The target is a diamond shape with either the left or right point cut off. Your task is to report which side is cut off. On some trials there is only noise and no target at all. You still have to make the same judgment, though. After you've made your judgment, press the space bar for the next trial.
After the last trial, the experimental window will close, and a new window should open that displays proportion correct reports for those trials with the target (guessing should be 0.5), and proportion reports of "right" for those trials that are noise only, as a function of ISI.
Additionally, several smaller windows will appear that show classification images for the noise trials. You may need to resize them to see the whole image. These probably will not have much useful information because there is no where near enough data.
With these stimuli on the computer in my office I get nice u-shaped masking functions.