A Researcher Who Took Tinnitus’s Silence Seriously
If you have just been told you have tinnitus — and that there is no cure — the search for progress can feel urgent and disorienting. What you may not know is that a small number of researchers are asking genuinely new questions about the condition. Linus Milinski, a neuroscientist at Oxford University’s Sleep and Circadian Neuroscience Institute (SCNi), is one of them. His work centres on a deceptively simple observation: tinnitus and sleep both depend on the same type of spontaneous brain activity. That overlap, he argues, is not coincidental. This article walks through what his research actually found, what it cannot yet tell us, and why it matters for someone living with tinnitus today.
What Linus Milinski’s Tinnitus Research Found, in Plain Language
Oxford neuroscientist Linus Milinski has shown, in a 2024 ferret study published in PLoS ONE, that neural markers of tinnitus are suppressed during deep NREM sleep — and that animals with more severe tinnitus also develop disrupted sleep, suggesting a bidirectional cycle where tinnitus worsens sleep and poor sleep prevents the brain from naturally dampening tinnitus activity (Milinski et al. (2024)).
This finding builds on a 2022 theoretical review in Brain Communications, co-authored with sleep neuroscientist Vladyslav Vyazovskiy and auditory neuroscientist Victoria Bajo Lorenzana, which was described by its authors as “the first to consider, at a functional level, how sleep might impact tinnitus, and vice versa” (Milinski et al. (2022)). Together, the two papers represent a sequenced research programme: first a theoretical framework, then an empirical test in animals. The evidence is preclinical and early-stage. It does not represent a treatment advance yet — but it is peer-reviewed science pointing in a direction no one had formally mapped before.
The 2022 Review: Building the Theoretical Framework
Before Milinski’s 2022 review, tinnitus research and sleep research occupied almost entirely separate territory. Clinicians knew that many patients with tinnitus slept badly; researchers studying sleep had no particular reason to look at phantom sounds. Milinski, Vyazovskiy, Bajo Lorenzana, and colleague Johannes Kohl changed that by asking whether the two fields might share a common biological mechanism.
The review they published in Brain Communications proposed a specific theoretical framework. Tinnitus arises when the brain’s auditory regions become persistently overactive — an attempt to compensate for reduced input from damaged hearing cells. This hyperactivity does not switch off easily. During deep NREM (non-rapid eye movement) sleep, however, slow oscillatory waves in the range of 0.5–4 Hz sweep rhythmically through the cortex. These waves produce what neuroscientists call “down states” — brief windows in which neuronal firing quietens almost completely. The theoretical proposal was that these down states might temporarily reset or suppress the phantom signal that tinnitus generates.
The review also raised a second possibility: that sleep deprivation, by reducing the frequency and depth of these slow-wave cycles, could keep the auditory system in a heightened state, making tinnitus more intrusive. And it went one step further, suggesting that sleep-based neural plasticity may contribute to the consolidation of tinnitus in its early stages — which implies a possible early therapeutic window before the brain’s response to tinnitus becomes entrenched.
This was theoretical, not clinical. The paper presented no original experimental data. Its value was in formally connecting two fields, proposing a mechanism testable by experiment, and laying groundwork that could justify empirical work — which is precisely what came next.
The 2024 Ferret Study: Putting the Theory to the Test
For the 2024 study, Milinski’s team needed an animal with an auditory brain structure that resembles a human’s more closely than a mouse or rat does. Ferrets fit that requirement: their auditory cortex has greater structural similarity to the human auditory cortex than rodents, making them a more credible model for translating findings toward human biology.
Eight adult ferrets received mild noise exposure designed to induce tinnitus. Tinnitus was assessed using gap-detection tests: an animal with tinnitus perceives a phantom sound that “fills in” a brief silent gap in a background noise, raising its detection thresholds in ways that can be measured behaviourally. Auditory brainstem responses tracked hearing integrity separately. Three of the eight ferrets received chronic EEG and EMG implants, allowing the team to monitor sleep architecture continuously over weeks and months after noise exposure.
The findings confirmed the core prediction. Animals showing signs of tinnitus consistently developed sleep impairments: more frequent wake episodes, fragmented NREM, and reduced time in deep slow-wave sleep. Neural markers of tinnitus — elevated cortical firing rates and heightened evoked responses — were measurably reduced during NREM sleep. As the paper puts it, “sleep may transiently mitigate tinnitus” (Milinski et al. (2024)).
A clinically significant detail, reported in the full paper body, is the dissociation between tinnitus severity and hearing loss. The paper notes that “behavioural and evoked activity measurements suggested distinct degrees of tinnitus and hearing impairment between individuals” (Milinski et al. (2024)). According to reporting on the full paper, it was tinnitus severity — not hearing loss per se — that predicted sleep disruption. This matters clinically: it means the sleep effects are being driven by the phantom perception itself, not simply by the degree of auditory damage. Note that the publicly available abstract uses the phrase “hearing loss and/or tinnitus” — the dissociation to tinnitus severity specifically is reported in the full paper.
The limitations here deserve equal weight. The total sample was eight ferrets, with only three receiving the chronic implants needed to measure sleep architecture properly. Individual variability between animals was pronounced. The study used a cross-case design, not a controlled group comparison. And because ferrets cannot describe a subjective experience, tinnitus can only be inferred from behaviour — not measured directly. These are real constraints on what can be concluded, and they make replication in humans the obvious next step.
