Intense vagal discharge often follows stimulus application during electroconvulsive therapy (ECT). Related periods of asystole during ECT have been reported sporadically in psychiatric journals, but to date not in the anesthesia literature. We report here two cases of prolonged asystole that occurred in our facility in spite of the fact that published suggestions for its prevention were followed. With careful monitoring of these patients--including echocardiography for one patient--we document the onset of asystole at the exact time of ECT stimulus application. With these data, we discuss why asystole is likely to result from a direct central pathway rather than via a baroreceptor reflex, and discuss a neuroanatomic pathway potentially responsible for our findings. We also demonstrate that high-dose atropine (0.8 mg) can effectively prevent most cases of asystole in susceptible patients, and that administration of esmolol following cessation of seizures effectively reduces the elevated heart rate without causing asystole or bradycardia.