The present study examined operant extinction in a controlled setting using a human operant paradigm. Participants watched a preferred video. During the video, either the video or audio portion of the video was selectively removed, on average every 15 s. Participants could restore the video by pressing a force transducer. In one group, relatively low forces were required (250 g) and in the other relatively high forces were required (750 g). At the 20th and 30th minute during the session, the video or audio was removed but the participants could not restore the component for 30 s. The results showed that responding during the probe increased relative to 30-s periods prior and following the probe, characteristic of an extinction burst. The results also showed that overall we saw increases in force under high force conditions during extinction when presses no longer produced sound or video, and force changed little during the low force conditions. We conclude that extinction bursts are a robust phenomenon that can be demonstrated in humans. Additionally, the topographies, i.e. force, established during baseline and the modality of the consequence appear to be two variables determining the short-term course of extinction.