Every year, 50,000 Canadians will have a stroke. Of those who survive, approximately onethird will have severe upper limb paralysis [1] and will not respond to conventional therapy [2]. Remarkably, Functional Electrical Stimulation (FES) therapy has successfully restored arm and hand function in stroke patients with severe hemiplegia [2-4]. However, FES therapy does not require consistent involvement of the CNS or a controlled inter-stimulus latency;​ both of which are critical aspects of neuromotor rehabilitation [5, 6]. An application using a BCI to control FES therapy may address both of these limitations. The hypothesis that several hand movements used during stroke rehabilitation can be detected from EEG was investigated in this study. Three different hand movements were classified with average accuracies between 65-75% across participants. This is the first account of specific hand movement detection using pre-motor EEG, and may assist in future designs of EEG-based BCI+FES therapy for stroke patients.