I think BenF is correct here. If you look at Trig Sens as hysteresis, then you want the trigger to recognize the hysteresis of the modulation but not the carrier signal itself. Because the modulation modifies the carrier amplitude, then the carrier must also be present within this hysteresis. Therefore the hysteresis is moot. When the carrier signal passes the trigger level, that is because the modulation pushed it up there. They are one and the same.
The only thing that I can imagine as a use for this Trig Sens in an edge triggered situation is to find amplitude glitches that exceed the normal signal parameters and single sweep mode trigger on those events. In this application, the hysteresis would be set outside the expected waveform voltage excursions (outside on top or bottom or both) and only abnormal events would be triggered.