Milinski described the core finding plainly: “We could actually see these sleep problems appear at the same time as tinnitus after noise exposure. This suggested, for the first time, a clear link between developing tinnitus and disrupted sleep” (Cassella (2025)).
The Bidirectional Cycle: What It Means for You
The picture emerging from Milinski’s work is not simply that tinnitus makes sleep harder (though it does). The more significant implication is that the relationship runs in both directions.
Tinnitus creates persistent hyperactivity in auditory brain regions. That hyperactivity resists the neural quietening required to enter and maintain deep NREM sleep, keeping the brain in lighter, more fragmented sleep stages. When deep sleep is reduced, the slow oscillatory cycles that would naturally dampen the tinnitus signal occur less often or less deeply. The next day, the phantom sound is more intrusive. Reduced sleep also raises stress levels, and, as Milinski has noted, “when we do not sleep well, we become more vulnerable to stress, and stress is one of the strongest factors known to worsen tinnitus. Stress can even trigger tinnitus to begin with” (Cassella (2025)).
His summary of this cycle is direct: “Tinnitus can make sleep worse, and poor sleep may, in turn, make tinnitus worse. It may be a kind of vicious circle, although I do not believe it is unbreakable” (Cassella (2025)).
A 2025 human EEG study from South China University of Technology provides early corroboration of this picture in people. Bao and colleagues studied 51 tinnitus patients against 51 age-, sex-, and hearing-matched controls using full overnight polysomnography. The tinnitus group showed persistent daytime hyperarousal — enhanced high-frequency gamma brain activity across all waking conditions — and struggled to suppress that hyperactivity as they transitioned into sleep. Across all sleep stages, the tinnitus group showed elevated gamma and beta power compared to controls, and reduced slow-wave activity in the deeper sleep stages. The authors described this as a “neuroplastic overgeneralization of wake hyperarousal into sleep” (Bao et al. (2025)). Critically, deep sleep did eventually reduce the hyperactivity — consistent with what Milinski observed in ferrets. The authors concluded that “sleep is a critical therapeutic target to interrupt the 24-hour dysfunctional cycle of tinnitus” (Bao et al. (2025)).
The therapeutic implication — that improving sleep quality might reduce tinnitus severity by restoring the brain’s natural suppression mechanism — has not yet been tested in a clinical trial. This is a hypothesis with a credible mechanistic basis, not a confirmed intervention.
What This Research Does — and Doesn’t — Tell Us
A clear-eyed reading of where the evidence stands is worth more than optimism that outruns the science.
The limitations are real. The 2024 ferret study rested on three chronically implanted animals. Individual variability was large enough that the authors used a cross-case design rather than group statistics. Ferrets are a better auditory model than rodents, but they are not humans, and the subjective experience of tinnitus cannot be measured in an animal — only inferred. No clinical trial has yet tested whether sleep enhancement reduces tinnitus severity in people. The causal direction is not fully resolved: the 2024 data suggest that tinnitus causes sleep disruption, and that poor sleep worsens tinnitus, but confirming the full bidirectional loop in humans requires larger prospective studies. The US VA/DoD clinical practice guideline, updated in 2025, does not include any sleep-targeted treatment for tinnitus — reflecting the evidence as it stands today (Sherlock (2025)).
What the research has achieved is different in kind from a clinical advance, but genuinely meaningful. For the first time, a mechanistic connection between tinnitus and sleep has been proposed at the theoretical level, tested empirically in a relevant animal model, and partially corroborated in a controlled human EEG study. The field has moved from “patients report sleeping badly” to “here is a specific neural mechanism that may explain why, and which direction the causality runs.” That is a different kind of progress.
Milinski’s current research direction extends this: his team is now investigating whether sleep may influence the very development of tinnitus in its earliest stages. If so, there may be a window early after onset when sleep-based interventions could reduce the likelihood that tinnitus becomes chronic — before the brain’s compensatory responses become entrenched. This remains a hypothesis, but it is one now being actively pursued with peer-reviewed funding from RNID and the Wellcome Trust.
The practical takeaway is not to wait for a clinical trial. Prioritising sleep quality — through established approaches like consistent sleep timing, reducing stimulants before bed, and addressing anxiety — targets a mechanism that the evidence now describes. It is good self-care that also happens to align with an active research hypothesis.
Key Takeaways
Linus Milinski’s Oxford research established that tinnitus and sleep share the same spontaneous brain activity circuits, and that deep NREM sleep may temporarily suppress the neural signals driving the phantom sound. The ferret study, though small, was the first to test this empirically — and a 2025 human study from South China University of Technology (Bao et al. (2025)) supports the finding in people.
The evidence is preclinical and early-stage. It is not a cure, and no clinical trial has confirmed that improving sleep reduces tinnitus severity. But it is credible, peer-reviewed science pointing toward a real mechanism.
For someone with tinnitus today: prioritising sleep quality is not just symptom management. It may address a pathway that research is actively mapping. The absence of a cure does not mean the absence of progress — and Milinski’s work shows what genuine progress in this field looks like.
