Place · Level 3
Tinnitus · mostly not the ear ringing
响的落点在从耳蜗往上那条通路 · 耳朵那一侧常常只是输入变少 · 中枢把增益拧大, 本来听不见的底噪也被一起放大 · CBT 改变的是困扰不是声音 · 银杏锌褪黑素指南直接写了不推荐 · 突发听力下降要尽快就医
Story path
Chapter 1
Tinnitus · mostly not the ear ringing
Tinnitus · mostly not the ear ringing
Tinnitus sounds like the ear ringing, but treating it as a tiny speaker lodged in your ear gets the location wrong at step one. The evidence points the source further up — along the auditory pathway running from the cochlea inward: the cochlear nucleus in the brainstem, the inferior colliculus in the midbrain, all the way to the auditory areas of the cortex. What happens out at the ear is often just this: less input arriving.
Start with the input end. The cochlea is a coiled tube; uncoiled it runs about 33 to 35 mm, varying noticeably between people. Stretched inside it is the basilar membrane, narrow and stiff near the base (where sound enters) and progressively wider and more flexible toward the apex. Incoming sound travels along it as a wave running from base toward apex, with high frequencies peaking at the base and low frequencies at the apex. In other words, frequency is laid out as position inside the cochlea: whichever stretch moves most tells you which band of frequency came in. That place-to-frequency map is laid down by the gradient in the membrane's width and stiffness — but only in rough form. What actually sharpens the tuning is active amplification by the outer hair cells; passive mechanics alone cannot account for tuning this sharp at the periphery.
Here's the crux: hearing damage is rarely the whole ear declining together. Certain stretches go first. The classic shape of early or moderately advanced noise-induced hearing loss is a notch sitting between 3 and 6 kHz, usually centred near 4 kHz, with some recovery by 8 kHz. It counts as typical, and describes the early stage only — not as a notch means noise damage. The more accurate description: one stretch of the cochlea is now sending up a thinner signal, while its neighbours on either side carry on as usual.
What happens next is central, and the direction is counterintuitive. A review summarises the evidence this way: after cochlear output drops, activity in more central auditory structures is paradoxically enhanced at suprathreshold intensities — a phenomenon called central gain enhancement. Animal experiments show the individual parts moving: markers of inhibitory transmission (glycinergic and GABAergic) decline persistently, neurons' intrinsic excitability goes up (for example, reduced Kv7 potassium conductance in dorsal cochlear nucleus neurons), and spontaneous firing in the central auditory system increases.
In plain terms: when one band's signal weakens, the system turns up the gain knob on that band — the signal comes back a little, and the noise floor you were never meant to hear gets amplified along with it. This is the most-discussed explanatory route to the persistent tone. Three boundaries have to be stated alongside it. First, those electrical recordings were made directly in animals; this cannot be written as 'your brain is doing this right now'. Second, this is not a uniform whole-brain turn-up — the magnitude differs at each level of the pathway, so do the time courses, and multiple mechanisms are likely contributing at once. Third, the jump from gain being raised to therefore tinnitus is a step the review itself labels a hypothesis, not established causation.
So what makes anyone think the sound isn't generated in the ear? There's a direct test: if it truly came from the ear, cutting the auditory nerve should end it. The literature notes that surgically sectioning the auditory nerve does not eliminate tinnitus in every case, and that this is precisely what favours a central rather than peripheral origin. Note the wording — it fails to eliminate it consistently, which is not the same as cannot eliminate it. The first says the periphery isn't the whole answer; the second would be a claim the evidence can't carry.
The same review immediately walks half of it back: it is now well established that many forms of tinnitus reflect a complex interaction between peripheral and central mechanisms within the auditory pathway. So not the ear ringing does not mean nothing to do with the ear. The gap in that stretch of cochlea is often the trigger; the perception of sound is maintained centrally.
This mechanism carries a very practical implication. If what's ringing is the central response to missing input, then pouring something into the ear is aimed at the wrong target by construction; the interventions current evidence can support act on how the brain processes and relates to that sound. The only one that has ever reached moderate certainty is cognitive behavioural therapy — it may reduce the negative impact tinnitus has on quality of life, but the certainty runs only low to moderate, and what it changes is the distress, not the sound switching off. The pages that follow open up that effect size, and explain why the supplement route doesn't lead anywhere.
