10Coding of intensity & loudness
How does the nervous system represent how loud a sound is? It has two tricks: each fibre fires faster as a sound gets louder, and louder sounds also recruit more fibres into action. Between them they let us hear across a range of more than a hundred decibels — a range that collapses in hearing loss and is rebuilt, very differently and much more narrowly, by a cochlear implant. The loudness physiology in this module is the direct ancestor of the T and C levels every implant is programmed with.
TCTwo codes for intensity
Loudness is carried by two neural codes working together. The first is firing rate: as a sound grows louder, each auditory-nerve fibre fires faster, up to saturation (Module 9). The second is recruitment of fibres — not to be confused with the clinical sign of the same name — meaning that louder sounds excite a wider population: more fibres at the centre of the excitation, and a spreading skirt of fibres tuned to neighbouring places. Total loudness reflects the summed activity across the whole responding population.[2012]
This is why no single fibre's saturation limits our hearing: even after the most sensitive fibres have maxed out, rising level keeps recruiting higher-threshold fibres and widening the active region, so the population code keeps growing.
TCHow loudness grows
Perceived loudness does not grow in proportion to sound pressure or even to decibels; across most of the range it grows roughly as a power functionof intensity (Stevens' law), so that a 10 dB increase corresponds approximately to a doubling of loudness. Loudness also sums across frequency: a sound spread over many frequency bands is louder than the same energy in one band, because it engages more of the cochlea. These facts make loudness a property of the whole pattern of neural activity, not of any single channel.[2013]
CThe dynamic range
The usable dynamic range of normal hearing — from the faintest audible sound to the loudest tolerable one — exceeds 100 dB. Speech occupies only about a 30 dBwindow within it (from soft high-frequency consonants to loud low-frequency vowels). In sensorineural hearing loss the range shrinks from both ends: the threshold rises while the uncomfortable level changes little, so the whole of speech must be packed into a much smaller window — the core problem a hearing aid's compression, or an implant's loudness mapping, has to solve.[2009]
FTRecruitment in hearing loss
When outer hair cells are lost, loudness growth becomes abnormal: thresholds rise, but once a sound is audible its loudness climbs unusually fast toward discomfort — clinical recruitment (introduced in Module 7). Compare the three regimes below: the compressive, wide-range normal curve; the steep, squeezed recruiting curve of cochlear loss; and the very narrow electric window.
FTElectric loudness coding
A cochlear implant codes loudness mainly by the amplitude of the current it delivers (and, to a lesser extent, by pulse rate and the number of electrodes active). But the range between a barely-audible current and an uncomfortably-loud one — between threshold (T) and the comfortable maximum (C) — is strikingly small, and loudness grows steeply, even expansively, across it. Electric hearing therefore reverses the cochlea's compression: where the normal ear squeezes a wide input into a comfortable loudness, the implant must do that squeezing in the processor, because the nerve's electrical loudness window is so narrow.[2009]
Crucially, electric loudness growth — though steep — is set by the device, not by damaged recruiting hair cells. Within the T–C window the clinician controls the mapping, so loudness can be made orderly and comfortable in a way a recruiting acoustic ear cannot. The implant trades the cochlea's effortless compression for a narrow but fully controllable range.
TThe physiology behind T and C
Everything in this module reappears in the programming booth. Setting T and C levelsis precisely the act of measuring the electrical dynamic range of this module; the processor's input dynamic range and compression are how the wide acoustic range is mapped into it; and the electrical stapedius reflex is an objective read-out of where comfortable loudness sits. The objective-measures chapter (Module 9 there) is, in this sense, applied loudness physiology.[2009]
Intensity is one of the nerve's two great dimensions. The other is frequency — and how the nerve codes pitch, by place and by timing, is the next module.
How does this electrical loudness behaviour compare with normal acoustic loudness coding, and why?
Besides each fibre firing faster, how else does the auditory nerve signal that a sound is louder?
Roughly how large is the usable dynamic range of normal hearing?
Why does a cochlear implant have a very narrow electrical dynamic range (small T–C window)?