Describir: An inter-subject model to reduce the calibration time for motion imagination-based brain-computer interface.