Start with the input end. The cochlea is a coiled tube; uncoiled it runs about 33 to 35 mm, varying noticeably between people. Stretched inside it is the basilar membrane, narrow and stiff near the base (where sound enters) and progressively wider and more flexible toward the apex. Incoming sound travels along it as a wave running from base toward apex, with high frequencies peaking at the base and low frequencies at the apex. In other words, frequency is laid out as position inside the cochlea: whichever stretch moves most tells you which band of frequency came in. That place-to-frequency map is laid down by the gradient in the membrane's width and stiffness — but only in rough form. What actually sharpens the tuning is active amplification by the outer hair cells; passive mechanics alone cannot account for tuning this sharp at the periphery.
Here's the crux: hearing damage is rarely the whole ear declining together. Certain stretches go first. The classic shape of early or moderately advanced noise-induced hearing loss is a notch sitting between 3 and 6 kHz, usually centred near 4 kHz, with some recovery by 8 kHz. It counts as typical, and describes the early stage only — not as a notch means noise damage. The more accurate description: one stretch of the cochlea is now sending up a thinner signal, while its neighbours on either side carry on as usual.
What happens next is central, and the direction is counterintuitive. A review summarises the evidence this way: after cochlear output drops, activity in more central auditory structures is paradoxically enhanced at suprathreshold intensities — a phenomenon called central gain enhancement. Animal experiments show the individual parts moving: markers of inhibitory transmission (glycinergic and GABAergic) decline persistently, neurons' intrinsic excitability goes up (for example, reduced Kv7 potassium conductance in dorsal cochlear nucleus neurons), and spontaneous firing in the central auditory system increases.
In plain terms: when one band's signal weakens, the system turns up the gain knob on that band — the signal comes back a little, and the noise floor you were never meant to hear gets amplified along with it. This is the most-discussed explanatory route to the persistent tone. Three boundaries have to be stated alongside it. First, those electrical recordings were made directly in animals; this cannot be written as 'your brain is doing this right now'. Second, this is not a uniform whole-brain turn-up — the magnitude differs at each level of the pathway, so do the time courses, and multiple mechanisms are likely contributing at once. Third, the jump from gain being raised to therefore tinnitus is a step the review itself labels a hypothesis, not established causation.
So what makes anyone think the sound isn't generated in the ear? There's a direct test: if it truly came from the ear, cutting the auditory nerve should end it. The literature notes that surgically sectioning the auditory nerve does not eliminate tinnitus in every case, and that this is precisely what favours a central rather than peripheral origin. Note the wording — it fails to eliminate it consistently, which is not the same as cannot eliminate it. The first says the periphery isn't the whole answer; the second would be a claim the evidence can't carry.
The same review immediately walks half of it back: it is now well established that many forms of tinnitus reflect a complex interaction between peripheral and central mechanisms within the auditory pathway. So not the ear ringing does not mean nothing to do with the ear. The gap in that stretch of cochlea is often the trigger; the perception of sound is maintained centrally.
This mechanism carries a very practical implication. If what's ringing is the central response to missing input, then pouring something into the ear is aimed at the wrong target by construction; the interventions current evidence can support act on how the brain processes and relates to that sound. The only one that has ever reached moderate certainty is cognitive behavioural therapy — it may reduce the negative impact tinnitus has on quality of life, but the certainty runs only low to moderate, and what it changes is the distress, not the sound switching off. The pages that follow open up that effect size, and explain why the supplement route doesn't lead anywhere.
Mechanism · where the gap comes from (1) · the parts that fail
If the main storyline is one stretch losing input, the next question is how that input goes missing. At least three places inside the cochlea can fail, and they fail in completely different ways.First, the amplifier. Outer hair cells contain a motor protein called prestin that changes the cell's length as membrane potential changes, actively amplifying vibration in the organ of Corti. Mice lacking prestin lose that electromotility, and hearing sensitivity drops by more than 40 dB. Forty decibels means the faintest audible sound pressure rises roughly 100-fold. Three things not to misread: this figure is from mammals (mice), not measured the same way in humans; 100-fold refers to pressure amplitude, not to sounding 100 times louder — perceived loudness is not linear in pressure; and those knockout mice also had outer hair cells only about 60% of normal length with reduced stiffness, so the 40 dB cannot be booked entirely to lost electromotility.
