Predict brain responses to video, audio, and text.
Recent revisions are kept. Restoring one creates a new revision.