One thing has always puzzled me in auditory nerve physiology. Most auditory nerve fibers saturate quickly when sound level is increased. This implies that, at least for tones, firing rate carries no information at high level. But then it is also known that above a few kHz (depending on species), neural firing in response to tones also becomes imprecise (loss of phase locking), so that spike timing does not carry information. So in other words: at high levels, high frequency auditory nerve fibers spend a lot of energy firing spikes that carry close to zero information. It is odd.
Now in a recent study in J Neurosci, Temchin and Ruggero show some data that suggest (although it remains a bit speculative) that spontaneous firing of fibers may reflect mechanical vibrations rather than neural noise. In my mind, this implies that there is shared variability in the firing of neighboring fibers, and therefore, the loss in spike timing precision in high frequency may hide precise relative timing between neighoring fibers, and this would carry information about the temporal fine structure of sounds.