Second, the battery. The endolymph in the cochlea carries a potential of about +80 mV, and that potential is the voltage source driving hair-cell mechanoelectrical transduction. Its causal chain has three links: transporters in the stria vascularis spend energy maintaining the potassium gradient, the gradient produces a diffusion potential, and that potential drives transduction — the pumps do not pump out 80 mV directly. Which is why it depends so heavily on metabolism: anoxia, or blocking the sodium-potassium pump with ouabain, or blocking the sodium-potassium-chloride cotransporter with bumetanide, all make the potential fall. That +80 mV was measured in guinea pigs — it is not a number recorded inside a human ear.
Third, ageing doesn't take a single route. A classic temporal-bone histopathology study sorted age-related hearing loss into four types: sensory (organ of Corti and hair cells), neural (spiral ganglion cells — the cell bodies of the auditory nerve), metabolic or strial, and cochlear conductive. One honest distinction belongs here: the first three are types where the corresponding structural damage is visible under the microscope, whereas the fourth is defined by the shape of the audiogram — the presumed mechanism is altered cochlear mechanics, but no corresponding histologic change has been found. Also, pure types are the exception and mixtures are the rule: the study itself states that many individual cases don't separate into any one type but carry mixtures of these pathologies. So this is a framework for understanding, not a form to file yourself under.
Last question: why doesn't the gap fill itself in? Because adult mammalian cochlear hair cells essentially do not regenerate. There is a sliver of leeway in the newborn period — mice within the first postnatal week still harbour progenitor cells able to give rise to new hair cells — but that ability is progressively lost after birth, and in adult mammals no recovery has been observed following hair-cell loss. Note that this is not observed, not proven impossible; and the window data come entirely from mice, with no human counterpart.
One of the barriers is epigenetic: mouse supporting cells progressively accumulate DNA methylation at promoters of hair-cell genes, and that methylation overlaps with binding sites of Atoh1, a key transcription factor for hair-cell fate. Three qualifiers travel with it: everything is mouse, and the experiments were run on explanted newborn cochlear tissue; the paper's own wording is that methylation contributes to the failure, one layer among several barriers rather than a single switch; and therefore this must never be read as 'block methylation and the ear grows hair cells back' — current evidence cannot carry that sentence.
Put together: once the gap forms it largely persists — and that gain knob in the central pathway sits turned up, long-term, in front of a gap that will not close on its own.
liberman-2002-prestin-cochlear-amplifiernin-2008-endocochlear-potentialschuknecht-1993-presbycusis-typologyfranco-2017-cochlea-hair-cell-regenerationnguyen-2023-cochlea-methylation
Mechanism · where the gap comes from (2) · a normal audiogram is not a full input
Plenty of people with tinnitus get a hearing test and are told everything is normal, which pushes them toward 'it's psychological'. But the audiogram has two known blind spots, and they're worth opening up separately.Blind spot one: thresholds can recover while the wiring is already cut. A very clean experiment was run in mice: an 8-16 kHz octave-band noise at 100 dB SPL for two hours produced a threshold shift that was completely reversible, back to pre-exposure levels within two weeks. And no hair cells were lost, inner or outer, across follow-up of at least a year. It looks like nothing happened.
Yet in the region tuned to 32 kHz, the ribbon synapses between inner hair cells and the auditory nerve fell from about 16 per inner hair cell to fewer than 7 within 24 hours. Synapse counts had not recovered at 8 weeks, while the spiral ganglion cell bodies — the cell bodies of the auditory nerve — died much later: counts were near normal at two weeks, with significant loss appearing around one year and roughly a 50% reduction in that region by two years.
In plain terms: the threshold only reports whether the loudest wires still work. It cannot report how many wires have already been cut. And those cut wires are exactly what the central side experiences as an input gap.
Three limits hold here: this is mice, not people; the dose is laboratory-grade noise and must not be converted into everyday listening or headphone volume; and the direct synapse follow-up runs to 8 weeks without recovery, with anything longer resting on the inference from cell bodies dying off over the following years.
What about humans? An opinion paper written by nine specialists puts it bluntly: this synaptopathy has been demonstrated in numerous animal models, while in living humans it can only be confirmed post-mortem on temporal bones; a variety of non-invasive indirect measures have been tried, and the results are conflicting.
