RESUMO
OBJECTIVE: To assess the sensitivity of traditional neuropsychological tests and of a behavioral inventory of executive disorders in a large sample of patients with chronic severe traumatic brain injury. METHODS: A total of 112 patients were compared with 780 healthy controls from a larger database. The GREFEX battery included 7 widely used tests and the Behavioral Dysexecutive Syndrome Inventory (proxy rating). A previously described statistical methodology was used, controlling for age, education, and gender. Summary scores were computed and performance was dichotomized on the basis of 5th percentile cutoffs from controls' z scores. RESULTS: The frequency of cognitive impairment was high (55.4%) but lower than that of behavioral changes (81.5%). Double dissociations were observed between cognitive and behavioral assessments. Behavioral changes exhibited larger effect-sizes as compared with cognitive impairments. Logistic regression analysis showed that 3 cognitive tests (verbal fluency, Stroop reading, and Trail Making Test-B) and 3 behavior z scores (hypoactivity, anticipation, and hyperactivity) best discriminated patients from controls. CONCLUSION: Behavioral changes were more frequent and severe than cognitive deficits, at least as assessed with traditional testing. The present results also suggest that a shortened battery may provide a rapid screening method with reasonable sensitivity to detect deficits of executive functions in patients with severe traumatic brain injury.