The question of how to define a therapeutically adequate electroconvulsive therapy (ECT) has been under discussion since the early days of ECT. Although convention has asserted a demand for minimum seizure times, the complex electrophysiological conditions involved in developing a generalized seizure make it problematic for therapeutic efficacy of ECT to be linked only with seizure duration. Within the framework of an open clinical study of 40 patients, selected parameters of the ictal electroencephalogram (EEG) have now been examined with respect to differentiation between therapeutically effective and ineffective treatments. For this purpose a rating scale covering both quantitative and qualitative features of the ictal EEG was used. Although this study recorded no correlations between seizure duration and clinical improvement, correlations were established between clinical improvement, on the one hand, and the frequency of epileptic discharges and their slowing during the spike-wave phase as well as the stereotypy of the discharge or a "stable" pattern of rhythmic spike-wave or sharp wave complexes, on the other. The results suggest that several of these EEG parameters might be combined to form a marker for therapeutically adequate ECT, and that treatment might be controlled accordingly.