Hearing the Light

Tobias Moser, Daniel Keppeler, Christian Goßler and Ulrich T. Schwarz

Through optogenetics and new medical devices, optical stimulation of the auditory nerve could improve on current hearing-restoration technology, affording better speech recognition and perception of music.

figure[Getty Images]

More than 5% of the world population—432 million adults and 34 million children—suffer from disabling hearing impairment, making it one of the most frequently reported sensory deficits. Left untreated, hearing loss impacts one’s ability to communicate and is thought to pose a yearly global cost of US$750 billion. Hearing impairment reduces chances in the job market, causes social isolation and increases risks of depression and cognitive decline.

Most hearing impairment is so-called sensorineural hearing impairment. This condition stems from disorders of the cochlea—the spiral-shaped structure in the inner ear that receives sound vibrations through the middle ear and converts them via specialized “hair cells” to electrical signals for the brain. In addition to genetic causes, sensorineural hearing impairment frequently results from damage to the cochlea over the course of one’s life due to loud noise, drugs, ischemia, trauma or infection, which commonly causes degeneration of sensory hair cells and auditory neurons.

figure[Adapted from Getty Images] [Enlarge graphic]

For the foreseeable future, hearing aids and electrical cochlear implants will remain key means of hearing rehabilitation for most individuals, depending on the hearing impairment’s severity and causes. Hearing aids mainly address moderate hearing impairment, when speech recognition becomes limited. These devices analyze surrounding acoustic signals and provide the ear with an amplified, preprocessed version, via an acoustic speaker or bone coupling.

More profound hearing impairment or deafness that hearing aids can’t address calls for cochlear implants. These surgically implanted devices electrically stimulate the spiral ganglion neurons (SGNs)—the nerve cells that give rise to the auditory nerve. Thus the implants bypass the hair cells in the cochlea that stimulate the SGNs in normal hearing, but that are absent or dysfunctional in cases of hearing impairment.

The electrical cochlear implant is perhaps the most successful neuroprosthesis, with more than 0.7 million devices implanted worldwide to date. Yet hearing with cochlear implants is far from normal. While they can enable speech comprehension in quiet environments, users have trouble understanding speech in the noisy environments of daily life, and tracking melodies and appreciating music are challenging. This is commonly attributed to the wide spread of the electric current from each electrode contact in the implant, which restricts the frequency resolution of the sound encoding.

The improved spatial confinement and frequency resolution possible with an optical signal, rather than an electrical one, could offer a better way. In this feature, we look at approaches toward hearing restoration that combine optogenetics and specialized optical cochlear implants—focusing in particular on recent work in our own labs from proof-of-principle to clinical translation.

figureThe minute clusters of hair cells populate the spiral cochlea in mammals and detect sounds transmitted to the inner ear. Below these hair cells are connections to the auditory nerve. [D. Spears FRPS FRMS/Getty Images]

Shortcomings of current implants

To understand some of the problems of electrical cochlear implants, it helps to think of the cochlea as a circular staircase, with each step corresponding to an acoustic frequency. The sophisticated cochlear micromechanics lets humans discern about 2,000 steps (frequencies) along the cochlea’s so-called tonotopic axis. Each SGN is stimulated by one inner hair cell, and the sound frequency that best excites the neuron depends on the hair cell’s “place code”—its position along the tonotopic axis.

Because of the spread of the electric current from the electrode contact, conventional cochlear implants make poor use of this place code, and thus their sound encoding has poor frequency resolution. Activating one of the device’s 12 to 24 electrode contacts will activate many steps—that is, many different frequencies at a time. Moreover, neighboring electrode contacts can stimulate overlapping neuron populations, further limiting frequency discrimination.

To understand some of the problems of electrical cochlear implants, it helps to think of the cochlea as a circular staircase, with each step corresponding to an acoustic frequency.

A cochlear implant’s direct stimulation of the SGNs, as opposed to stimulation through the hair cells in the cochlea, comes with an additional disadvantage: the dynamic range of stimuli over which the physiological and perceptual response changes is considerably smaller than in normal hearing. The wide dynamic range of ordinary physiological sound encoding traces to a variety of mechanisms: amplification of weak and dampening of strong vibrations by the basilar membrane (the structure carrying the sensory hair cells in the Organ of Corti); adaptation by the hair cells, synapses and neurons; and “tiling” of the dynamic range via intensity coding by the SGNs. The neuron encodes the increased volume by sensing a change in the rate of stimulation (attributable to loudness) and by recruiting other neurons in the immediate neighborhood of the stimulated frequency spot.

