relevant visualization of where research happens: http://www.nature.com/polopoly_fs/7.8087.1355853889!/image/c...
i.e. of the ~300K papers published in the US, how many of those papers were authored and co-authored by immigrants (H1-B, or naturalized) etc.
I suspect that a greater % of these papers were written by foreign born researchers.
Say 25% of the researchers in the US were born in the US: The US column would drop significantly but since that other 75% is probably spread more or less evenly across the other columns, they would only all rise a little.
The only really dramatic change I would expect would be to mainland China. It probably wouldn't loose much, but would likely gain quite a lot.