In probabilistic conditions, people choose low-payoff alternatives on some trials, thus failing to maximize their payoffs. We suggest that such behavior implicates exploration of task rules by choosing risky options instead of exploiting more rewarding alternatives. We hypothesized that exploration would affect brain responses to feedback. Further, a shift to exploration develops gradually and, therefore, a decision to make an exploratory choice may be observed on trials preceding risky choices. We investigated beta power (16–30 Hz) in the magnetoencephalographic data from 62 healthy participants performing a two-choice probabilistic gambling with monetary gains and losses. The effects were found at 600–800 ms after feedback onset in frontal, central and occipital brain regions. On trials preceding risky choices we identified a decrease in beta power which implies a change in decision-making strategy and a shift towards cognitive flexibility and exploration. An increase in beta power during risky decisions indicates that reward learning mechanisms are implicated. Increases in beta power following losses in risky choices indicates at the process of updating the internal representation of the task. In summary, current findings reveal that the outcomes of exploratory trials are processed differentially, while there is no evidence of such processing on exploitatory trials. This corroborates the hypothesis that exploratory choices represent active probing into the surmised task rules. Current findings also suggest that the processing of outcomes preceding the exploratory trials is altered in such a way that subjects override their intention to use the utility model and reset their behavioral strategy.