✓ Results of multimodality evoked potential (MEP) studies recorded from 100 comatose patients soon after severe head injury were analyzed prospectively, using a previously established grading system, to assess the prognostic value of MEP's with respect to patient outcome, to evaluate the effect of clinically relevant sequelae of head injury on the prediction of outcome by MEP's, and to describe time to clinical recovery as a function of initial MEP grade. Graded MEP's, when recorded in the first few days after head injury, could predict patient outcome at 1 year with approximately 80% accuracy. Exclusion from the analysis of patients who died from causes unrelated to the brain and those with severe systemic complications that occurred after the evoked potentials were recorded improved the accuracy of outcome prediction to nearly 100%. The presence of a mass lesion requiring surgery reduces the probability of good to moderate outcome for a given MEP grade group by approximately 25% to 40% from that seen in patients without mass lesions. The clinical outcome predicted shortly after head injury by MEP grades may not be realized for many months. Patients with mild MEP abnormality (Grade I or II) generally reach their outcome by 3 to 6 months, whereas those with more severe deficits (Grade III) may not show improvement for at least 1 year.