One detail is easy to get wrong and worth correcting: the same paper notes there is little debate about the existence of age-related synaptic loss in the human inner ear — it is visible in post-mortem temporal bones. So what remains unresolved is not whether the lesion exists in people, but two other things: whether noise is a cause of it in humans, and whether it can be measured in the living. That asymmetry — animals can be sectioned, living people can only be inferred about — is exactly where a lot of 'hidden hearing loss' marketing plants its flag, and it's more useful to remember than any single number. Note too that this paper is from 2019, so even the state of 'the results conflict' may have moved since.
Blind spot two: the conventional audiogram stops at 8 kHz. Among 116 adults with normal conventional audiograms (every threshold from 0.25 to 8 kHz at 20 dB HL or better), 74 (64%) had elevated thresholds above 8 kHz.
Read that number carefully. It comes from a recruited convenience sample, not a population prevalence; mean age was 29.5 years, range 18 to 65. The outcome measured was speech perception in noise: 39 participants (34%) reported difficulty, and the extended high-frequency threshold average identified them correctly about 81% of the time, versus about 71% for the conventional-frequency average. That is same-time statistical discrimination, not prospective prediction, and certainly not causation.
One more boundary, stated firmly: this study did not measure tinnitus at all. So the only claim it supports is this — a normal audiogram does not mean the cochlear input is intact — and it cannot be borrowed as an explanation for tinnitus.
The point of this page is a judgement tool: when someone says the test was normal so it must be in your head, you at least know how much ground the word normal covers on that sheet of paper, and what it leaves out. This is not a licence to dismiss your test results; it is a way to walk back into the clinic with sharper questions.
kujawa-2009-cochlear-synaptopathybramhall-2019-cochlear-synaptopathymotlagh-zadeh-2019-ehf-hearing
Chapter 2
What helps · how far the evidence goes
What helps · how far the evidence goes
The mechanism points centrally, so interventions should too. But aiming in the right direction and already working are two different things.
The only intervention that has reached moderate certainty is cognitive behavioural therapy (CBT): a Cochrane review of 28 randomised trials totalling 2,733 people concludes it may reduce the negative impact of tinnitus on quality of life. Two things travel with that: what it changes is the distress tinnitus causes, not the sound itself — and at the 6-month and 12-month follow-up points, the evidence is missing.
The other side has already been stated by a guideline. The 2014 American Academy of Otolaryngology-Head and Neck Surgery Foundation clinical practice guideline states that clinicians should not recommend ginkgo, melatonin, zinc or other dietary supplements for patients with persistent, bothersome tinnitus. That also follows mechanically: if what is ringing is central gain, then taking something to open up the blood vessels in the ear is aimed at the wrong target from the start.
The next two pages open each side up: first the effect sizes and certainty of brain-directed interventions, then how strong that guideline grade actually is and who it covers.
The only intervention that has reached moderate certainty is cognitive behavioural therapy (CBT): a Cochrane review of 28 randomised trials totalling 2,733 people concludes it may reduce the negative impact of tinnitus on quality of life. Two things travel with that: what it changes is the distress tinnitus causes, not the sound itself — and at the 6-month and 12-month follow-up points, the evidence is missing.
The other side has already been stated by a guideline. The 2014 American Academy of Otolaryngology-Head and Neck Surgery Foundation clinical practice guideline states that clinicians should not recommend ginkgo, melatonin, zinc or other dietary supplements for patients with persistent, bothersome tinnitus. That also follows mechanically: if what is ringing is central gain, then taking something to open up the blood vessels in the ear is aimed at the wrong target from the start.
The next two pages open each side up: first the effect sizes and certainty of brain-directed interventions, then how strong that guideline grade actually is and who it covers.
Clinical · interventions aimed at the brain, and how far the evidence actually goes
The mechanism points centrally, so intervention should point centrally too. But aiming the right way and already working are two different things, so this page lays the evidence out, including the unflattering half.Cognitive behavioural therapy (CBT). A Cochrane review including 28 randomized controlled trials with 2,733 participants concluded that CBT may reduce the negative impact tinnitus has on quality of life — while at 6 or 12 months of follow-up, the evidence is absent.
Certainty has to be read per comparator rather than collapsed into one phrase: versus no intervention, low certainty; versus audiological care, moderate certainty; versus other active treatments, low certainty.
