Cochlear implants that use light sent sharper signals in gerbils

A cochlear implant that stimulates the hearing nerve with light, not electric current, produced more separate, more sound-like signals in the brains of gerbils. The study, from Tobias Moser’s group at the University Medical Center Göttingen and engineers at the University of Freiburg, was published on October 22 in EMBO Molecular Medicine.

The problem it is trying to fix

A standard cochlear implant has between 12 and 24 electrodes along the cochlea, the spiral where different nerve cells answer different pitches. The trouble is that current spreads through the cochlea’s fluid. As Moser wrote in IEEE Spectrum in 2022, it “causes broad activation of neurons across the frequency map of the cochlea.” Neighbouring channels blur together, so the brain gets a coarser picture of pitch than the number of electrodes suggests.

Light can be aimed more narrowly than current. The catch is that nerve cells do not respond to light on their own.

The design

So the team made them respond. An engineered virus carried the gene for ChReef, a light-sensitive protein, into the spiral ganglion neurons, the nerve cells that carry sound signals from the cochlea to the brain. The gerbils then received experimental implants with 5 to 10 miniature green LEDs.

The researchers recorded activity in the inferior colliculus, a relay station in the midbrain that keeps the cochlea’s pitch map. They compared three things: flashes from single LEDs, pulses from conventional electrodes, and ordinary tones played to the ear.

We could not read the full paper, which is behind a paywall. Nothing we could open gives the number of animals or the effect sizes. The account below comes from the abstract, the team’s earlier preprint and a detailed write-up in Hearing Health & Technology Matters.

What they found

The light worked at very low energy. Single channels drew a response from about 150 nanojoules, a tiny fraction of a joule. At higher intensity, one LED drove nerve activity as hard as a pure tone at 51 dB.

The channels stayed more distinct. Each LED lit up its own band of the brain’s pitch map, in the right order along the cochlea. A statistical test of how well each channel could be told from its neighbours came out better for light than for electricity, and light channels could be separated at smaller distances. The authors’ summary is that light stimulation “achieves improved spectral resolution over state-of-the-art electrical stimulation.” The responses looked more like those to real sound than electrical ones did.

The limits

Better than electrodes is not the same as natural hearing, and the light did not match the sharpness of real tones. At higher intensities some LEDs lit up more than one band. Some treated cochleae had fewer nerve cells than expected, which needs explaining before anyone tries this in people.

ChReef also switches off slowly, which may limit how well it follows fast changes in sound, and speech depends on timing. The study used single pulses, not the rapid streams a working implant would send. And it measured brain signals, not whether the animals heard better.

Moser co-founded OptoGenTech, a company set up in 2019 to bring the device to patients. We did not find the paper’s conflict-of-interest statement.

Why it matters

Nearly 737,000 cochlear implants had been fitted worldwide by 2019, by Moser’s count, and all of them stimulate with current. Device makers are working on the parts around it. MED-EL just launched an implant with no outside parts. Optical implants aim at the core problem, the blur, but they need gene therapy first. That is a field now reaching patients for one inherited deafness, as our OTOF brief reported.

What to watch next: in 2022 Moser wrote that the team aimed to begin a first-in-human study in 2026 to find the safest gene-therapy dose, with a phase 3 trial hoped for in 2028. We found no trial registration or announcement confirming that start.