The aim of this thesis was to study sources of inter-individual variation in brain structure and function in pre-adolescence. This thesis focused on several factors that contribute to this variation, including test-retest variability, biological factors such as age and sex, socioemotional skills, and childhood experiences. I used data from 1000 children (8-, 9- and 10-year-olds) participating in the first wave of the YOUth: Child & Adolescent cohort.
The first article contains a full description of the YOUth MRI protocol including test-retest reliability of brain measures derived with this protocol (Chapter 2). The information shared in this chapter, such as the state-of-the-art acquisition parameters, extensive quality control procedures and reliability measures, can be useful to the neuroimaging community by aiding to increase the reproducibility and harmonization across developmental neuroimaging studies. The test-retest study using the YOUth MRI protocol in young adults, indeed showed that test-retest reliability was comparable to literature for brain measures derived from all MRI modalities: structural T1-weighted imaging, diffusion-weighted imaging (DWI), resting-state functional MRI and task-based functional MRI. As expected, higher test-retest reliability was found for structural T1-weighted and DWI scans compared to functional brain measures.
In the second article different de-identification (defacing) procedures for anatomical MRI data are compared (Chapter 3). De-identification methods remove or blur identifiable facial characteristics on anatomical MRI scans. Implementing de-identification methods enables researchers to share data while protecting participants privacy. I show that de-identification methods introduce some variability in outcome measures, but the reproducibility was comparable to test-retest reliability and no large systematic effects were found. Furthermore, I show that some de-identification methods do not work well on child brain scans.
In the third article, I investigated associations between anatomical brain measures and adverse childhood experiences (ACEs) (Chapter 4). An association was found between 2 out of 11 adverse childhood experiences: 1) growing up in a family with alcohol or substance abuse issues, and 2) exposure to violence or crime within the family or neighborhood.
For the last article, I used functional neuroimaging data to investigate interrelations between different factors related to social cognition: behavioral emotion labeling, neural processing of emotional faces and social competence (Chapter 5). No relationship was found between these three factors. However, this study did reveal that older children and girls, on average, have better social skills and find it easier to label emotions on faces.
To better support children, it is important to understand the biological mechanisms linking childhood experiences to well-being in later life. Overall, associations between inter-individual differences in brain characteristics and behavioral or environmental phenotypes were low, despite the relatively large sample size. The inherently complex associations between the social environment and the developing brain ask for interdisciplinary collaborations, embracing diversity in all aspects, and pooling data to enhance statistical power. Understanding how childhood experiences convey risk for mental health problems later in life can help to identify malleable factors that could ultimately support children growing up.
This dissertation includes a summary in Dutch and a summary for young readers in Dutch.