Electrical excitation’s broad tonotopic spread means that electrical cochlear implants don’t capitalize on these mechanisms. Thus, while the devices can achieve good intensity discrimination in the lab, their limited discrimination of dynamic range makes it hard to understand speech amid the noisy background of real life. While efforts are afoot to improve cochlear-implant performance, for example via current steering, the potential for reducing the spread of electrical excitation seems rather limited.

Optogenetic treatments for sensory deficits

As light can be better spatially confined than electrical excitation, optical stimulation of the auditory nerve could well overcome the performance bottleneck of electrical cochlear implants. Our work, and that of others, suggests that future optical cochlear implants could increase the spectral selectivity of artificial sound encoding—and, thus, might improve speech recognition in background noise, as well as processing of melodies and music.

Two concepts for optical stimulation of the auditory nerve are currently being pursued. Work led by Claus-Peter Richter at Northwestern University, USA, aims at employing infrared direct neural stimulation, thought to activate neurons via a photothermal effect. The other mechanism, optogenetic stimulation of the auditory nerve, is being pioneered in our lab in Göttingen, Germany. This method involves photosensitizing of the auditory nerve by genetically modifying the cells to express light-gated ion channels.

Combining as it does genetics and optical stimulation for controlling cells with light, optogenetics has been a disruptive technology in the biology lab over the past two decades, with clinical applications only now starting to emerge (see “Optogenetics: Controlling Neurons with Photons” OPN, April 2018). The potential for optogenetics in addressing hearing impairment in humans can be inferred from the recent first clinical application of the technique, which dealt with problems with another sense—vision.

Unlike the ear, the eye currently has no means of sensory restoration comparable to a cochlear implant. Retinal implants typically have not restored vision above the criterion of legal blindness, and are no longer being produced. The pressing unmet need for vision restoration in degenerative disorders of the eye has remained, however, and the development of gene-therapy techniques with viral vectors, coupled with light stimulation mechanisms, has driven the development of optogenetic treatments for the eye.

The development of optical cochlear implants builds on long-standing experience with implementing and improving hearing rehabilitation via electrical implants.

These efforts began with a gene therapy safety trial, conducted by Allergan, that used an adeno-associated virus (AAV) as a vector to insert genes expressing the light-sensitive protein channelrhodopsin 2 in retinal output neurons. That was followed by the more recently launched PIONEER trial, led by GenSight Biologics, which has now provided a preliminary report indicating efficacy. The PIONEER trial has employed a combination of AAV-mediated gene therapy for expressing the red-light-activated channelrhodopsin Chrimson in retinal output neurons, coupled with a light-amplifying goggle, to address a specific, rare genetic cause of blindness, retinitis pigmentosa (see “Partial Sight Restoration via Optogenetics,” OPN,

From retina to cochlea

As with vision, optogenetic approaches to hearing restoration require multidisciplinary collaboration, as they combine gene therapy and a medical device: the optical cochlear implant. Moreover, cochlea and retina share advantages as targets for gene therapy; both require only tiny AAV doses, and both are “immune privileged,” occupying microenvironments in which the body’s immune response is locally inhibited. Yet optogenetic manipulation of the SGNs holds some specific challenges, as the neurons’ cell bodies, or somata, “hide” in the cochlea’s bony core.

figureAn optogenetic hearing restoration system (left) would use a surgically implanted optical device (center), similar to an electrical cochlear implant, for spectrally selective optical stimulation of auditory neurons (right) that have been genetically modified express light-sensitive proteins, or channelrhodopsins. [Courtesy of the authors]

To an extent, the development of optical cochlear implants builds on long-standing experience with implementing and improving hearing rehabilitation via electrical implants. Audio processing, transcranial signal and power transmission techniques pioneered for electrical cochlear implants—as well as the atraumatic insertion of a flexible, well-encapsulated stimulating implant into the fluid-filled natural cavity of the cochlea—can be similarly implemented for optical ones.

Switching to optical stimulation does, however, create both advantages and challenges. Further, translation of optical cochlear implants into the clinic will require demonstration of the feasibility, safety and efficacy of both the medical device itself and the required gene therapy. And a performance improvement of optical sound encoding over electrical sound encoding must be demonstrated—since clinical translation of the more complex optical cochlear implant can only be justified if it promises a major improvement of hearing restoration.

figureIn a preclinical µLED-based optical cochlear implant system tested in mice, a head-worn system includes a wirelessly controllable audio processor with a MEMS microphone, driver electronics and a battery board (rectangular inset, bottom to top), as well as a µLED array (blue) that is surgically implanted in the animal’s cochlea. [Courtesy of the authors]

Early progress

Proof-of-principle studies and further work with rodents, much of it in our labs, have made considerable progress toward these goals. One key step has been the demonstration of safe and stable, AAV-mediated expression of light-sensitive channelrhodopsins in the SGNs of mice, rats and gerbils upon intracochlear application of a single dose of the viral vector. Optogenetic stimulation of the genetically altered auditory nerve in these animals has been shown to activate the auditory pathway up to the cortex and to result in behaviors consistent with auditory perception.

