With an N-MOSFET has an N-doped channel between drain and source, it'll be conductive without any gate voltage applied. Consider you apply a negative charge to the gate, the resulting electric field pushes electrons out of the channel. As you add more charge, so many electrons will leave the channel that it undergoes inversion and effectively behaves as if it was P-doped, at which point the MOSFET stops conducting.
Would the inversion P-Doping region occur close to the oxide, or close to the interface between the N-channel (connecting the source and drain) and the P-substrate?