Witten index, defined as ${\rm Tr}(-1)^F$, makes know if supersymmetry is spontaneously broken or not for a given model. But it is known that supersymmetry can be also broken dynamically and one can think of a mechanism (yet to be implemented if any) a la Nambu-Jona-Lasinio model for chiral symmetry in QCD. In this case, how does Witten index change? I think it should reflect the fact that, even if we are not requiring explicitly a selected ground state breaking symmetry, particle masses are lifted and a gap equation is satisfied.
A more general question is if Witten index applies as well to dynamical breaking of supersymmetry, for all the models one can possibly conceive.
As usual, good references are welcome.
Thanks.