Neuroeconomics explores the neurobiological and computational basis of value-based decision-making. It aims to provide a biologically grounded account of human behavior applicable to both natural and social sciences. This review proposes a framework to integrate recent findings, highlight key problems, define a common lexicon, and guide future applications. Value-based decision-making is widespread, occurring in both simple and complex choices. Neuroeconomics combines models, tools, and techniques from economics, psychology, neuroscience, and computer science to understand decision-making processes. The framework divides decision-making into five types of computations: representation, valuation, action selection, outcome evaluation, and learning. Representation identifies potential actions and states. Valuation assigns value to actions, action selection compares values, outcome evaluation measures results, and learning updates processes based on feedback. These are conceptual categories, not rigid ones, with many open questions about their accuracy. Three valuation systems—Pavlovian, habitual, and goal-directed—exist. Pavlovian systems assign value to innate or learned responses. Habitual systems learn to value actions through repeated training. Goal-directed systems compute action-outcome associations and evaluate outcomes. These systems interact, with Pavlovian behaviors often being automatic, while goal-directed behaviors involve planning. The Pavlovian system values few responses, while the habitual system learns to value many. The goal-directed system evaluates outcomes based on their rewards. Neural mechanisms for these systems are being studied, with the medial orbitofrontal cortex and ventral striatum involved in valuation. The amygdala and nucleus accumbens are involved in Pavlovian responses. Habit systems rely on dopamine and cortico-thalamic loops. Goal-directed systems use the dorsomedial striatum and orbitofrontal cortex. Learning involves trial-and-error, with habit systems learning slowly and goal-directed systems updating values based on outcomes. Risk and temporal discounting affect decision-making. Risk involves uncertainty, while temporal discounting involves delayed rewards. The brain uses different mechanisms to handle these, with the striatum and prefrontal cortex involved in processing risk and time. Action selection involves comparing different actions. The brain uses models like the race-to-barrier diffusion process to make decisions. Competition among valuation systems can lead to conflicts, with the brain resolving control through experience. Outcome evaluation measures the desirability of outcomes, with the medial orbitofrontal cortex and insula involved. Learning involves updating values based on feedback, with prediction errors playing a key role. Neuroeconomics has applications in psychiatry, law, public policy, marketing, and artificial intelligence. Understanding decision-making processes can improve diagnoses, treatments, and policies. Future research aims to characterize the computational and neurobiological basis of decision-making processes.Neuroeconomics explores the neurobiological and computational basis of value-based decision-making. It aims to provide a biologically grounded account of human behavior applicable to both natural and social sciences. This review proposes a framework to integrate recent findings, highlight key problems, define a common lexicon, and guide future applications. Value-based decision-making is widespread, occurring in both simple and complex choices. Neuroeconomics combines models, tools, and techniques from economics, psychology, neuroscience, and computer science to understand decision-making processes. The framework divides decision-making into five types of computations: representation, valuation, action selection, outcome evaluation, and learning. Representation identifies potential actions and states. Valuation assigns value to actions, action selection compares values, outcome evaluation measures results, and learning updates processes based on feedback. These are conceptual categories, not rigid ones, with many open questions about their accuracy. Three valuation systems—Pavlovian, habitual, and goal-directed—exist. Pavlovian systems assign value to innate or learned responses. Habitual systems learn to value actions through repeated training. Goal-directed systems compute action-outcome associations and evaluate outcomes. These systems interact, with Pavlovian behaviors often being automatic, while goal-directed behaviors involve planning. The Pavlovian system values few responses, while the habitual system learns to value many. The goal-directed system evaluates outcomes based on their rewards. Neural mechanisms for these systems are being studied, with the medial orbitofrontal cortex and ventral striatum involved in valuation. The amygdala and nucleus accumbens are involved in Pavlovian responses. Habit systems rely on dopamine and cortico-thalamic loops. Goal-directed systems use the dorsomedial striatum and orbitofrontal cortex. Learning involves trial-and-error, with habit systems learning slowly and goal-directed systems updating values based on outcomes. Risk and temporal discounting affect decision-making. Risk involves uncertainty, while temporal discounting involves delayed rewards. The brain uses different mechanisms to handle these, with the striatum and prefrontal cortex involved in processing risk and time. Action selection involves comparing different actions. The brain uses models like the race-to-barrier diffusion process to make decisions. Competition among valuation systems can lead to conflicts, with the brain resolving control through experience. Outcome evaluation measures the desirability of outcomes, with the medial orbitofrontal cortex and insula involved. Learning involves updating values based on feedback, with prediction errors playing a key role. Neuroeconomics has applications in psychiatry, law, public policy, marketing, and artificial intelligence. Understanding decision-making processes can improve diagnoses, treatments, and policies. Future research aims to characterize the computational and neurobiological basis of decision-making processes.