Page 22 - Like me, or else... - Michelle Achterberg

Chapter 1
An additional difficulty when it comes to neuroimaging studies in childhood is that different studies seldom used the same experimental paradigm. This makes it difficult to study reproducibility of behavioral and neural findings. Indeed, the (lack of) reproducible results in psychological studies has received a lot of attention (Ioannidis, 2005; Schmidt, 2009; Open Science, 2015). Moreover, findings that show no evidence of significance when analyzed individually (i.e., due to small sample size and/or low statistical power) might provide stronger evidence when collapsed across samples (Scheibehenne et al., 2016). One particularly elegant way to examine a new paradigm is to use a pilot, test and replication design within the same project and combine results meta-analytically. However, to be able to divide a childhood sample into subsamples - again - requires a large sample size.
All of these factors were taken into account when we designed the longitudinal twin study of the Leiden Consortium on Individual Development (L- CID), Samen Uniek in Dutch. The L-CID study consists of two cohorts (early childhood and middle childhood) that are being followed for six constructive years, with annual home or lab visits (Euser et al., 2016). The majority of studies in the current thesis (Chapters 2, 4, 5, 6, and 7) are based on data from the middle childhood cohort. Specifically, I made use of the data of the first wave, and a follow up measure two years later. The study included 512 children (256 families) between the ages 7 and 9 at time point 1 (mean age: 7.94±0.67; 49% boys, 55% MZ). This large sample size provides sufficient statistical power to examine childhood brain development, specifically when taken into account that neuroimaging data in developmental samples are more prone to data loss and artifacts due to movement (O'Shaughnessy et al., 2008).
Dissertation Outline
The large sample size of the L-CID study allowed me to test for within-sample replication, thereby contributing to the debate about reproducibility of neuroscientific patterns (Open Science, 2015). In doing so, I first examined the SNAT paradigm using a design with built-in replication and meta-analysis. In chapter 2, I tested the SNAT paradigm in separate pilot, test and replication samples and combined the results meta-analytically. The aim of this study was to detect robust behavioral patterns and neural signals related to social feedback, a crucial first step in examining social evaluation processing in childhood. Next, in chapter 3, I investigated neural processes of social evaluation in adults, were I additionally investigated brain-behavior associations to shed light on individual differences in the neural mechanisms for social emotion regulation. Unraveling these neural patterns in adults provided an index to compare the results in middle childhood with.
20