From a proof of concept to a system
The first cochlear implants showed that electrical stimulation of the hearing nerve could give a deaf person a sense of sound. They did not show that the sound was good enough to understand speech. The following decades of work were mostly engineering: how many electrodes to use, where to put them, how to turn sound into electrical patterns, and how to make the equipment small enough to wear all day.
Why one electrode was not enough
The inner ear’s cochlea is a fluid-filled spiral that sorts sound by pitch. High pitches stimulate the base of the spiral, and low pitches stimulate the far end. A single electrode excites a broad stretch of nerve fibers at once, which gives the person a sense of rhythm and loudness but very little sense of pitch.
Single-channel devices of the 1970s and early 1980s, notably those developed in Los Angeles by William House and engineers, were useful for awareness of sound and for lip-reading support. Some users could also recognize a few words in good conditions. Understanding open speech without visual cues was uncommon.
The multichannel idea
The next step was to place several electrodes along the cochlea so that different electrodes could stand for different pitches. Several groups worked on this at once. In Melbourne, Graeme Clark’s team developed a multi-electrode array and an implant that was first used in a patient in 1978. In Vienna, Erwin and Ingeborg Hochmair developed a multichannel device in the late 1970s. Researchers at Stanford and in San Francisco explored multielectrode designs as well.
Multichannel implants raised new problems. Current from neighboring electrodes spreads and overlaps in the fluid-filled cochlea, so two electrodes can blur into each other. Designers had to decide on the shape of the electrode carrier, how deeply it should go, and how to avoid damage during insertion. Different groups chose different answers, and the choice of design remained a topic of debate for years.
By the mid-1980s, multichannel systems from Australia received approval in the United States for adults, and by around 1990 approvals had broadened to include children. Exact dates differ by country and device.
Speech processing strategies
Hardware alone did not decide success. The processor that converts sound into pulses mattered at least as much. Early systems tried to preserve the sound wave itself as an electrical signal on a few channels. Others extracted particular features of speech, such as the main frequencies in vowels, and sent only those. These feature-based approaches improved results in some users.
A further change came in the late 1980s and early 1990s, when researchers including Blake Wilson and colleagues reported strategies that used a bank of filters and sent very rapid pulses to different electrodes one at a time. Because the electrodes were never activated simultaneously, the current spread was less of a problem. These interleaved strategies, and later variations that selected the strongest frequency bands in each moment, are the basis of many modern systems. Some people who had received implants for years improved greatly after switching to a better strategy using the same implanted hardware, which is one reason the processor was designed to be replaceable.
Making the external parts smaller
Early users carried a body-worn processor, about the size of a deck of cards or larger, connected by a cable to a coil on the head. The implanted part picked up signals through the skin using magnetic coupling and radio waves. Advances in silicon chips and low-power electronics made the processor smaller. In the 1990s, behind-the-ear versions became available, and later some designs merged microphone, battery and processor into one small piece. Recent models include wireless links to phones and other devices.
Battery size and life remained a limit. Behind-the-ear units run on small batteries or rechargeable cells, and designers had to balance size against hours of use.
Electrodes, surgery and preservation
Improvements to the implanted part were less visible but important. Slimmer, softer electrode carriers were designed to reduce damage to the delicate inner ear, with the hope that any natural hearing left in the low frequencies could be preserved. This led to systems that combine electrical stimulation for high pitches with ordinary acoustic amplification for low pitches in people with partial hearing. Surgical techniques also changed, with smaller incisions and more standardized approaches. As with other implantable devices, long-term reliability depends on sealing the electronics against body fluids and designing connectors that survive for decades.
What continued to develop
Modern research aims to improve hearing in noise and music, which remain hard for many users. Some manufacturers now use machine learning in sound classification, a form of artificial intelligence, to adjust settings automatically in different environments, although the effect varies between people. Others are exploring electrode designs that reduce the overlap between channels or that deliver drugs to protect the nerve. Compared with the early days of cardiac pacing, which also began with large external boxes and gradually shrank, the bionic ear followed a similar path of miniaturization.
Outcomes differ greatly from person to person. They depend on age at implantation, how long the person was deaf, the condition of the hearing nerve and the amount of rehabilitation, so the figures that appear in marketing should be read cautiously.