Understanding autonomous behaviours in children: an investigation of self-directed cognitive control development

Gaining autonomy is a key aspect of growing up. However, little is known about how children engage cognitive control in an autonomous (or self-directed) fashion to achieve goals. The aim of this dissertation is to better understand the cognitive processes supporting self-directed control development, examining in particular the potentially prominent role of goal identification. In a first study, we disentangle task selection and task execution by adapting the double registration procedure for use with children, and explore whether these processes follow a similar or different developmental trajectory and to what extent they are influenced by task self-directedness regarding performance costs. We show that task selection development lags behind task execution development, task selection yields both mixing and switch costs whilst task execution yields switch costs only, and that both task selection and task execution are affected by task self-directedness. In a second study, we adapt a paradigm widely used in the adult literature to investigate self-directed control, the voluntary task-switching paradigm, and explore task selection development and the contributions of proactive and reactive control to this form of control. Our results indicate both younger and older children struggle with task selection as compared to adults, and that reactive and proactive control may be related to self-directed control. Further, younger children did not struggle to switch tasks, indicating that switching per se is unlikely to be the main source of difficulty for children when engaging cognitive control. In a last study, we specifically target and disentangle context-tracking and task selection by varying the difficulty on each process in two experiments. We find that, although both processes contribute to self-directed control performance, progress with age is mostly associated with improvements in context-tracking. These results are then discussed and leads to the elaboration of a tentative theoretical model of self-directed control development