In July 1991, Blake Wilson, Charles Finley, Dewey Lawson, Robert Wolford, Donald Eddington, and William Rabinowitz published “Better speech recognition with cochlear implants” in Nature (Vol. 352, pages 236-238). The paper introduced continuous interleaved sampling, or CIS, the sound-processing strategy that turned the cochlear implant into a device that lets many deaf users understand spoken language.
A cochlear implant bypasses damaged hair cells and stimulates the auditory nerve directly with electrodes placed in the inner ear. Earlier processors drove several electrodes with analog signals at the same time, which caused interactions between channels that smeared the information the nerve received. The CIS strategy instead delivered brief pulses to each electrode one at a time in rapid sequence, so that no two electrodes were active simultaneously. This non-overlapping stimulation sharply reduced channel interaction and produced large, immediate gains in speech recognition for the people tested.
The cochlear implant is the most successful neural prosthesis ever built, restoring a useful sense to hundreds of thousands of people, and CIS or its descendants became the standard way these devices encode sound. It stands as an early, working proof that an engineered signal could be substituted for a biological one at the interface to the nervous system.
For a general reader, this milestone is the practical ancestor of today’s brain-computer interfaces: long before electrodes were decoding speech from the cortex, a careful signal-processing idea was already writing intelligible sound into the auditory nerve.