Creating video simulation where user identifies bad behaviour, unprompted.

I'm wondering if this is possible in Captivate, or if anyone out there has any experience doing this:

A video plays on a slide - it is a small vignette of two people interacting with each other.  As the conversation moves along, the user has to identify good and bad things that are happening in the video, and get scored on them.  (Like thumbs up and thumbs down buttons below the video or something.) The video does not pause and the user is not prompted at any particular time.  The user has to identify the cue points on the fly.

I don't know if this is possible in captivate.  It is easy to pause the video, and then put up a few buttons on the screen that the user has to click, and then continue the video - but that's not what we want.

Any insight would be appreciated.