I'm not an author of the paper - just pointing out that they were trying control for all variables excepting the 'matrix' - so they made both bread and pasta with the *same* flour - all they changed was cooking method, and probably yeast for the bread. No idea if they used durum / semolina; I didn't dig into the study itself (my bad), just Jason Fung's summary. Did you? At least the methods section? I suspect they could have gotten a differential response between 'fresh' and air-dry pasta.