There has a great deal of research that has shown that maternal depression has a long-term effect on the child. Longitudinal studies have also shown that maternal anxiety can affect the child, including recent studies from ALSPAC that have shown that maternal anxiety during pregnancy can have a long term effect on the child.

There is considerable interest in the idea that pregnancy and the postnatal period represent a "critical period" of development for the mother-infant relationship and for the infant. If the "critical period" idea is correct, then maternal mental health problems that occur during this period should be particularly harmful to the infant compared to maternal mental health problems that occur at other periods of the child's life. Other studies, however, have shown that chronic and recurring forms of maternal mental health problems have a stronger and more lasting effect than "critical periods" of maternal poor mental health.

In this proposal, we aim to investigate longitudinal PATTERNS of maternal anxiety. A patterns approach to mental health may be helpful in understanding what kinds of individuals struggle most, and if there are particular patterns or timepoints that are particularly important for child outcomes.

We expect that within mothers who experience anxiety, the pattern of anxiety across time will differ. We will use newer statistical procedures to model these different groups, or classes, of patterns. For example, some mothers may have chronic anxiety, some may have recurring patterns, and some may have episodic patterns. Other studies of patterns or trajectories of mental health problems (primarily depression) have detected low, high, decreasing and increasing patterns across time. Once we have established a well-performing model of maternal anxiety, we want to assess whether there are individual factors that predict the type of pattern of anxiety mothers have. For example, are mothers who live in poverty more likely to experience recurring or chronic patterns of anxiety? We also want to see if the pattern of anxiety varies based on other factors that might be happening for mothers. For example, we expect that stressful life events and depression may vary across time with anxiety, although one may precede the other in time. We will use lagged analyses to examine these effects. We also think that some baseline factors might change how closely life events, anxiety and depression co-vary. Women who have a history of abuse, for instance, may have a stronger link between life events and anxiety over time than those who don't have that history. Lastly, and importantly, we want to see if different patterns of anxiety predict child outcomes. We will use cross-lagged analyses to examine if there are reciprocal effects of child behaviours on maternal anxiety, and vice versa.