No, they can't actually record everything that happens in neurons. They observe small set of changes, and based on those changes and currently understood theory of brain they reconstructed full picture* (Kinda like archeologists reconstruct structure of dinosaur bones based on other knowledge such as bird bones).
* I have no idea how accurate that technique or my description of it is, but in general that is what happens. We got a model of something, we record set of measurements, and then we compute what internal state the model needs to be in for us to observe that set of measurements.
* I have no idea how accurate that technique or my description of it is, but in general that is what happens. We got a model of something, we record set of measurements, and then we compute what internal state the model needs to be in for us to observe that set of measurements.