The issue with Sets E and F had less to do with overall LDI increases (although there was some evidence of that), but more to do with the distribution of difficulty across the lure bins for creating the low-to-high similarity bins. Instead of using sets E or F directly (we have not published using those sets without converting them to C/D using the LureBinConversion, which can be found on our website), I recommend selecting stimuli across sets that fall into the easier lure bins. I am attaching here, a list of the stimuli and their lure bin status, so you could pull those images to create your own sets and then use the Matlab script to run the experiment.
Hope this helps!
Shauna