neuroscience - Why do neurons have a negative resting potential?

Neurons expend the majority of their energy powering ion pumps to maintain the chemical gradients that power their electrical activity. To have a negative resting potential, neurons leak potassium across the membrane, which seems like a terrible waste of energy to me. I would like to know what benefit a neuron receives in exchange for this seemingly unnecessary metabolic load.

I am not asking how the resting potential is achieved. I am also not interested in the trivial answer: that the voltage-gated channels are configured to require a transition across the -40mV or so threshold in order to fire an action potential. It seems to me that this threshold is arbitrary; if there was no advantage to maintaining this gradient then neurons would have evolved to avoid it.

Any ideas? Or better yet, pointers to places where this has already been answered?

My best guess so far looks like this: The total range of available voltages is more-or-less fixed from -90 to +50mV. We want to avoid getting too close to either end, since the channels become less effective near their reversal potentials, so maybe the effective range is more like -70 to +30 (to go outside that range, we must sacrifice speed). Within that 100mV range, we leave the bottom 30mV or so for EPSP integration, and the other 60mV for action potentials. Now, if the resting potential was 0mV, the available dynamic range for integration and spiking would be much smaller which probably translates to making the output noisier.

Answer

Essentially all animal cells maintain an ionic balance causing a resting potential of about -70 mV in order to maintain their internal environment including pH, ion concentrations, osmotic pressure and volume. (Lodish, Molecular Cell Biology) Neurons developed from existing types of cells and it's unlikely that the cost of maintaining resting potential in the neuron could have driven evolution of an entire alternative system to provide the homeostasis supported by the existing system.

Note that the depolarization of the membrane at any particular place during an impulse is very short so the impact of the ion flows have only limited effect on the cell's overall internal environment beyond requiring adjustment by ion pumps for the losses involved. It also is not obvious how the wave-like characteristic of a nerve impulse could be generated if the membrane did not carry a nonzero potential; there would be no stored energy (in the form of ion gradients) available to make the pulse swiftly spread across the membrane.

Another point to consider is that the perhaps 15% of a neuron's energy spent on replacing leaking ions may have some hidden usefulness. This paper suggests that neurons may actually be set up to leak at a rate higher than they would if it were minimized.