Your description is a little vague on design, but I'll take a shot and suggest that you could simply randomize the trial order.