To induce a seizure for electroconvulsive therapy (ECT), an electrical charge is delivered above seizure threshold. The means and criteria used to determine the electrical dosage are subject to debate. Nonetheless this is an important issue because effectiveness and side effects have been shown to be influenced by the electrical charge used. The objective is to review data available in the literature on seizure threshold and ECT and determine the eventual consequences for practical determination of stimulus dosing. A comprehensive review of the literature is based on the search of electronic databases (Medline, INSIT) and a manual search; 72 references out of a total of 96 selected were used for this review. Seizure threshold varies widely between subjects receiving ECT (600% mean variation), however a majority of subjects of all ages have a threshold below 150 mC. Only a few individuals have very high thresholds (400 to 800 mC). ECT has an anticonvulsive effect as threshold increases during a course of ECT. Many factors influence threshold and all are not known. Among those that have been documented are: the characteristics of the current used (longer stimulus duration with same dosage gives lower thresholds); electrode placement (bilateral gives higher thresholds than unilateral placement); age (explains 12 to 26% of threshold variance); gender (which inconsistently gives higher thresholds for males); and other factors such as anesthetic drugs, concurrent psychotropics, and some morphological characteristics. Different methods are used to determine an individually adapted dosage. Two are recommended: titration and age. The age method is based on the fact that age is an important factor influencing threshold. The titration method is based on the observation of a very important variation in threshold between individuals that is not explained by age. We discuss the pros and cons of each method.