Feedback plays an important role when learning to use a Brain-Computer Interface (BCI). Here we compare visual and haptic feedback in a short experiment. By imagining left and right hand movements, six subjects tried to control a BCI with the help of either visual or haptic feedback every 1s. Alpha band EEG signals from C3 and C4 were classified. The classifier was updated after each prediction using correct class information. Thus feedback could be given throughout the experiment. Subjects got better at controlling the BCI during the experiment independent of the feedback modality. Haptic feedback did not present any artifacts to the signals. More research is required on haptic feedback for BCI-applications because it frees visual attention to other tasks.