Differences between RMSE from Regress function and basic fitting toolbox

I'm looking at the relationship between modelled and observed values of some variables. I initially used the basic fitting toolbox on my scatter plots, but the RMSE values looked very large, and so I went back and used the regress function:

https://www.mathworks.com/help/stats/regress.html , where the outputs in "b" are m and c, and in "stats" are R^2, F, P, and RMSE^2

The values of m, c and R^2 do not change between the wizard and the function, but the RMSE values from the regress function are approximately 150 times smaller. Why?