Cochlear implants: how they help people hear, and what is still unproven

This article is general information, not medical advice.

A cochlear implant is an electronic device that can give a sense of sound to people who are deaf or have severe to profound hearing loss. It is considered when hearing aids give only limited benefit. The World Health Organization estimates that over 5% of the world's population, or 430 million people, need rehabilitation for disabling hearing loss (fact sheet dated 3 March 2026), and it lists cochlear implants among the hearing technologies used in rehabilitation.

Cochlear implants matter because they can help some people who would otherwise get little help from sound. They are also a useful case study for wider neurotechnology, since they show what happens when electrical signals are sent directly to a nerve. Anyone considering an implant should speak to a hearing service or specialist.

How it works

A National Institute on Deafness and Other Communication Disorders (NIDCD) page describes the implant as having an external part that sits behind the ear, with a microphone that picks up sound and a speech processor, and a second part that is surgically placed under the skin. An electrode array collects electrical impulses from the implant and sends them to different regions of the auditory nerve. The Gloucestershire Hospitals NHS Foundation Trust leaflet says the implant bypasses the damaged part of the inner ear, and that signals are sent directly to the hearing nerve and then to the brain.

The brain has to learn to interpret these signals. NIDCD says that hearing through an implant "is different from normal hearing and takes time to learn or relearn", and that an implant "does not restore normal hearing".

What is established in routine use

• Scale. NIDCD reports about 736,900 registered devices implanted worldwide as of December 2019, including roughly 118,100 in US adults and 65,000 in US children. NIDCD says these are FDA estimates based on manufacturer reports. The figures count devices rather than people, and they are several years old (the NIDCD page was last updated on 13 June 2024).
• Outcomes vary. NIDCD states that "not everyone performs at the same level with this device".
• Surgery. NIDCD says that surgical implantations are "almost always safe, although complications are a risk factor", as with any surgery.

Approvals reported by the manufacturer

These approvals are reported in Cochlear's own announcements. We have not checked them against FDA records, so they are reported claims, not checked regulatory records.

• Lower age limit. On 18 March 2020 Cochlear announced FDA approval to lower the age for implantation from 12 months to 9 months in the United States. The approved use is for children aged 9 to 24 months with bilateral profound sensorineural hearing loss who show limited benefit from hearing aids in both ears. Cochlear says the approval covers all current Nucleus implant models. NIDCD separately states that implants have been FDA approved for eligible children from 9 months of age since 2020.
• Single-sided deafness. On 11 January 2022 Cochlear reported that the FDA approved Nucleus implants for unilateral hearing loss and single-sided deafness. The company said implants were already approved for moderate to profound hearing loss in both ears, and that this approval let it offer cochlear implants, alongside its existing options, to people with hearing loss in one ear for the first time.

What studies show, and their limits

Music and noise. A 2004 review by Hugh McDermott summarises earlier melody tests, and the evidence here is old. In a 2002 study of 12 well-known tunes (Gfeller and colleagues, cited in the review), 49 implant users averaged about 19% correct, against about 83% for listeners with normal hearing. In a 2004 study (Looi and colleagues, cited in the review) of 15 Nucleus implant users, the users recognised about half of 10 familiar melodies, while normal-hearing listeners scored nearly 100%. The review also reports that implant users tend to rate the quality of musical sounds as less pleasant than listeners with normal hearing. It says many users of recent systems understand speech reasonably well even in moderately noisy conditions, while earlier feature-extraction processors had difficulty in situations with high levels of background noise. The review states that the feature-extraction approach to processor design "is now obsolete", so today's devices may perform differently.

A fully implanted device. Lefebvre and colleagues (2025) studied a totally implantable cochlear implant, which has no external processor. It was a non-randomised, single-group feasibility study of six adults, four of them women, at two centres in Liège and Munich. Participants were followed for 52 weeks. The primary outcome was adverse events. One anticipated serious device effect occurred, swelling at the implant site with signs of infection, and it resolved after treatment. Speech perception in quiet and in noise was comparable to that with the external SONNET processor. The authors say their conclusions can only be tentative because of the small sample. Three authors are employees of MED-EL, which funded the study and supplied the implants.

Light instead of electricity. Dieter and colleagues (2020) tested optical implants in gerbils. The animals' hearing nerve cells were made light-sensitive with a viral vector. Of 35 injected animals, 15 (about 43%) gave robust responses. Only 206 of 528 individual micro-LEDs (about a third) evoked responses. The experiments were acute, and the authors state that biocompatibility and long-term stability still need to be evaluated with chronic implantation. The paper discloses that two authors are co-founders of a company, OptoGenTech.

Open problems

• Variable results. Some people do well and others get much less benefit. The sources reviewed do not explain all of the difference.
• Surgery and recovery. Implantation is surgery, and NIDCD notes that complications are possible. The learning period after the implant is switched on is also part of the process.
• Batteries and replacement. The sources reviewed for this article do not cover battery life, processor upgrades or device replacement, so we give no figures. Questions about these points should go to an implant team.
• Cost and access. The Gloucestershire leaflet (reference GHPI1828_05_24, review due May 2027) says that to be considered, a person must have severe to profound hearing loss and receive only limited benefit from hearing aids. It also says: "In the UK, most adults are offered a cochlear implant for one ear. Those who are registered blind will be offered cochlear implants for both ears." This is one NHS trust's patient leaflet, not national guidance. NIDCD notes that in the US the process can be expensive and insurance may not always cover it. The sources reviewed here do not compare referral and funding across health systems.
• Music and noisy settings. As noted above, these remain difficult, and the evidence is dated.
• Small studies. The fully implanted device was tested in six adults without a separate comparison group, and the optical work was done in animals.
• Company-reported approvals. Several milestones come from manufacturer announcements, not from regulatory documents.

What to watch next

• Longer follow-up and larger studies, with comparison groups, for fully implanted devices.
• Chronic safety data for optical implants, and whether they work outside animal models.
• Updated device counts, since the most recent worldwide figure cited here dates from December 2019.
• Newer processing strategies tested specifically for music and speech in noise.