ran the statistics comparing the distribution of major Roman roads with city populations 1200-1850 and with the amount of light you see across Europe in a satellite image.
Which is interesting, but I am not sure it says much about Roman roads. It does say something rather dramatic about the population structure of western Europe over the past 2,000 years: most of the great Roman cities are still great cities, and within the empire few new places have appeared to rival them.
To test for the impact of Roman roads in particular you would have to somehow identify territories that are very similar except for the presence or absence of a road, and I am not sure how that would be done. Anyway these guys did not do it.