Cochlear Implant Atlas
CI Atlas · From Sound to Stimulation · Module 06

6CIS — the breakthrough

This is the past in 'past, present and future' — the moment sound coding turned a promising device into a working one. By the late 1980s multichannel implants existed, but driving their electrodes all at once, with continuous analogue currents, let the electric fields collide in the cochlea and smear the channels together; speech results were disappointing. The fix, continuous interleaved sampling, was deceptively simple: deliver brief pulses to the electrodes one at a time, in a rapid non-overlapping sequence, so no two fields are ever active at the same instant. The improvement in speech recognition was large and immediate, and it came not from new hardware but from a smarter way to use the hardware already there. CIS is the template most modern strategies still build on.

FTThe problem before CIS

Early multichannel processors used simultaneous analoguestimulation — several electrodes driven at once with continuous currents. In the conducting fluid of the cochlea, those currents summed and interacted, so the electrodes did not act as independent channels and the place code was blurred. The hardware promised many channels; the stimulation method threw much of that promise away.

CIS's key idea — fire the electrodes in turn, never together, so their fields don't collide

E1E2E3E4time →fields kept separate

Before CIS, driving electrodes simultaneously let their currents sum in the cochlea, smearing the channels together. Continuous interleaved sampling (Wilson and colleagues, 1991) delivers brief pulses in a rapid non-overlapping sequence, so only one electrode is ever active at an instant. The gain in speech recognition was immediate and large — a software change, not a new electrode, and the template most modern strategies still build on.

CThe interleaving idea

Continuous interleaved sampling (CIS) replaced continuous analogue currents with brief biphasic pulses, and — the key move — staggered them in time so that only one electrode is ever active at any instant. By interleaving the pulses, the fields never overlap in time, so they cannot sum: each channel is delivered cleanly, one after another, fast enough that the brain hears them as simultaneous.

The electrical output up close — a charge-balanced biphasic pulse train

phase widthamplitude = loudnessrate = pulses/s

Each pulse is biphasic — a negative phase followed by a positive one — so the net charge is zero and the tissue is not damaged. The programmer's knobs map onto its dimensions: amplitude sets loudness (within the narrow threshold-to-comfort window), phase width trades off with amplitude for charge, and the rate sets how finely the envelope is sampled in time. CIS, ACE and the rest are all just patterns of pulses like these. Schematic.

CThe result

The effect was dramatic. Wilson and colleagues reported large, immediate improvements in speech recognition with CIS over the simultaneous strategies of the day — a step change, published in Nature in 1991, that helped turn the cochlear implant into a reliably effective treatment. Crucially, it was achieved with the same electrodes: the gain came from how they were driven, not how many there were.[1991]

TWhy it still matters

CIS established two enduring lessons. First, channel interaction is the enemy, and managing it is central to good coding (Module 7). Second, the strategy matters as much as the hardware — a theme the history chapter also drew out (Chapter 1). Nearly every modern strategy, from ACE to fine-structure coding, is a descendant of CIS: pulsatile, interleaved, and built on the same insight.

One budget, shared — more channels means each is pulsed less often

total budget8 channels × 1,800 pulses/s each
Per-channel rate1,800 pulses/s

Because interleaving fires the electrodes one at a time, a fixed total pulse rate must be divided among the active channels. Use more channels and each is refreshed less often (a coarser temporal sample); use a high per-channel rate and fewer channels fit in the budget. This spectral-vs-temporal tug-of-war — visible in the SPEAK/ACE settings of Module 8 — is a direct consequence of the interleaving that made CIS work. Schematic.

Case 8.6 · A leap without new hardware
A historian of the field notes that around 1991 speech results from multichannel implants improved sharply, yet the electrode arrays had not changed. She asks what produced the leap.

What was the key change?

Self-assessment — Module 62 questions
Question 1 · Foundation

What problem did continuous interleaved sampling (CIS) solve?

Question 2 · Trainee

What is the enduring lesson of the CIS breakthrough?

Tracked locally in your browser — see /progress for the dashboard.