I looked up the full text of this study ... a few things I noticed.
First off, the sample was 20(!) former MDMA users (abstinent for a year) and a control group of 20 non-users. This is a very small sample size (which is unfortunately kind of necessary, given the amount of effort and resources necessary to run an fMRI study). So it's unclear whether or not enough information has been gathered to generalize this data.
Thirdly, the upper end of MDMA usage on this study was pretty high. The subject who had rolled the most had reportedly rolled 155 times, with a reported lifetime intake of 10 grams (which seems low for 155 rolling episodes, but whatever). The average was 33 reported lifetime uses, and about 2.6g reported lifetime intake. With no way of knowing how often they rolled, we have no way of generalizing this data with spacing out rolls, redosing, etc in mind. It's possible (although improbable) that one of the people had only eaten 1g in their entire life but did it all over the course of a week on consecutive days, which would obviously do a LOT more damage than eating 2-3g over the course of a few years, with 30-60+ day breaks in between. Basically, this information isn't enough to generalize with regards to recreational use, aside from simply saying that your risk goes up as you do more.
Finally, looking at the results ...
This is a pretty loose correlation.