9The auditory nerve

TThe spiral ganglion

The cell bodies of the auditory nerve sit in the spiral ganglion, coiled in the bony core of the cochlea — roughly 30,000 neurons in the human ear. They come in two types. Type I neurons (90–95%) are large and myelinated, and each contacts a single inner hair cell; about ten of them converge on each inner hair cell. Everything we know about auditory-nerve physiology comes from type I fibres, and they carry essentially the whole signal. Type II neurons (5–10%) are thin and unmyelinated and contact many outer hair cells; their function remains one of the genuine mysteries of cochlear physiology.[1972]

CTraffic on this cable is not one-way. Two brainstem efferent systems run back down to the cochlea. The medial olivocochlear (MOC) bundle synapses on the outer hair cells and dials down the amplifier (Module 7). The lateral olivocochlear (LOC) bundle is subtler: it terminates on the type I afferent dendrites themselves, just beneath the inner hair cells, and slowly adjusts their excitability — tuning the fibres' dynamic range and, crucially, balancing sensitivity between the two earsso that binaural cues stay calibrated. It is a reminder that even the nerve's own output is under central control.[2010, 2006]

TCTuning curves

Probe a single fibre with tones and you find it is frequency-selective: it has a low threshold for one frequency — its characteristic frequency (CF) — and needs more and more level as you move away. Plotting threshold against frequency gives the classic sharp, V-shaped tuning curve, with a sensitive tip at CF, a steep high-frequency skirt, and a gentler low-frequency tail. A fibre's CF is set by where along the cochlea its inner hair cell sits — tuning is tonotopy, read out one fibre at a time. Drag the CF below.[1982]

The sharpness of that tip is not passive: it depends on the cochlear amplifier. Lose the outer hair cells and the tuning curve broadens and its tip rises — the fibre becomes both less sensitive and less selective, the neural face of sensorineural hearing loss.

CSpontaneous rate & the rate-level function

Auditory-nerve fibres fire even in silence, at a spontaneous rate, and they fall into classes by it (switch the widget to the rate-level view). As level rises above threshold, firing rate climbs along an S-shaped function and then saturates. Only the rising part codes level unambiguously, so each fibre's usable dynamic range is limited — and it differs by class. High-spontaneous-rate fibres have low thresholds and narrow ranges; low-spontaneous-rate fibres have high thresholds and wide ranges.[1978]

CThe dynamic-range problem

Here is a puzzle with a clinical payoff. A single fibre saturates within perhaps 20–40 dB, yet we hear across more than 100 dB. How? By selective listening: soft sounds are carried by the sensitive high-SR fibres, and as those saturate the high-threshold low-SR fibres take over the loud end. Splitting the range across complementary fibre classes lets the population code an intensity range no single fibre could. Module 10 develops loudness coding in full.[2009]

CTwo-tone suppression & adaptation

Two more properties matter. Two-tone suppression: a second tone near CF can reducea fibre's response to its best tone — not neural inhibition but a cochlear nonlinearity, present in nearly every fibre and pervasive for complex sounds like vowels. Adaptation: a fibre fires hardest at a sound's onset and then settles to a lower sustained rate, emphasising changes and onsets — which carry much of the information in speech.[2012]

FTThe nerve the implant drives

Electric stimulation engages this same nerve, but reads and writes it differently. A current pulse excites spiral-ganglion fibres directly and synchronously — bypassing the hair cell, the synapse, and the slow chemistry — so the electrically-driven volley is far more tightly time-locked than an acoustic one (which is why the ECAP is so large). But electric stimulation is also blunter: it activates a broad swath of fibres at once rather than the cochlea's exquisitely narrow tuning, and it has no spontaneous-rate classes to share out the intensity range. Those two facts — broad excitation and the missing dynamic-range trick — underlie two of the central limitations of electric hearing: coarser frequency resolution and a much narrower dynamic range.[2009]

With the nerve's code in hand, the next two modules take its two great dimensions in turn: how it encodes intensity (Module 10) and frequency (Module 11).