4Steep and Fast: How Loudness Grows with Electric Current

FFrom eighty decibels to a handful

A normally hearing ear moves from the faintest audible sound to an uncomfortably loud one across roughly eighty decibels. Drive the auditory nerve directly with electric current and that comfortable range collapses to something on the order of only a few decibels of current change between threshold and the upper comfortable level. The whole usable span of electric stimulation is therefore extraordinarily narrow.

The reason is simple: an electric pulse delivered through an implant electrode skips the basilar membrane and the outer hair cells entirely. Those structures normally provide a gentle, level-dependent gain that is large for soft sounds and shrinks for loud ones, compressing a vast acoustic range into the neural code. Bypass them and the compression is gone, so a small change in current produces a large change in the number and firing rate of the neurons recruited, and hence a large change in loudness.[2020]

TPower, exponential, and what charge has to do with it

Plotted on log-current versus log-loudness axes, the electric loudness function is far steeper than the acoustic power law (whose Stevens exponent is near 0.3). Investigators have modelled the electric growth as a power function, as an exponential of current, or as a hybrid of the two; for most users any of these fits reasonably over a restricted range, but the exponential form captures the shallow tail near threshold and the rapid rise toward comfort more consistently.

Loudness depends on charge, but not purely on charge. The membrane is a leaky integrator: lengthening a pulse delivers more charge yet some leaks away during the pulse, so doubling current raises loudness more than doubling pulse duration. Loudness-balancing studies confirm that halving current is not offset by doubling duration, and that the effect is level dependent. This is why current amplitude, not pulse width, is the primary loudness control in clinical fitting.[2020][2022]

CRate, pulses, and summation add up

Loudness is not set by single pulses alone. The central auditory system integrates excitation over a window of about seven milliseconds, so as pulse rate rises above roughly 100 pulses per second more pulses fall inside that window and the sound grows louder; current must be lowered to keep loudness constant when rate increases. Working against this, neural refractoriness means each pulse in a fast train recruits fewer fibres, so the net rate effect is a moderate rise in loudness.

Loudness also sums across electrodes. When several channels are active within a stimulation period, the combined percept is louder than any single channel, and the current on each must be reduced, sometimes by several decibels, to match a single-channel reference. This summation is larger near threshold and is only weakly affected by electrode spacing, a direct consequence of overlapping neural excitation rather than simple current addition.[2020]

CSetting C and M levels on a steep slope

Because the slope is so steep, every microamp matters. A small overshoot when setting the upper comfortable level (C in Cochlear terminology, M in others) can push a channel from pleasant to startling, while an undershoot wastes audible range. Clinicians therefore loudness-balance neighbouring channels so that a sweep across the array sounds even, and set the upper levels to a consistent, comfortable loudness judgement rather than a fixed current.

The steep growth also explains why electric hearing is unforgiving of mapping errors and why automatic gain control upstream is essential: the processor must squeeze a wide acoustic input into this narrow electric corridor before the steep loudness function ever sees it. Knowing the shape of the function lets the audiologist anticipate that a channel set even slightly high will dominate the percept and that small, deliberate current steps are the right tool during programming.[2022]