Thanks for this! I will try this method, too. I just had success with using fixation cross as a separate trial, then blocking them such that the even numbers were fixation crosses and odd numbered trials were my stimuli/"real" trials. Here is my abbreviated syntax for future viewers:
<expressions>
/fixduration = rand(750,1500)
</expressions>
<trial fixation>
/stimulustimes = [0=fixation]
/trialduration = (expressions.fixduration)
/correctresponse = (noresponse)
</trial>
<trial ANG_congR>
/stimulustimes = [0=ANGfacesR,NEUfacesL; 125=eraser,probeR]
/correctresponse=("I")
/validresponse=("E","I")
/pretrialpause=0
/timeout=3000
</trial>
<block practiceblock>
/trials=[1=blockpause; 2,4,6,8,10,12,14,16,18,20=fixation; 3,5,7,9,11,13,15,17,19,21=noreplace(ANG_congR)]
/preinstructions=(intro1, intro2, intro3)
/recorddata=false
/postinstructions= (begin)
</block>