Measuring the rest potential of a membrane is relatively easy and straight forward: put one electrode inside the neuron, the other one (the "counter electrode") outside the neuron.
Because everything is at equilibrium (steady state), it's not important where exactly inside the neuron you place the first electrode, and even less important where outside the neuron you place the counter electrode: in the extracellular fluid next to the neuron or inside a container filled with extracellular fluid placed anywhere. One may even connect the other electrode to the ground (a lightning rod), if one does it consistently (over all measurements).
But now consider you want to measure the spread of depolarization along the membrane by measuring the membrane potential at two different locations of the membrane: You place one electrode underneath the membrane at point $p_1$ (and its counter electrode to the ground). And you place another electrode underneath the membrane at point $p_2$ (and its counter electrode to the ground).
Now it seems possibly important where exactly inside the neuron you place the two electrodes: the distance from the membrane might make a significant difference. (But maybe not.)
What probably would make a significant difference: Since the ionic compositions of the extracellular fluid in the vicinity of $p_1$ and $p_2$ will differ, it might be important to connect the counter electrodes not to the ground (or to some volume of normal extracellular fluid) but to corresponding points $p_1'$, $p_2'$ directly opposite to $p_1$, $p_2$ on the other side of the membrane.
I emphasized might be important because in practice it may be not important. That's what my question is for:
Is it OK, even for the measurement of the dynamics of depolarizations, to connect the counter electrodes to the ground (if only consistently). What would be the difference in the measured voltages, and can these be determinstically be cancelled out?