I've just read the article, not the paper (yet). From the article it's pretty clear that they are talking about chords, not (just) the spectral composition of sounds. Since the invention of tempered tuning in the 17th century, the notes in a chord are no longer in an even numbered ratio, but usually in a multiple of the twelfth root of two. It is rather the case that people today consider a chord to sound strange if it is made up of whole-number ratios, which again indicates habituation. Besides, Pythagoras' theory also referred to specific tone intervals, not the spectral composition of arbitrary sounds. Anyway, let me first read the paper to see what they actually claim to have found out.