AbstractVirtual reality (VR) offers a powerful tool for investigating cognitive processes, as it allows researchers to gauge behaviors and mental states in complex, yet highly controlled, scenarios. The use of VR head‐mounted displays in combination with physiological measures such as EEG presents new challenges and raises the question whether established findings also generalize to a VR setup. Here, we used a VR headset to assess the spatial constraints underlying two well‐established EEG correlates of visual short‐term memory: the amplitude of the contralateral delay activity (CDA) and the lateralization of induced alpha power during memory retention. We tested observers' visual memory in a change detection task with bilateral stimulus arrays of either two or four items while varying the horizontal eccentricity of the memory arrays (4, 9, or 14 degrees of visual angle). The CDA amplitude differed between high and low memory load at the two smaller eccentricities, but not at the largest eccentricity. Neither memory load nor eccentricity significantly influenced the observed alpha lateralization. We further fitted time‐resolved spatial filters to decode memory load from the event‐related potential as well as from its time‐frequency decomposition. Classification performance during the retention interval was above‐chance level for both approaches and did not vary significantly across eccentricities. We conclude that commercial VR hardware can be utilized to study the CDA and lateralized alpha power, and we provide caveats for future studies targeting these EEG markers of visual memory in a VR setup.