We thank Moreira-Almeida & Junqueira de Souza for their interesting correspondence on our paper. Indeed, we regard childhood environmental factors as highly important to the intergenerational pathways for the transmission of depression. Nevertheless, maternal depression during pregnancy in itself has been identified as a significant risk factor for offspring depression, not only in our sample, but also in other samples of varying demographics and size. ^{Reference Pearson, Evans, Kounali, Lewis, Heron and Ramchandani1,Reference O'Donnell, Glover, Barker and O'Connor2}

There are plausible and documented mechanisms linking a mother's depression in pregnancy with her child's increased vulnerability to experiencing maltreatment; namely, changes to the mother–child attachment relationship, maladaptive caregiving behaviours, interparental conflict and increased offspring reactive temperament. ^{Reference Lereya and Wolke3} Such mechanisms likely operate by compromising levels of care and protection afforded by a mother, as well as directly affecting stress resiliency in her developing child, thereby increasing her child's vulnerability to being exposed to, and experiencing, episodes of maltreatment. Depression is a disorder with a recurrent course, ^{Reference Eaton, Shao, Nestadt, Lee, Lee, Bienvenu and Zandi4} thereby meaning the likelihood of depression after birth is elevated following an episode during pregnancy. Additionally, the temporal precedence of antenatal depression to childhood maltreatment and further maternal depression after birth adds to the logic as to why antenatal depression should be considered as a primary risk factor in the intergenerational transmission of depression, and the aforementioned childhood adversities as mediators to this trajectory.

As Moreira-Almeida & Junqueira de Souza correctly highlight, in our multiple hierarchical regression models, maternal depression during pregnancy was not found to predict offspring adulthood depression in the context of childhood adversity factors. A major limitation of inferring causality from such multiple regression models is the issue of multiple confounding, particularly in the case of correlated factors with temporal sequence (i.e. maternal depression during pregnancy with (i) offspring child maltreatment (OR = 2.4), and (ii) maternal depression during offspring childhood (OR = 4.8)). Mediation analysis provides an analytic strategy which not only allows for the quantification and evaluation of a mediated effect, but also reduces issues of confounding through minimising the overloading of regression models with multiple confounding variables. ^{Reference MacKinnon, Lockwood, Hoffman, West and Sheets5} Indeed, when we applied mediation analysis to our data to understand better the trajectory of effect for the intergenerational transmission of depression, we found that maternal depression during pregnancy indirectly predicted offspring depression in adulthood through elevated exposure to child maltreatment. Our conclusion is therefore that maternal depression during pregnancy is a primary risk factor for the intergenerational transmission of depression, and represents a unique time point for intervening to break this intergenerational cycle. We believe that prioritising pregnant women with depression for psychological intervention will not only help to stem the cascade of depression from one generation to the next, but also has the potential to reduce rates of childhood adversity where risk is augmented by antenatal depression.