A couple "work-around" options in order to use the "overwrite the image" approach. If the late response (>675ms) can be expected within a second window for 675ms--i.e., and overall RT of 1350ms, then you could eliminate the clear675 function and instead use a minmax value of 0-675! for the style on that trial. That would cause automatic continuation from the first opportunity to respond to the second. Importantly for your purpose, you would not have to clear the screen between stimuli--you could simply cover the initial image but no the labels. In your subsequent data crunching, you would then need to add the two RTs together. This raises again the issue though of what you want to happen if they *do* respond in the first 675ms. Do you want an overall fixed response window of 1350ms? Or do you want them to continue to the next trial immediately? If the latter, then you could put a skip function in for each response that exits them from the trial immediately to the next. Let me know if you'd like elaboration on any of this or if it's not what you are looking for.