Presently disclosed are systems and methods for beamforming training in WLANs. In various embodiments, there are unified MIMO beamforming training procedure, which includes a training period in which an initiator transmits multiple unified training frames for performing a transmit-beamforming training of the initiator and a receive-beamforming training of one or more responders; a feedback period in which each responder replies with a beamforming-feedback response; and an acknowledgement period during which the initiator transmits respective acknowledgement frames to the one or more responders from which responses were received. Rules for restricted random access in various slots of the feedback period may be implemented, to address response contention between multiple qualifying responders.