I. Introduction
Recent studies on human RL found that humans use both MB and MF reinforcement learning strategies [1], [2]. However, classification of these different learning strategies is not investigated so far. The aim of this paper is to build and optimize a simple classifier which would attempt to distinguish between these two RLs. Specifically, we classified two types of trials, each of which is designed to promote MB and MF RL, respectively, from EEG data recorded from human subjects brains during performing a two-stage Markov decision task.