The Null hypothesis of t-test in non-linear least-squares ?

[Jay Leo]

I'm reading a paper named "Flows of knowledge from universities and federal laboratories:Modeling the flow of patent citations over time and across institutional and geographic boundaries", in which the function like:
P = ai*aj*ag*exp(-b1*bi*bj*bg*x)*(1-exp(-b2*x)), in which ai~ag,b1,bi~bg and b2 are parameter for estimating.
In this paper, the Non-linear least-squares regression results table shows the t-statistic for H0 (Parameter = 1).
My question is what is the H0 of the t-test in normal non-linear least-squares in Stata (Parameter = 0? or 1?) ? And how to make the H0 as Parameter = 1, like the above paper?

[Clyde Schechter]

Here's a simple example of how it's done:

sysuse auto, clear
nl (price = {a1}*mpg^{a2})
nl, coefl
test _b[a1:_cons] = 1

[Jay Leo]

Thanks