Cohorts like ALSPAC typically collect data on their participants over several years, but since data collection is usually both expensive and burdensome these data collection events tend to take place every few years, measuring or recording information at a particular instance in time e.g. via questionnaires or clinic visits. Hence, these data contain a limited amount of information on phenotypic variability across the life-course, and restricts the research questions that can be asked using these data. There is much more scope to exploit existing and emerging technologies to collect data ‘continuously’ over the longer term in cost-effective and less burdensome ways.

Digital health devices have been successfully used to collect data on specific traits over a number of days (e.g. physical activity measured with accelerometers), but these devices tend to each focus on particular traits such that collecting data in this way is expensive (having to buy specific devices to collect specific phenotypes), and many types of phenotypes do not lend themselves to this type of data collection, in particular, those that can only (currently) be collected via self-report. Recent advances in artificial intelligence and voice recognition technologies means it is now feasible to use voice-based systems to collect self-reported data continuously over several days or weeks in a less burdensome way. However, to date, voice-based data collection has not been used in epidemiology.

A second potentially valuable source of data comes from our pervasive use of the world wide web (the ‘web’). ALSPAC has included items in questionnaires (e.g. “Have you sought help or advice regarding your sex life from the internet in the last year?”), but collecting web usage information passively using a technological approach over a potentially long period of time (weeks, months or even years), has the potential to provide a very large and currently untapped source of health-related information, if collected in ALSPAC.

In this study we aim to assess feasibility and acceptability of a voice-based approach to data collection and passive collection of web usage data. We then plan to collect these data in ALSPAC participants.

Only a small proportion of pregnant individuals will continue to smoke during their pregnancy. Identifying the key traits which predict this behaviour will allow the development of more efficient screening procedures and interventions for this vulnerable group, improving outcomes for both mother and foetus. While both psychological vulnerabilities (such as depression) and socioeconomic risk factors (such as material deprivation) have been considered individually in their relationship with smoking during pregnancy, the relative importance of psychological versus socioeconomic factors has not been determined. In addition, very few studies have considered the complex inter-relationships between socioeconomic status (SES) and psychological wellbeing, and how these might contribute to smoking in pregnancy. The purpose of the present project is to use the ALSPAC dataset to address these two issues. Pregnant women in this dataset will be classed as either continuing smokers (smoked prior to pregnancy and continued during second and/or third trimesters), quit smokers (smoked prior to pregnancy but not during second or third trimesters), and never smokers (did not smoke prior to pregnancy or during pregnancy). Differences between these groups in psychological variables (depression, anxiety, and experience of stressful events during pregnancy) and socioeconomic variables (educational attainment, financial difficulties and neighbourhood deprivation) will be tested using multiple logistic regression. The second phase of this project will use network outcome analysis to map the complex inter-relationships between psychological and socioeconomic risk factors and smoking in pregnancy. This will allow us to identify the best targets for intervention.

Gastrointestinal infections are common, with 1 in 4 people in the UK population experiencing an episode each year, which causes an estimated annual cost to the individuals, the National Health Service and the wider economy of £1.5 billion. This research will take a life-course approach to the assessment of the causes and consequences of gastrointestinal infections in the community, from early childhood into early adulthood – continuing into the next generation (ALSPAC-G2 study: Children of the Children of the 90s). This will allow us to provide important insights into changes in causes and consequences of gastrointestinal infections across the individual life-course but also over time at the population level. The findings can then inform public health interventions to reduce the burden of gastrointestinal infections, especially among the most vulnerable groups.

The Clinical Interview Schedule – Revised (CIS-R) is a structured diagnostic measure developed from the Clinical Interview Schedule (CIS), a standardised interview designed to assess common mental disorders among community settings. The CIS was designed for use by clinical interviewers and required expert judgement to determine psychopathology. The CIS was standardised to enable interviewers without this expert knowledge to administer, the resulting CIS-R can thus be self-completed and returns results comparable to those from standardized interviews. The CIS-R has been validated across a number of populations and ages, and studies suggest that the instrument remains valid across a number of cultural settings and age groups. To date, little work has been done to validate the CIS-R administered using different modes of assessment.

