Abstract / Summary
A viral video of a dangerous stunt can make a viewer's palms sweat and heart race, even though they know they are safe. Such videos let us study emotion as it naturally unfolds, building, persisting and receding. Here we characterise the response to one such film moment by moment and present the open-source browser platform we built to do so. Viewers rate valence and arousal continuously on an affect grid embedded in the video. Each rating is stored with the frame on screen, the film pauses when viewing is interrupted, and participants are trained and their input checked. The film was a 13.5-min silent first-person recording of unsecured climbing. An exploratory sample (n = 21) informed preregistered tests in an independent confirmatory sample (N = 124). Arousal rose as the climb progressed, held a plateau for several minutes and then declined towards baseline, with valence most unpleasant during the plateau. This time course was highly reproducible across samples. Somatic anxiety increased substantially after the film (dz = 1.03), and participants with larger increases had rated it as more arousing and unpleasant. The film thus carries viewers through a reproducible, interoceptively charged emotional episode, and the platform can characterise other stimuli. Researchers can ask how the shift from a calm to an aroused state changes brain connectivity, interoception and emotion, and whether it is heightened in anxious people. Platform, data and reference time courses are openly available.