Objective. Although family, twin and adoption studies have suggested that lifetime major depression (MD) is a heritable condition, nearly all these studies have relied for the diagnosis on long-term human memory, which is fallible and potentially biased. Could the estimates of heritability of MD be biased by the well-demonstrated genetic influences on memory?

Methods. Both members of 858 female–female twin pairs from a population-based registry were personally interviewed at least three times over 9 years. The interview assessed a history of depressive onsets and recoveries in the last year to the nearest month. We examine heritability of MD using four recall intervals: last year, last 6 months, last 3 months and current month.

Results. Examining the occurrence of one or more depressive episodes across all three interviews, heritabilities of MD (95% CI) for the four time periods were: 0·41 (0·27–0·54), 0·41 (0·26–0·55), 0·35 (0·16–0·52) and 0·34 (0·11–0·55). These heritability estimates did not differ significantly from one another. A similar pattern was found if heritability was assessed for the number of interviews in which criteria for MD were met.

Conclusion. Modelling results suggest that the heritability of MD was not influenced by the duration of the required recall. Genetic influences on human recall do not appear to contribute substantially to estimated heritability of MD.