It seems that you have a similar issue to what I was doing recently. Perhaps this will help:
in short, it might be enough to use the following:
# check if the mouse is pressed within the current one:
if mouseresp.isPressedIn(your target image name):
thisExp.addData('correct', '1')
else:
trhisExp.addData('correct', '0')
However, I’m not sure this will work the same, since your 4 images appear to be text objects? Are you presenting your participants with images or words?