Central gain is the brain's hearing pathway turning up its own sensitivity after the ear sends it less signal, and it is a leading explanation for how hearing loss can lead to both tinnitus and hyperacusis.
A 2014 review sets out the core observation. After noise or an ototoxic drug damages the cochlea, the signal leaving the ear is weaker, yet activity in more central hearing structures is paradoxically enhanced in response to sounds above threshold. That compensatory increase is called central gain enhancement, and the review proposes it as a possible mechanism for hyperacusis and tinnitus. It also concludes that several mechanisms, working over different timescales and frequency ranges, probably contribute.
The idea is well supported in animals and partly supported in people. What it has not yet done is explain why two people with the same hearing loss can end up one with tinnitus and one without.
How the brain turns up the volume
Neurons tend to hold their average firing rate steady. When input falls, they compensate by becoming more responsive. A 2006 computational model showed that this homeostatic compensation can produce hyperactivity after outer hair cell loss, and can amplify inputs from outside the hearing system too.
A 2011 review by Noreña describes the trade-off. Central gain adapts sensitivity to the weaker input, preserving stable firing and efficient coding. Maintaining that balance at all costs can mean amplifying "neural noise", the random background activity in the pathway, and that amplified noise may be heard as tinnitus.
A 2015 review from Manchester spells out how one mechanism could produce two symptoms. Abnormally high gain may raise spontaneous firing, which could be heard as tinnitus, and raise the response to real sounds, which could be felt as sound intolerance. Hyperacusis, a reduced tolerance for ordinary sound, is the second of those.
Gain is not only a fault. In a 2016 mouse study, destroying more than 95% of the synapses between hair cells and the auditory nerve virtually eliminated the auditory brainstem response and the startle reflex, yet the mice could still detect tones almost normally. Sound-evoked activity in the cortex rebounded to control levels or beyond. The restored hearing covered simple features coded by firing rate, but not those needing precise timing, such as speech.
What the animal studies show
Animal work tests the model most directly, with one constant limitation: animals cannot report tinnitus, so it is inferred from behaviour, as how tinnitus is generated explains.
- Loudness tracks gain. In a 2019 rodent study, high doses of sodium salicylate caused temporary hearing loss and signs of hyperacusis. Changes in how quickly loudness grew were correlated with changes in sound-evoked brain activity within individual animals, which the authors called strong support for the central gain model of hyperacusis.
- But not always enough. In rats exposed to noise for 12 weeks, responses in the auditory cortex and amygdala were proportionately much larger than the cochlea's reduced output, yet still below the level needed to fully account for the hyperacusis the rats showed.
- Gain is not the dividing line. A 2019 review in Neuron reports that animals with equivalent cochlear damage appear both with and without behavioural signs of tinnitus. Only those that develop tinnitus show synchronised spontaneous firing in brainstem neurons fed by the auditory nerve.
What the human studies show
People cannot be given cochlear damage for research, so human studies use temporary deprivation or compare groups carefully matched for hearing.
Earplugs. In a 2012 study, 18 volunteers with normal hearing wore a silicone earplug in one ear for 7 days, simulating a mild high-frequency hearing loss. 14 of them reported phantom sounds. In the 11 who had a stable sound on day 7, it was mostly high-pitched, in the range the plug blocked most. In every case it disappeared when the plug came out.
Follow-up studies measured the gain itself. After 7 days of plugging, the acoustic reflex, a middle-ear muscle response controlled by the brainstem, was triggered at a lower sound level in the plugged ear, and sounds were rated louder in both ears. Most changes had gone within 24 hours of removal. The authors took the mismatch between the two measures as a sign of more than one gain mechanism. A 4-day version found the reflex back to baseline within 2 hours, but could not reproduce the loudness change, and one brainstem wave moved in the opposite direction to the one predicted.
Silence. The same principle may explain phantom sounds in silence. In a 2008 study, 53 normal-hearing young adults spent 4 minutes in an anechoic chamber, and 83% reported hearing at least one sound.
Loudness in tinnitus ears. A 2013 study compared loudness judgements in 124 ears with tinnitus and 106 ears without, matched for hearing loss. Sensitivity to sound was enhanced in the tinnitus ears, including those with normal audiograms, a result the authors read as compatible with maladaptive central gain.
Brain imaging. A 2010 functional MRI study of people with clinically normal hearing thresholds found that those with reduced sound tolerance showed elevated sound-evoked activity in the auditory midbrain, thalamus and primary auditory cortex. Tinnitus was linked to elevated activity in primary auditory cortex only, which the authors suggested may reflect attention drawn to hearing.
Hidden damage. A 2011 study found a smaller auditory nerve wave but a normal brainstem wave in people with tinnitus and normal audiograms, which the authors read as the brainstem restoring a weakened input. A 2017 study of young adults matched for hearing up to 14 kHz found no reduction in that nerve wave. Whether this hidden hearing loss is common in people is still disputed.
