We make choices between different options every day, and thereby we frequently estimate values for different options. The value, which represents the benefit or detriment to the decision maker, is subjective and difficult to measure. For decades, economists have been trying to understand human subjective value functions. They used many different models to study human behaviour and economic decision-making. More recently, neuroscientists started to use these economic models to understand the neural basis of decision-making with human and animal experiments. Nevertheless, how the brain encodes the subjective of options composed of multiple components (e.g. different reward types) has not been systematically studied. In this thesis, I investigated the neural mechanism of two different two-component rewards: (1) fat-sugary milkshake reward in humans using fMRI neuroimaging and (2) probability-magnitude liquid reward in rhesus monkeys using single-cell electrophysiology. In the first experiment, human participants were asked to make choices between two bundle reward options, which each contain two milkshakes. The participants were scanned with fMRI to study how the brain integrates the two-component reward into a scalar brain signal following revealed preference theory, an economic theory widely used for studying subjective preference and choices. In the next experiments, rhesus monkeys were trained to make choices between two visual stimuli (options) using a left-right joystick. In each option, there were different probabilities of getting different amounts of liquid rewards. Because risky decision-making is an important research topic for economists and neuroscientists, some economic theories, including expected utility theory and prospect theory, were developed to model this kind of behaviour. In this study, I analysed how well our monkeys follow these economic theories; in other words, how well these economic theories can explain our monkeys’ behaviour. I also compared the behaviour of monkeys to previous human studies, using machine learning and statistical tools, across different probabilities of rewards. By testing how much monkeys and humans are consistent in risky decision-making, we can obtain a better understanding of how monkeys can be used to study subjective values. Dopamine neurons were recorded in our monkeys with single-cell electrophysiology. We can thereby understand, at a single-cell level, how the brain computes and updates the subjective value of risky choices. We can also investigate whether/how the reward prediction error signal of dopamine neurons can reflect different economic theories and choices. By using all these experiments and analyses from humans and monkeys, I hope this thesis can provide some insights into how the brain encodes subjective economic value and choices.