The effect size is where honesty matters most, because the structure is awkward. The scale here is the Tinnitus Handicap Inventory (THI, 0 to 100, lower is better), and an improvement of about 7 points is generally taken as the smallest change a patient would notice. The one arm with moderate certainty (versus audiological care) came in at -5.65 points — short of that 7-point line — while the one arm that cleared it (versus no intervention, roughly -10.91 on the same scale) carries only low certainty. In one line: the more certain result isn't large enough, and the larger result isn't certain. This isn't talking it down — it is the only option on this page that has ever reached moderate certainty.
The safety side is clean: the review found CBT probably results in little or no difference in adverse effects, with a single participant reporting a single event across the seven studies that measured this.
One more thing that gets misread: these outcomes measure the negative impact of tinnitus on quality of life, not the sound going away. Within the central-gain frame that's actually coherent — if what rings is a gain setting that has been turned up, the first thing that can change is likely how the system relates to that signal, not the signal itself.
Does auditory training count? A systematic review assessed individual computer-based auditory training: 11 of 13 included articles showed learning generalising to untrained measures of speech intelligibility, but the improvements were small and not robust, published evidence ran from very low to moderate study quality, and the authors concluded it cannot reliably be used to guide intervention at this time. Two limits: the review covers individual self-administered computer-based training, not group aural rehabilitation or clinician-led work; and it was published in 2013.
And hearing aids? A hard line belongs here. Within the central-gain frame, restoring peripheral input and narrowing the gap should in principle weaken the very thing driving the gain up — a reasonable direction, but the current evidence does not measure that. A Cochrane review (5 randomized trials, 825 participants) assessed hearing aids for adults with mild to moderate hearing loss and found improvement across three outcomes — hearing-specific quality of life, listening ability, and overall health-related quality of life — on moderate-quality evidence. But those outcomes belong to hearing loss, not tinnitus, and this evidence cannot be read as hearing aids treating tinnitus. Note as well that the review graded the evidence on adverse effects as very low (measured by only one small study), so 'moderate quality' does not carry over to safety.
To close the page: 'the mechanism points centrally' is a directional judgement — useful for ruling out things that plainly miss the target, and unable to conjure a high-certainty therapy out of nothing. Admitting that is precisely where honest education parts ways with sales copy. What to actually do, and whether to do it, is something to decide with a clinician.
henshaw-2013-auditory-trainingferguson-2017-hearing-aids
Debunked · ginkgo, zinc, melatonin — and why the guideline says don't
If what rings is central gain, then 'take something to open up the blood vessels in the ear' misses the target by construction. And in this case both a guideline and a systematic review have already finished the sentence for us.The guideline side. The 2014 clinical practice guideline on tinnitus from the American Academy of Otolaryngology–Head and Neck Surgery Foundation states, in Statement 11, that clinicians should not recommend Ginkgo biloba, melatonin, zinc, or other dietary supplements for treating patients with persistent, bothersome tinnitus.
Three things about that statement, none of them skippable.
First, its grade is a recommendation (against) — not a strong recommendation. The same guideline reserves the strong tier for other statements, so writing this up as 'strongly advised' overstates the source.
Second, the guideline grades its own underlying evidence as quality C: the relevant randomized trials and systematic reviews show extreme heterogeneity, and most of the trials raise significant concerns about methodology and subject selection. So the correct reading is not proven effective, which is not the same as proven ineffective — scientifically these are different sentences, and both marketers and debunkers like to blur them.
Third, the panel disagreed internally, and said so in print: the majority felt harm clearly predominated over benefit, a minority felt they were in equilibrium, and none perceived a preponderance of benefit. Writing the uncertainty down is more credible than pretending to unanimity.
The scope matters too: this statement addresses adults, with primary persistent bothersome tinnitus, using supplements as a treatment for tinnitus. It does not say zinc is unimportant to normal cochlear physiology, and it does not say these substances fail in other indications.
Ginkgo later got a more detailed accounting. A 2022 Cochrane review located 12 studies with 1,915 participants in total. Now set that number down — the studies that could actually be pooled for tinnitus symptom severity were 2 studies, 85 people: at three to six months, Ginkgo biloba may have little to no effect compared with placebo (Tinnitus Handicap Inventory, 0-100, lower is better, mean difference -1.35, 95% CI -8.26 to 5.55), on very low-certainty evidence, and the authors concluded there is uncertainty about both benefits and harms.
