Data was collected from electrophysiological recordings of brain activity and behaviour in freely-moving mice. Brain electrical signals were recorded from micro-wire bundles implanted in the hippocampus. Behavioral data were collected from an accelerometer incorporated in the head-stage pre-amplifier of the electrophysiological setup. Since the head-stage was connected to the implant during recordings, the accelerometer captures the head movements (instantaneous accelerations). Each instance, representing a time point where the brain state has to be classified, contains two types of inputs: - A power spectrogram of the local field potential (LFP, representing local electrical actifity) around the time point to classify (+- 50 s). - A segment of the movement magnitude time series, extracted from an accelerometer on the mouse head. Each brain state is associated with a particular combination of patterns of LFP power and movement, allowing classification.