"I'm not entirely sure what you mean by sampling pixels"
I think I'm confusing some things. I can understand that each pixel of a sensor can output values from lets say 0 to 1V where 0V is dark, 1V is light. So you need a shutter-time because else the sensor will only output values of 1V.
I was thinking todays cameras can sample these values at a very high frame-rate. So when you sample these values over time and store them you can expand the range. But it seems like "bracketing" is doing something like this. I did not know. Thanks for the info!