Much of human behavior is characterized by the extraordinary ability to quickly reconfigure our mind, inhibit prepotent responses, and switch between different tasks, often referred to as cognitive control (or executive functions). When defining cognitive control, it is often contrasted with arguably more low-level forms of learning, such as stimulus-response learning. However, this traditional definition on cognitive control might have put us on the wrong path when trying to understand the training of, development of, and impairments in these control processes. Therefore, in this talk, I will present a set of studies that focus on how optimal cognitive control processes, decision making strategies, or reinforcement learning parameters can be learned through selective reinforcement or contextual cueing. Together, these results argue for a more integrative learning perspective on cognitive control.