To achieve this effect, you will need to put all of your fixation points and images on a single line of your .csv input file. Whenever you have more than one row of stimuli in a .csv input file, DirectRT must show a blank screen because it needs time to compile all of the stimuli for the upcoming trial (see here in the DirectRT manual for more information: http://www.empirisoft.com/directrt/h...l_interval.htm.
I've attached an example .csv file that should do what you want. Let me know if you have any other questions.