Previously the CIS-R has been administered to ALSPAC participants via a computerised assessment during clinic time. To minimise participant burden during clinic time and maximise response rate to the CIS-R, it has been suggested that the assessment could be completed online, outside of the clinic. The CIS-R has not currently been validated as an online task, therefore we propose to investigate whether responses to this measure differ according to the setting in which it is completed, and whether it remains a valid diagnostic measure when completed online outside of the clinic setting.

Despite the common belief that teachers can impact various aspects of students' lives, the empirical evidence on the breadth and strength of their influence is relatively unclear. This project aims to examine the relative strength of the influence of teachers' characteristics and behaviours as perceived by multiple raters (i.e., teachers themselves, students, and parents) on multiple experiences and outcomes associated with the teacher (e.g., job satisfaction) and their students (e.g., academic achievement).

Ultimately, the similarity between parents and their children can come from two sources: nature and nurture. Recently, methods have been introduced to identify the effects of “genetic nurture” - effects of the parentally-provided environment which can be explained by the parents’ genes through their influences on parental behavior (rather than through genetic transmission to their children). We have developed an extension of these methods which may allow us to find specific genetic factors which make up this effect, for nuclear families with genotype data available on at least 2 members (e.g., sibling pairs or parent-offspring pairs). This method involves imputing missing parental genotypes from the available genotypes of each nuclear family. Because of its detailed longitudinal phenotype data and genotyped mothers and children, ALSPAC is an ideal dataset to examine these effects across a number of traits important to human health and disease. The results of this study will better help us understand the important role of the parental environment in childhood development and identify specific genetic and environmental risk factors (and their interaction) for physical and mental health outcomes in children.

Social withdrawal (SocW) – defined as reduced social interaction – is a key feature of psychotic disorders and multiple other psychiatric outcomes. Despite growing recognition of its negative consequences, SocW remains one of the least studied human traits, and its underlying mechanisms remain subject of wide speculation. The overall goal of this project is to examine elucidate the role of SocW in the development of psychosis.

By combining clinical data with increasingly powerful ‘genetic risk scores’ (i.e. the sum of risk genes that have shown to be associated with a behavioral trait), we are now able to test whether SocW is a cause or merely a consequence of symptoms in psychosis development. Using the UK Biobank we will create a powerful genetic risk score for SocW, and use this score to examine the role of SocW in psychosis in two large developmental cohorts: The Avon Longitudinal Study of Parents and Children (ALSPAC) and Philadelphia Neurodevelopmental Cohort (PNC). Jointly, these studies followed nearly 16,000 young individuals from childhood to young adulthood, through ages encompassing typical onset of and peak onset age of psychosis. In ALSPAC, we will also test the association between changes in psychotic symptoms and changes in SocW behaviorally, and explore the effects of SocW and established environmental risk in the path to psychosis.

Overall, we aspire to better understand the well-established association between SocW and psychosis, determine critical windows during which individuals may be more or less prone to the influence of SocW, and facilitate much-needed preventive targeted therapies.

Cannabis is widely used for recreational and medical purposes with usage increasing rapidly following its legalization in parts of the world[1]. While there are well-documented effects of cannabis on the users themselves, the impact of its consumption during pregnancy on offspring is much less clear. Previous studies have reported that maternal cannabis usage during pregnancy is associated with lower birth weight and there is the suggestion of impaired neurodevelopment in both observational cohorts in humans and experimental studies in animal models[2-7]. However, how maternal cannabis exposure leads to these adverse health outcomes in offspring remains unclear.
Epigenetic mechanisms play an important role in regulating gene expression and cell homeostasis. Previous experimental studies showed that DNA methylation in offspring could be modified by maternal cannabis exposure[8-10]. In some cases this leads to dysregulation of the immune system of the fetus[11-13]. However, no human studies have yet been conducted, and no specific epigenetic signatures have been identified related to cannabis exposures. We hypothesize that maternal cannabis exposure during pregnancy alters the epigenetic profiles and leads to a specific epigenetic signature of the fetus through in utero exposure, and the epigenetic alterations can then lead to impaired neurodevelopment in the offspring. Therefore, we propose to investigate the association of maternal cannabis exposure around the time of pregnancy with DNA methylation profiles in offspring at birth in the ALSPC cohort.