The article reviews the evolution of research on decision-making and choice in cognitive neuroscience, focusing on the dichotomy between reflective and reflexive control. The authors trace the development of this field through four generations of studies, from early cognitive maps to more recent model-based and model-free reinforcement learning approaches. They highlight the importance of neural mechanisms, such as the dorsomedial and dorsolateral striatum, and the role of dopamine in modulating these systems. The review also discusses the computational and statistical characteristics of goal-directed and habitual control, and the ongoing efforts to understand their interactions and integration. The authors conclude by sketching the potential directions for the fifth generation of research, emphasizing the need to explore the dynamic interplay between goal-directed and habitual processes in complex and changing environments.The article reviews the evolution of research on decision-making and choice in cognitive neuroscience, focusing on the dichotomy between reflective and reflexive control. The authors trace the development of this field through four generations of studies, from early cognitive maps to more recent model-based and model-free reinforcement learning approaches. They highlight the importance of neural mechanisms, such as the dorsomedial and dorsolateral striatum, and the role of dopamine in modulating these systems. The review also discusses the computational and statistical characteristics of goal-directed and habitual control, and the ongoing efforts to understand their interactions and integration. The authors conclude by sketching the potential directions for the fifth generation of research, emphasizing the need to explore the dynamic interplay between goal-directed and habitual processes in complex and changing environments.