Hidden in there is the exact trick this site most wants you to spot: using 1,915 to prop up a conclusion carried by 85. And the interval is wide enough to hold both possibilities: the lower bound of -8.26 means it cannot rule out a clinically meaningful benefit (it clears the 7-point line), and the upper bound of 5.55 means it cannot rule out a small harm either. So the honest verdict is we don't know: not shown to work, and not powered to show it doesn't. Anyone citing this review as proof that ginkgo is disproven is overreading it too, just in the opposite direction.
Harm isn't zero, and that part usually gets skipped. The guideline spells out the mechanism: ginkgo's flavonoids and terpenoids have antiplatelet effects and inhibit clotting when combined with anticoagulants; reported events include bleeding, hematoma, apraxia, permanent neurologic deficits, and death; it can also interact with thiazide diuretics to raise blood pressure. On that basis the guideline flags older adults, people prone to bleeding, and anyone taking anticoagulants as those who should particularly avoid it.
So the real ledger isn't 'useless but harmless'. It's: something not proven to work, carrying a specific documented bleeding risk, while occupying the time and confidence that could have gone toward a direction with evidence behind it.
One closing note: this guideline was published in 2014, so cite it as 2014 rather than blurring it into 'the latest guideline'. Whether to take any supplement — especially if you are already on medication — is a judgement for a clinician who knows your situation.
sereda-2022-cochrane-tinnitus
Chapter 3
Red flags · when to see a clinician
Red flags · when to see a clinician
Tinnitus itself is usually not an emergency. But it can appear alongside something that does need prompt medical attention: sudden sensorineural hearing loss.
The definition given is a hearing drop of 30 dB or more across at least 3 consecutive frequencies, arising within 72 hours. One point is very easy to read backwards — the 72 hours describes the window in which the loss occurs, not 72 hours you have in which to seek treatment.
The next page sets out how to read each of those numbers correctly, along with the other signals that mean stopping and seeing a clinician.
The definition given is a hearing drop of 30 dB or more across at least 3 consecutive frequencies, arising within 72 hours. One point is very easy to read backwards — the 72 hours describes the window in which the loss occurs, not 72 hours you have in which to seek treatment.
The next page sets out how to read each of those numbers correctly, along with the other signals that mean stopping and seeing a clinician.
Red flag · when to stop reading and see a clinician
Tinnitus by itself is usually not an emergency. But it can appear alongside something that does warrant prompt medical attention: sudden sensorineural hearing loss.The definition given in the literature is a rapid-onset decrease in hearing occurring over a 72-hour period, of 30 decibels or more, affecting at least three consecutive frequencies. One thing is very easy to read backwards, so say it outright: the 72 hours is the window in which the loss develops, not 72 hours you have in which to get treated.
Those three numbers are easy to misuse in other ways too, so three readings come with them.
First, the 30 dB across 3 frequencies is a research and trial-entry convention, not a clinical gate. The guideline itself states that in clinical practice, expanding the definition to cases with less than 30 decibels of loss may be considered. Meaning a drop short of 30 dB can still be this — never use the number as a reason not to get seen.
Second, this is a number from an audiogram; you cannot measure it yourself. And because most people have no pre-illness audiogram, the practical comparison is against the thresholds of the opposite ear. The accurate phrasing is that a clinician confirms it with an audiogram — not that you estimate how many decibels you've lost.
Third, the urgency has two layers. One is that it may require ruling out vestibular schwannoma (acoustic neuroma), stroke, and malignancy. The other is that, as the literature notes, hearing recovery may be better when it is recognised and managed promptly. The actionable time anchors are: an audiogram within 14 days of symptom onset, and corticosteroids may be offered within 2 weeks.
So in the context of this scene, the dividing line runs like this: a long-standing, bilateral tinnitus with no other symptoms is the central-gain story told above; a sudden, mainly one-sided tinnitus accompanied by clearly worse hearing on that side is the kind of thing to have a clinician look at promptly, not to watch for another two weeks at home.
One last thing said plainly: this scene explains a mechanism. It is not a diagnosis and it does not replace a clinician. Its real uses are two. When someone tells you to take ginkgo to open up the blood vessels in your ear, you know at which step that sentence stops matching the mechanism. And when one ear suddenly muffles while ringing, you know that is not a signal to wait.
Understanding the mechanism will not make tinnitus stop. But it lets you stay calm — because you know where the sound comes from; and it lets you stay hard to sell to — because you know which promises never landed on the target in the first place.