When should highlight boxes appear and disappear in an e-learning to a user? Should highlight box disappear after the action to a user or before?

Good question.  I don't know of any standard, but here's how I use highlights.  If I'm describing something on the screen with the voice over, I highlight that object as I'm describing it.  If an action is to take place on the screen, I highlight the place where the action is to take place, then remove the highlight just before the action - if it's just a demonstration.  If it's a simulation where the user takes the action, instead of a highlight, I use a smart shape with a "pointer" pointing to the place to take the action, and leave it there until they do take the action.