The article discusses the distinction between reflective and reflexive decision-making in cognitive neuroscience, highlighting the roles of goal-directed and habitual behaviors. It traces the evolution of research from early debates on cognitive maps and stimulus-response theories to modern studies on model-based and model-free reinforcement learning. The first generation of studies established a dichotomy between goal-directed behavior, which relies on knowledge of action-outcome relationships and current motivational relevance, and habitual behavior, which is learned through reinforcement and is less flexible. The second generation adapted these paradigms to human neuroimaging, validating findings from rodent studies. The third and fourth generations used model-based and model-free reinforcement learning to analyze decision-making, revealing the computational and statistical characteristics of these systems. The fifth generation of research is now addressing the interaction and competition between these systems. The article also discusses the neural substrates of these behaviors, including the role of the hippocampus in spatial navigation and the dorsolateral and dorsomedial striatum in habitual and goal-directed actions, respectively. Human studies using fMRI have shown that the ventral striatum is involved in model-free control, while the orbitofrontal cortex is involved in model-based control. The article concludes that both systems are essential for adaptive behavior, with model-based control providing efficient solutions to complex environments and model-free control facilitating automatic responses. The interplay between these systems is crucial for decision-making, and understanding their interaction is an active area of research.The article discusses the distinction between reflective and reflexive decision-making in cognitive neuroscience, highlighting the roles of goal-directed and habitual behaviors. It traces the evolution of research from early debates on cognitive maps and stimulus-response theories to modern studies on model-based and model-free reinforcement learning. The first generation of studies established a dichotomy between goal-directed behavior, which relies on knowledge of action-outcome relationships and current motivational relevance, and habitual behavior, which is learned through reinforcement and is less flexible. The second generation adapted these paradigms to human neuroimaging, validating findings from rodent studies. The third and fourth generations used model-based and model-free reinforcement learning to analyze decision-making, revealing the computational and statistical characteristics of these systems. The fifth generation of research is now addressing the interaction and competition between these systems. The article also discusses the neural substrates of these behaviors, including the role of the hippocampus in spatial navigation and the dorsolateral and dorsomedial striatum in habitual and goal-directed actions, respectively. Human studies using fMRI have shown that the ventral striatum is involved in model-free control, while the orbitofrontal cortex is involved in model-based control. The article concludes that both systems are essential for adaptive behavior, with model-based control providing efficient solutions to complex environments and model-free control facilitating automatic responses. The interplay between these systems is crucial for decision-making, and understanding their interaction is an active area of research.