This study examined the ictal electroencephalographic (EEG) characteristics of four forms of electroconvulsive therapy (ECT) known to differ in efficacy. Previously, we demonstrated that titrated, low-dose right unilateral ECT reliably produces generalized seizures of adequate duration, but is remarkably weak in antidepressant effects. Using a new rating scale, we found that specific features of the ictal and immediate postictal EEG varied significantly with ECT stimulus intensity and electrode placement. The low-dose right unilateral condition differed from more effective forms of ECT in having the longest polyspike phase duration, averaging twice that of the other conditions; it was also the condition least likely to manifest bioelectric suppression immediately following seizure termination. In contrast, high-dose bilateral ECT--a treatment with particularly rapid antidepressant effects--resulted in the greatest peak slow-wave amplitude in both hemispheres. Total seizure duration did not differ among the four treatment conditions. These findings indicate that seizure duration is not a useful marker of therapeutic efficacy, and instead provide preliminary evidence that other features of the EEG may be more useful markers of treatment adequacy.