Project 1

Aims: To investigate the relationship between trajectories of maternal perinatal depression, and childhood neurocognition and social cognition.

Hypotheses: It is anticipated that more severe trajectories of maternal perinatal depression over time will be significantly associated with childhood deficits in neurocognition and social cognition in the study child.

Exposure variables: Maternal perinatal depression

Outcome variables: Social cognition (Theory of mind, locus of control, emotional recognition, social communication) and neurocognition (IQ, executive function, verbal comprehension, memory, attention, inhibition, reasoning)

Confounding variables: Family psychiatric history, mother's age at delivery, gestation, family ethnicity, parity, marital status, housing tenure, social class, disposable income, maternal education, urbanicity, smoking/drinking in and pre pregnancy, problematic partner relationship, maternal affective disorder (anxiety, suicidality), abuse of drugs, involvement with crime, child's gender, birthweight, breastfeeding in first year, childhood psychiatric disorder

Project 2

Aims: To investigate the relationship between trajectories of maternal perinatal depression and later psychotic-like experiences, at-risk mental state, and psychosis.

Hypotheses: It is anticipated that more severe trajectories of maternal perinatal depression over time will be significantly associated with the risk of later psychotic-like experiences, at-risk mental state, and psychosis in the study child.

Exposure variables: Maternal perinatal depression

Outcome variables: Psychotic-like experiences, at-risk mental state, psychosis

Confounding variables: Family psychiatric history, mother's age at delivery, gestation, family ethnicity, parity, marital status, housing tenure, social class, disposable income, maternal education, urbanicity, smoking/drinking in and pre pregnancy, problematic partner relationship, maternal affective disorder (anxiety, suicidality), abuse of drugs, involvement with crime, child's gender, birthweight, breastfeeding in first year, childhood psychiatric disorder

Project 3

Aims: To investigate whether the relationship between between trajectories of maternal perinatal depression and later psychotic-like experiences, at-risk mental state, or psychosis are mediated by cognition in childhood.

Hypotheses: It is anticipated that the association between more severe trajectories of maternal perinatal depression and the risk of later psychotic-like experiences, at-risk mental state, or psychosis in the study child will be stronger when the study child experienced cognitive deficits in childhood as well.

Exposure variables: Maternal perinatal depression, social cognition (theory of mind, locus of control, emotional recognition, social communication), neurocognition (IQ, executive function, verbal comprehension, memory, attention, inhibition, reasoning)

Outcome variables: Psychotic-like experiences, at-risk mental state, psychosis

Confounding variables: Family psychiatric history, mother's age at delivery, gestation, family ethnicity, parity, marital status, housing tenure, social class, disposable income, maternal education, urbanicity, smoking/drinking in and pre pregnancy, problematic partner relationship, maternal affective disorder (anxiety, suicidality), abuse of drugs, involvement with crime, child's gender, birthweight, breastfeeding in first year, childhood psychiatric disorder.

Planned analyses

We plan to examine trajectories of maternal depression over several time points using a latent class analysis. Maternal depression symptom data from pregnancy up to the study child at age 12 will be analyzed using group-based trajectory models in order to identify parents with chronic high levels of depressive symptoms. In Projects 1 and 2, we plan to run logistic regression to estimate odds ratios with a 95% confidence interval for the cognitive and psychosis-related outcomes. Multiple mediation analyses will then be carried out in order to explore the extent to which these associations are mediated by specific variables (eg. IQ, theory of mind). In order to examine the multivariate relationships in Project 3, we plan to carry out structural equation modeling. To overcome the issue of missing data, we plan to conduct sensitivity analyses using a range of auxiliary variables to carry out multiple imputations by chained equations.

We also plan to liaise with Jon Heron and Jonathan Evans in Bristol on their work with perinatal depression and trajectories.