What if a cochlear implant could be retuned using the patient’s own sense of pitch, so the implanted ear sounds more like the normal-hearing ear? This study tests that idea in twelve people with single-sided deafness.

Rehabilitating single-sided deafness has been debated in comparative studies and systematic reviews. Many patients experience difficulty localizing sound sources, understanding speech in noisy environments, and enjoying concerts or live music.

Without the head-shadow effect, binaural squelch, and binaural redundancy effect, everyday listening situations can prove exhausting for both children and adults. Mild to moderate tinnitus in the affected ear can also impair quality of life.

Cochlear implantation aims to rehabilitate the hearing of the impaired ear rather than transfer information to the contralateral side, restoring access to binaural cues. In adult patients with single-sided deafness, cochlear implantation has been demonstrated to improve sound localization and speech comprehension in noisy environments, while also significantly reducing tinnitus severity.

The perceived pitch of a tone depends not only on the place of stimulation in the cochlea, called tonotopy, but also on the stimulation rate of the vibration. A mismatch between frequency and place encoding may negatively affect auditory processing.

Using place-dependent stimulation rates in a cochlear implant has been demonstrated to improve tonotopic pitch perception in patients with single-sided deafness. Involving patients in optimizing their frequency maps and using their subjective pitch sensation may serve as a valuable tool for the audiologist.

Patients with single-sided deafness and bimodally aided cochlear implant users may be ideal candidates because they are motivated to minimize disturbing interaural frequency mismatch and achieve more natural bilateral hearing and better speech comprehension, especially in noisy environments.

The study included twelve patients with acquired postlingual single-sided deafness: six male and six female, with pure-tone air conduction thresholds of twenty-five decibels hearing level or less in the non-implanted ear. All participants had CI five twenty-two or CI six twenty-two cochlear implants and Nucleus CP nine-ten, CP nine-fifty, or CP one-thousand speech processors.

They were experienced users with one to five years of daily usage and eight to seventeen hours of use per day. Although the participants could differentiate the pitches of all twenty-two stimulating channels, they were not satisfied with the sonority and general tone of their devices or with speech comprehension in noise, especially chattering speech-noise.

Other fitting methods had not provided satisfying results, including changing to NRT-based mapping, changing coding strategy, lowering stimulation rate, or lowering the number of maxima. Figure one outlines the interaural pitch-matching setup, linking the cochlear-implanted side to a computer running Custom Sound Pro and the normal-hearing side to an active loudspeaker.

An iPad frequency generator supplies the acoustic signal, while an oscilloscope is also connected in the signal path. This arrangement matters because it lets the authors compare electrical stimulation on the implanted side with controlled acoustic stimuli presented to the normal-hearing side.

The tested cochlear-implant channel was activated in Custom Sound Pro six point three using intermittent pulse trains at a rate of nine hundred pulses per second, with a stimulus duration of two hundred fifty milliseconds and an inter-stimulus duration of two hundred fifty milliseconds.

The loudspeaker volume for the acoustic tone on the contralateral side was adjusted to the perceived level of the cochlear-implant electric stimulus, based on the patient’s subjective discernment. The audiologist then adjusted the pitch of the continuous acoustic tone to approximate the perceived pitch of the electric stimulus, relying on the patient’s feedback.

Patients were instructed to compare pitches, not volume levels, so the task tested pitch matching rather than loudness matching. Figure two plots measured central frequencies for Patients one through twelve across channels twenty-two, seventeen, eleven, five, and one, alongside star markers for the default FAT.

The authors obtained each channel’s average central frequency from at least four to five randomized approximations between zero point one and twenty kilohertz, with patients’ eyes closed to reduce subjective bias. This visual matters because it shows how individually measured channel frequencies compare with the factory default values.

Figure three shows measured central frequencies for Patient one across channels twenty-two, seventeen, eleven, five, and one, with a third-degree polynomial fitted through those points. The inset reports the fitted equation and an R-squared value of zero point nine nine nine nine three, indicating how closely the curve represents these measurements.

This fitting step matters because the procedure used such approximations to calculate a new frequency allocation for the cochlear-implant channels. Figure four gives an example of frequency allocation table generation, plotting frequency against channel number.

The squares represent the applied new central frequencies for Patient one, while the stars show the central frequencies from the default FAT; both sets span channels twenty-two through one and rise toward higher frequencies. This illustrates how the authors rearranged channel frequencies while preserving the map’s overall frequency structure, supporting the later evaluation of hearing outcomes after the map switch.

The pre-fitting and post-fitting PTA four measurements showed no difference, with a mean change of minus zero point thirty-one decibels and a standard deviation of two point ninety-seven decibels. This result substantiates that changing the frequency allocation table does not affect hearing thresholds when the T-levels were correctly set before the procedure.

The SRT fifty percent measurements showed a similar outcome, with an average difference of zero point twenty-one decibels and a standard deviation of three point seventy-one decibels. The important change appeared in the WRS at sixty-five decibels: word recognition score demonstrated a significant improvement with the experimental map, with a mean change of minus nine point fifty-eight percent, a standard deviation of four point ninety-eight percent, and a matched-pairs t-test comparison below zero point zero zero one.

SRT fifty percent measures the minimum loudness needed to gain enough vocal information to decide which phrase was presented, whereas WRS at sixty-five decibels requires clearly understanding monosyllabic words without further clues. This may explain why WRS improved while SRT fifty percent did not.

Figure five compares word recognition scores before the pitch-matching session with scores two weeks afterward for patients one through twelve, plus an average. Testing presented speech to the implanted ear at sixty-five decibels and noise to the opposite ear at sixty decibels.

The chart shows patient-level variability, while the surrounding text notes that speech comprehension and perceived hearing quality improved for nearly all individuals, with heterogeneous results. The SSQ twelve questionnaire showed significant improvement in speech intelligibility, sound localization, and sound quality, with a mean increase of zero point ninety-six points and a standard deviation of zero point forty-five points.

Figures five and six display the respective results for the twelve patients: WRS at sixty-five decibels in Figure five and SSQ twelve in Figure six. Figure six compares SSQ twelve questionnaire scores before pitch matching with scores two weeks after fitting, for patients one through twelve and the average.

The paired bars show that hearing quality was assessed individually, while the error bars on the average summarize variability across patients. This matters because the authors report noticeable heterogeneity: even when improvement was observable in nearly all patients, some participants—including patients four and eleven—had already reported high SSQ twelve scores before fitting yet remained dissatisfied with the implant’s unnatural sound.

The results demonstrate that the proposed pitch-matching method may increase quality of hearing and speech comprehension in cochlear-implant patients with single-sided deafness. The method may also usher in positive results for bimodal patients, or for pitch-matching the two cochlear implants in bilateral patients.

The disadvantages include that the method is time-consuming and requires the patient to cooperate fully with the clinician at a more demanding level. Understanding and performing the task as a comparison of pitches, rather than loudness levels, is crucial to achieving beneficial results.

As an optional tool, the procedure can be implemented in cochlear-implant fitting software together with automated computations for reallocating the frequency bands. Subjective pitch matching produced a new frequency map whose hearing thresholds stayed essentially stable, while word recognition in noise and self-rated hearing quality improved significantly.

The method is promising, but it is time-consuming and requires careful patient cooperation.