The prefrontal areas of the cortex have been implicated in decision making and flexible behaviour in different species spanning from rodents to non-human primates and humans. In particular, the orbitofrontal cortex has been associated with value representation and integration and the prelimbic cortex with strategy changes. However, the neuronal mechanism underlying such cognitive processes remains elusive. We aim to gain further mechanistic insights into how neuronal circuits of the prefrontal cortex compute decisions, both at a single neuron and population level, and how circuit operations can be related to flexible behaviour at different timescales.We train mice to perform an adapted version of the Iowa Gambling task. During the task, the mice have to choose between a big but uncertain or a small and certain reward. The probability of the big reward changes throughout the session, making different choices optimal in different parts of the session. In order to maximize the reward, the animals need to dynamically evaluate the outcome of their choices. While the mice are performing the task, we use silicon probes to record neurons in the prelimbic and orbitofrontal cortices and we identify GABAergic interneurons by optotagging.On a behavioural level, we observe that mice often decide based on the experience over a longer time frame rather than on a trial-by-trial basis. Furthermore, we identify firing patterns of single neurons and neuronal populations correlating with task performance and choices.Funding: Grant I 5458 of the Austrian Science Fund and FOR5159 of the German Research Foundation