To take progress further, we have developed multichannel optical cochlear implants that use microscale blue-light-emitting diodes (µLEDs) to stimulate optogenetically modified neurons. A low-weight (less than 15 g), battery-driven, wirelessly controllable system captures and processes the sound externally and wirelessly operates the implanted µLEDs. Behavioral experiments suggested that the system restored hearing in rodent models of human deafness.

But how well does the system’s performance compare with electrical cochlear implants? To find out, we benchmarked the optical cochlear implants against a sister electrical system that differs only in the driver electronics and that employs clinical-style electrode arrays for stimulation. Both computer simulations and live-animal experiments suggest that the spectral selectivity of optogenetic stimulation approaches that of normal hearing, and dramatically exceeds that of state-of-the-art electrical cochlear implants. The experiments thus corroborated the near-physiological spectral resolution potentially realizable via µLED-based multichannel optical cochlear implants.

figureElectrophysiological recordings from the auditory midbrain of rodents showed that, upon stimulation by a pure tone, the spectral confinement of the neural activity elicited by optogenetic stimulation of sensory ganglion neurons from a 200-µm optical fiber (center) was comparable to that in normal acoustic hearing (left). In contrast, monopolar electrical stimulation (right), as with a conventional electrical cochlear implant, resulted in broad activation of the cochlea and, consequently, the midbrain. [Courtesy of the authors]

Further, it is worth noting that the optical cochlear implant implementations tested thus far have not yet been optimized for spatially selective stimulation. The system we tested uses blocks of four 50-µm-sized µLEDs or single LEDs of 200×250-µm size, without any optical shaping of the beam from these Lambertian emitters. Determining the maximum achievable spectral selectivity will require future experiments employing emitters with narrow Gaussian beam profiles.

To this end, we have recently developed experimental red-light optical cochlear implants that combine laser diodes and polymer waveguides. Red-light stimulation—which has become possible due to the recent discovery of the channelrhodopsin Chrimson—not only promises better spectral selectivity (less scattering), but reduces the risk of potential phototoxicity to the nerve cells.

Temporal fidelity: The need for speed

The slow off-kinetics of the red-light activated Chrimson, however—the closing time constant of which, after illumination is switched off, is 25 ms at room temperature—highlights a key issue that our lab and others have had to address in developing optogenetic hearing restoration. SGNs fire hundreds of spikes per second, and with sub-millisecond precision, allowing the spike timing to faithfully report the temporal fine structure of the stimulus. This raises the risk, in optogenetic hearing restoration, of trading the excellent temporal fidelity of SGN stimulation amenable to electrical cochlear implants for the high spectral resolution of optical cochlear implants.

figureFast Chrimson can provide near-physiological firing rates in optogenetic stimulation. At a stimulation frequency of 100 Hz, each stimulus (amber bars) elicits a neural spike (magenta). At higher rates (300 and 500 Hz) adaptation sets in; the neurons cannot follow each of the presented stimuli. [T. Mager et al., Nat. Commun. 9, 1750 (2018); CC-BY 4.0]

Therefore, we and others have engineered and applied channelrhodopsins with ultrafast deactivation, to enhance the temporal fidelity of optogenetic sound encoding. For Chrimson specifically, variants have been created that show fast (3 ms at body temperature) to ultrafast (1 ms at body temperature) deactivation. These Chrimson variants should be able to drive SGNs reliably and with good temporal precision up to stimulation rates of approximately 200 Hz.

Some channelrhodopsins—such as the blue-light-activated Chronos, which has a time constant of 0.7 ms at body temperature—can, in principle provide even faster kinetics. The shorter the channel’s open time, however, the less charge transfer is yielded per photon absorption, which translates into bigger power budget. Thus optogenetic hearing restoration needs to balance temporal fidelity and the power budget of coding. Limiting the stimulation rate to 200 Hz seems justified based on what’s known from lab studies of speech understanding as a function of stimulation rates in electrical cochlear implants.

Beyond inherently improved spectral resolution and comparable temporal fidelity, can optical cochlear implants improve on the limited intensity coding of electrical implants on the single-neuron level? Recent investigations revealed a 1-7 dB (mW) dynamic range for optical stimulation, compared with 1-2 dB (current) for electrical stimulation. Because the two stimulation modalities are difficult to compare directly, future experiments on the neural-population level or behavior will need to compare optogenetic stimulation to actual acoustic hearing.

Toward clinical trials

While early progress on optical cochlear implants is encouraging, much remains to be done before the first human trials, currently planned to start before the end of 2025. As the product combines gene therapy and a medical device, both components need to be further developed and tested—and the safety and efficacy of the combination has to be assessed.

