This pilot study developed a set of codes designed to capture the "nonaphasic" but characteristic discourse deficits that may be present following prefrontal cortex damage (PFCD). The codes were utilized based on narrative sample elicitation to investigate between-group differences in two study populations: patients with left, right, or bi-frontal PFCD and age and education-matched healthy comparison group participants. Narrative samples were coded on indices of content units, thematic units, story grammar features, and discourse errors, and analyzed using CLAN. Results of this study support the original deficit hypotheses. The coding schema demonstrated fair to good inter-rater reliability, stronger performances by the healthy comparison group across all four levels of analysis, and poorer performance overall on the retell phase than the tell phase. Qualitative analysis revealed relatively few discourse errors associated with the healthy comparison group, while various classic discourse errors were associated with the PFCD group.