Why the same hearing loss does not always cause tinnitus
This is the model's main gap, and it agrees with what tinnitus and hearing loss describes. A 2019 review titled "Tinnitus: Does Gain Explain?" notes that dramatic gain changes follow damage to the ear whether or not tinnitus occurs. It finds compelling evidence that ear damage raises gain, but says attributing that to tinnitus is usually hampered by the lack of hearing-matched human controls or damaged animals without tinnitus. The few attempts to classify individuals by gain have not succeeded.
The clearest human test came from earplugs. In a 2019 study, 44 normal-hearing adults wore an earplug in one ear for 4 or 7 days, and 30 reported tinnitus by the end. The acoustic reflex threshold in the plugged ear fell by 5.9 dB in those with tinnitus and 6.3 dB in those without, essentially the same. The authors concluded that brainstem gain may be necessary but not sufficient, with further changes higher in the pathway needed.
Candidates for that missing ingredient include synchronised firing, better-preserved inhibitory circuits in older people with hearing loss who have no tinnitus, and, from the 2019 gain review, focused attention, failed sensory gating and altered sensory predictions.
Tinnitus and hyperacusis: one mechanism or two?
The two often occur together, which is part of why a shared mechanism was proposed. A 2015 modelling study notes that about 85% of hyperacusis patients also have tinnitus. But its model fitted hyperacusis best with gain applied to sound-driven activity after spontaneous activity was subtracted, in contrast to tinnitus theories built on amplified spontaneous firing.
A study from the same group measured audiograms and loudness discomfort levels in 381 patients whose main complaint was hyperacusis. Hearing was on average mildly reduced at high frequencies, but more than a third of the ears tested had normal thresholds. Discomfort levels averaged around 85 dB HL and were lowered across the whole frequency range regardless of the pattern of hearing loss, with a moderate correlation of 0.36 between thresholds and discomfort levels. The authors suggested a generalised increase in gain in hyperacusis, unlike the frequency-restricted changes thought to cause tinnitus.
A 2020 model by Zeng puts the split in similar terms: added central noise is likely to generate tinnitus, and multiplied central gain is likely to generate hyperacusis. A 2022 review adds that hyperacusis is linked to more than twenty non-auditory disorders, including migraine, autism and head trauma, and that amplification may extend into brain regions for emotion, memory and stress. Misophonia versus hyperacusis covers how the two differ.
What it means for treatment
If lost input turns gain up, restoring input might turn it back down. The 2006 model predicted that added sound could reverse the hyperactivity. In a 2006 cat study, animals kept after noise trauma in quiet, or in sound enriched only in low frequencies, developed increased spontaneous firing and synchrony. Those kept in sound enriched in high frequencies did not, which the authors read as the absence of putative neural signs of tinnitus. That was sound given straight after the injury, in animals.
In people, the strongest controlled result is for loudness tolerance. A 2015 randomised placebo-controlled trial enrolled 36 people with sensorineural hearing loss whose reduced sound tolerance made hearing aids hard to use, 9 per group. They were not recruited for tinnitus. Counselling plus sound generators raised the level judged uncomfortably loud by about 12 dB. By the trial's criterion of at least 10 dB improvement, 82% responded with full treatment, against 25% with counselling and a placebo device, 40% with sound generators alone, and 50% with the placebo device alone.
A 2022 tutorial concluded that the evidence for sound therapy reducing gain is stronger for hyperacusis than for tinnitus. A 2020 review agreed that basic science supports the model but that clinical evidence lacks controlled trials.
For tinnitus, the trial evidence is thin. A 2018 Cochrane review of 8 studies with 590 participants found no evidence that hearing aids, sound generators or combination devices beat a waiting list, placebo or information alone, because none of the included studies provided data for those comparisons; the device-against-device evidence it could assess was low quality. A 2014 Cochrane review of hearing aids found one trial and no evidence to support or refute their routine use. A 2026 umbrella review of 44 systematic reviews did find that hearing aids and sound or music therapy consistently improved tinnitus outcomes.
In practice, the model is consistent with the steps usually offered: have a hearing test, consider amplification where there is hearing loss, and avoid long stretches of silence, using background sound such as the sound library if it helps. None of the trials above shows a device resetting central gain in people with tinnitus. Sound therapy and the treatment comparison set out the options, and how tinnitus is diagnosed covers the hearing tests. Earplug overuse covers when protection helps and when it may backfire.
When hearing loss needs checking now
A 2015 review describes gain changes after deprivation as slow adaptations, but some hearing changes are urgent. NICE says people with tinnitus and hearing loss that developed over 3 days or less in the past 30 days should be seen within 24 hours; sudden hearing loss explains why. The when to seek care checklist sorts the other signs that need prompt or emergency attention.