Gene therapy

On the gene-therapy side, continued work must focus on designing the best genetic construct to render the auditory nerve light sensitive—that is, one that enables robust optogenetic activation of SGNs, at physiological firing rates and with low light requirements. Also, the viral vector used to transduce this optogenetic code into the SGNs must be optimized for safe, efficient delivery of its genetic payload.

AAVs will likely be the vectors of choice, since they enable foreign protein to be expressed at high levels and over long periods. Further, AAVs have little risk for adverse reactions of the target cells, have been used in various clinical trials for gene therapy (including retinal dysfunction), and have been successfully used to restore auditory function in animal models of deafness. Implementing adequate promoters and applying the virus locally in the ossified cochlea will help limit the transduction to the SGNs, and thus exclude opsin expression beyond the auditory nerve.

Inspired by retinal optogenetics, tissue explants and inner-ear organoids might help evaluating the construct of choice and optimize efficiency and specificity in human SGNs. Upon successful implementation of the gene-therapy part of optogenetic hearing restoration, the effect of chronic illumination of intracochlear neural tissue needs to be evaluated. Safety limits for light intensity will need to be defined, to avoid tissue heating or phototoxic effects. These considerations should foster the conservation of cochlear structures and off-target tissues, contributing to safety for the patient.

The medical device

For the medical-device component, while communication and processor technology can readily be adapted from electrical cochlear implants, restrictions in the temporal domain due to opsin kinetics, as discussed above, and the increase in stimulation channels relative to electrical devices will likely require novel coding strategies.

An ambitious research agenda will be needed to move optical cochlear implants from the lab to the clinic. But we believe that the result will be well worth the effort.

The minimal duration and intensity of a light pulse sufficient for SGN activation will depend on channel kinetics of the yet-to-be-defined optogenetic tool. The sound intensity coding will need to use the full range of optogenetically evoked SGN activation at the neurons’ natural firing rates until the upper limit is reached, where firing saturates or the behavior of the experimental subject signals discomfort. The energy of a single pulse will need to be balanced between its duration (which needs to be limited to allow high stimulation rates) and its intensity (which is restricted by biosafety limits), and must also allow for reasonable battery lifetimes of for the optical cochlear implant. The patterns of optical sound encoding will then need to be mapped onto a set of independent optical stimulation channels larger than the number for electrical implants to encode spectral information. A range of 50 to 100 optical emitters seems a reasonable target to aim for.

All of this suggests an ambitious research agenda to move optical cochlear implants from the lab to the clinic. But we believe that the result—a richer, more natural experience of conversation and music for patients with hearing impairments—will be well worth the effort.

This work was funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreement no. 670759 – advanced grant “OptoHear”) to T.M.

Tobias Moser and Daniel Keppeler are with the University Medical Center Göttingen, the German Primate Center and the Max Planck Institute of Experimental Medicine, Germany. Christian Goßler and Ulrich T. Schwarz are with the Technical University of Chemnitz, Germany. All authors are co-founders of the start-up company OptoGenTech GmbH, Göttingen, Germany.

References and Resources

  • “Deafness and hearing loss,” World Health Organization,

  • L.M. Friesen et al. “Speech recognition in noise as a function of the number of spectral channels: Comparison of acoustic hearing and cochlear implants,” J. Acoust. Soc. Am. 110, 1150 (2001).

  • A.D. Izzo et al. Laser stimulation of the auditory nerve, Lasers Surg. Med. 38, 745 (2006).

  • H. Fastl and E. Zwicker. Psychoacoustics: Facts and Models (Springer-Verlag New York, 2007).

  • V.H. Hernandez et al. “Optogenetic stimulation of the auditory pathway,” J. Clin. Invest. 124, 1114 (2014).

  • N.C. Klapoetke et al. “Independent optical excitation of distinct neural populations,” Nat. Methods 11, 338 (2014).

  • H. Ahmed et al. “Emerging gene therapies for genetic hearing loss,” J. Assoc. Res. Otolaryngol. 18, 649 (2017).

  • C. Wrobel et al. “Optogenetic stimulation of cochlear neurons activates the auditory pathway and restores auditory-driven behavior in deaf adult gerbils,” Sci. Transl. Med. 10, eaao0540 (2018).

  • A. Dieter et al. “Near physiological spectral selectivity of cochlear optogenetics,” Nat. Commun. 10, 1962 (2019).

  • J.-A. Sahel et al. “Partial recovery of visual function in a blind patient after optogenetic therapy,” Nat. Med. 27, 1223 (2021).

For extended references and resources, go online:

Publish Date: 01 October 2021

Add a Comment

Share this Article