2From Acoustic to Electric: What Is Lost in Translation

FWhat the healthy cochlea gives for free

Normal hearing is effortless precisely because the cochlea does an enormous amount of work before the brain is involved. It separates a complex sound into its frequency components with great sharpness, it follows the fine timing of the waveform, and it squeezes an immense range of sound intensities into a form the nerve can carry. None of this is conscious, and none of it survives the loss of hair cells.

When we move a listener from acoustic to electric hearing, we are not simply making sound quieter or distorted. We are removing the cochlea’s signal processing and replacing it with a blunt instrument. The modules ahead examine the consequences one by one; this module names the four things being given up.[2001][2004]

TThe cochlear amplifier and its instantaneous compression

The outer hair cells are not passive. Through electromotility, driven by the motor protein prestin, they feed mechanical energy back into the basilar membrane, amplifying faint sounds far more than loud ones. The input-output function of the basilar membrane is therefore strongly compressive near the characteristic frequency: a soft sound is boosted, a loud one barely so. Damage or remove the outer hair cells and that function straightens toward linear, the response shrinks at low levels, and tuning broadens.

This compression is instantaneous and local, happening sample by sample at each place along the cochlea. It is the mechanism that lets a 1 kHz tone be both just-detectable and uncomfortably loud across roughly 120 dB while individual nerve fibres span only 20 to 40 dB. Electric stimulation has no equivalent. Whatever compression an implant provides must be added artificially in the processor, and it can only ever approximate the cochlea’s elegant, frequency-specific original.[2001][1997]

TCrude, broad, near-synchronous activation

Acoustic stimulation drives nerve fibres through hair-cell synapses, so each fibre fires somewhat independently, with stochastic timing that smooths the population response and carries fine temporal detail. A single electrode, by contrast, injects current that spreads through the conductive fluids and excites a broad swath of fibres at once. Place specificity is poor, so two electrodes can sound similar in pitch, and the population fires in near-lockstep with each pulse rather than with the natural, jittered independence of acoustic firing.

This highly synchronous, spatially broad activation is efficient at making sound audible but impoverished in detail. It collapses the cochlea’s fine frequency map into a handful of overlapping channels and replaces graded, stochastic firing with a coarse, deterministic pulse-locked response. Sharp tuning, independent channels and rich temporal fine structure are the casualties.[2004][1983]

CFraming the perceptual consequences

Each thing lost has a perceptual signature the clinician will recognise. The missing cochlear amplifier shows up as a narrow, steeply growing loudness range that the processor must compress into. The broad electrical spread shows up as a limited number of truly distinct channels and as channel interaction and masking. The loss of fine temporal detail shows up as poor music and pitch perception and difficulty in noise.

Framing the whole chapter this way gives a single organising idea: electric hearing is acoustic hearing minus the cochlea’s processing. Everything the rest of the chapter measures, and everything the engineer and audiologist later try to restore, is an attempt to compensate for one of these four deficits in software and electrode design.[2004][2008]