Despite ongoing controversy, there has never been a large-scale, prospective study of the cognitive effects of electroconvulsive therapy (ECT). We conducted a prospective, naturalistic, longitudinal study of clinical and cognitive outcomes in patients with major depression treated at seven facilities in the New York City metropolitan area. Of 751 patients referred for ECT with a provisional diagnosis of a depressive disorder, 347 patients were eligible and participated in at least one post-ECT outcome evaluation. The primary outcome measures, Modified Mini-Mental State exam scores, delayed recall scores from the Buschke Selective Reminding Test, and retrograde amnesia scores from the Columbia University Autobiographical Memory Interview-SF (AMI-SF), were evaluated shortly following the ECT course and 6 months later. A substantial number of secondary cognitive measures were also administered. The seven sites differed significantly in cognitive outcomes both immediately and 6 months following ECT, even when controlling for patient characteristics. Electrical waveform and electrode placement had marked cognitive effects. Sine wave stimulation resulted in pronounced slowing of reaction time, both immediately and 6 months following ECT. Bilateral (BL) ECT resulted in more severe and persisting retrograde amnesia than right unilateral ECT. Advancing age, lower premorbid intellectual function, and female gender were associated with greater cognitive deficits. Thus, adverse cognitive effects were detected 6 months following the acute treatment course. Cognitive outcomes varied across treatment facilities and differences in ECT technique largely accounted for these differences. Sine wave stimulation and BL electrode placement resulted in more severe and persistent deficits.