and findings were compared to the mean performance of a group of 15 normal functioning young adults. For thematic coherence, global and local coherence were evaluated and scored separately on different 5-point rating scales. Cohesion analysis involved identification of three types of cohesive ties (closed-class appropriate, open-class appropriate, incomplete cohesion) with scores computed as proportions of total spoken words. Lexical error production included identification of verbal paraphasias and indefinite terms, also computed as proportions of total words spoken. Outcomes & Results: The results revealed that mean coherence ratings for each TBI participant were consistently lower than those observed for the control group, regardless of coherence type (local, global) or discourse topic (family, work). Furthermore, the TBI participants demonstrated greater difficulty with global than local coherence and showed more performance variability among participants in global as compared to local coherence. However, all five TBI adults displayed minimal deficits in lexical production and cohesion. Conclusions: Overall, the findings revealed dissociation of microlinguistic and macrolinguistic cognitive functions in the TBI participants, thus providing support for the distinction between and independent organisation of microstructural and macrostructural discourse components.