Objective: Transient cognitive impairment during electroconvulsive therapy (ECT) can be a reason to discontinue ECT in depressed elderly patients. We hypothesized that both white matter hyperintensities and medial temporal lobe atrophy contribute to transient cognitive impairment during ECT. Methods: In 81 elderly patients with depression, magnetic resonance images (MRI) were recorded before ECT. We rated white matter hyperintensities (WMH) with the Age-Related White Matter Changes scale (ARWMC). Cognitive impairment during ECT was measured weekly with the Mini Mental State Examination (MMSE), 2 days after each session. Results: The mean MMSE score at baseline for all patients was 25.5 points, the lowest MMSE score during ECT was 23.3 points, and the mean MMSE score after ECT was 26.3 points. Stratification for the ECT method showed no significant difference in the lowest MMSE scores of patients with or without WMH, receiving unilateral ECT (22.5 points versus 23.9 points). There was a difference in the lowest MMSE scores in patients who switched from unilateral ECT to bilateral ECT (18.7 points in patients with WMH versus 22.0 points in patients without WMH). Conclusion: Depressed elderly patients with WMH who receive bilateral ECT are at increased risk of transient cognitive impairment. Our findings show, however, that cognitive impairment improves when ECT is continued. This implies that ECT does not have to be discontinued when patients experience transient cognitive impairment during ECT.