Abstract-We study the feasibility of cognitive radio (CR) communication in the presence of a K-user multi-input multioutput (MIMO) interference channel as the primary network. Assuming that the primary interference network has unused spatial degrees of freedom (DoFs), we first investigate the sufficient condition on the number of antennas at the secondary transmitter under which the secondary system can communicate while causing no interference to the primary receivers. We show that, to maximize the benefit, the secondary transmitter should have at least the same number of antennas as the spatial DoFs of the primary system. We then derive the secondary precoding and decoding matrices to have zero interference leakage into the primary network while the signal-to-interference plus noise ratio (SINR) at the secondary receiver is maximized. As the success of the secondary communication depends on the availability of unused DoFs, we then propose a fast sensing method based on the eigenvalue analysis of the received signal covariance matrix to determine the availability of unused DoFs or equivalently spatial holes. Since the proposed fast sensing method cannot identify the indices of inactive primary streams, we also provide a fine sensing method based on the generalized likelihood ratio test (GLRT) to decide the absence of individual primary streams. Simulation results show that the proposed CR sensing and transmission scheme can, in practice, provide a significant throughput while causing no interference to the primary receivers, and that the sensing detects the spatial holes of the primary network with high detection probability.Index Terms-Cognitive radio, K-user MIMO interference channel, interference alignment, null space sensing, spatial holes, eigenvalue-based sensing, GLRT detector.