keywords: anton confabulation hallucination
I have had tinnitus for fifteen years. I have been told it is either a hearing problem or something wired badly in my brain. Neither explanation has ever satisfied me, and neither has offered anything useful. The sound changes: sometimes a square wave jumping between frequencies, sometimes a low pulsating throb, sometimes nothing at all. Some days I can stop it for thirty minutes by stretching my jaw wide open, after which it returns exactly as it was. Nobody has explained why.
In 2022, during Melbourne’s pandemic lockdown, I had a splitting headache as I went to sleep. When I woke, l looked at my phone, and I got angry. My phone was broken – I thought… The right side of the screen was blank. Great… I have a splitting headache and now my phone is broken. I looked up in the bathroom mirror and then it hit me. Half of my face was gone. Not gone, just bits of it were missing – on the right side. My head still looked normal but I only had one eye. I looked down at my hand and I only had three fingers… I shut my eyes and felt my hand: all five fingers were there and I touched my face and sure enough – I felt two eyes. I went to find my wife.
She also had only one eye.
She drove me to hospital emergency. Due to pandemic protocols she was not permitted to enter with me. I went in alone – was this madness or had I gone blind?
They handed me forms to fill out. I could not read them. Not because of my headache, but because the right half of every page was white. I assumed the forms were poorly printed. I said nothing about it, filled in what I could see, and waited.
They diagnosed a severe migraine and sent me home. It was not a migraine. I had had a stroke.
It was in fact two strokes, one in the left occipital lobe and one in the cerebellum, neither of which was detected until weeks later when, as I began to recover and the confabulation started to resolve, an MRI at a different hospital revealed the lesions, and sent me straight back to hospital to the stroke ward. By that point I had been living with the damage for weeks, navigating a world that my nervous system was reconstructing from the inside, plausibly enough that neither I nor the emergency physicians had recognised it as neurological.
That was for me, the birth of the Myelin Mind project.
This is the first thing the Myelin Mind thesis predicts and the standard account cannot explain. The confabulation was not noise. It was not obviously wrong. It was a coherent, plausible version of reality, generated by intact white matter drawing on a lifetime of accumulated visual experience, convincing enough to be mistaken for the world itself. A broken wire produces static. My nervous system produced a tree that looked exactly like a tree, a face that looked exactly like a face, a hand that looked exactly like a hand, but with three fingers.
What I could and could not see
The confabulation was not random. It was selective in a way that, once you have the framework, is precisely predictable.
Natural spaces were reconstructed faithfully. Trees looked exactly as trees should look. Landscapes were intact. The physical world, the environment that every vertebrate navigates, was present and coherent on both sides of my visual field, even on the right side that my damaged occipital cortex could no longer directly serve.
Faces were reconstructed with extraordinary fidelity. When I looked at a person I could see their full face. Except for the right eye. The right eye falls in the right visual field, the territory the lesion had interrupted, and the confabulation could not reach the myelinated representation of that specific location. So it reproduced the face, faithfully, minus the element it could not access. Not a distorted face. A face with an absence. The category was intact. The element was missing.
Cars were reconstructed but imperfectly. They had wheels only on the left side. They floated. My myelinated visual history of cars is decades deep and the confabulation had material to work with, enough to produce something recognisable, but not enough to complete the object beyond what was directly accessible. A car, approximately. Levitating on two wheels, and not seeming strange.
When I looked at my hand it had three fingers. 
The hand is one of the most heavily myelinated perceptual structures in the human nervous system, the cortical map disproportionately large, the accumulated experience of a lifetime of manual activity deeply inscribed in white matter. Yet the confabulation produced three fingers rather than five. The myelinated structure that constitutes hand was intact and the result did not look grotesque. It looked like a hand. The number was wrong. The category was right.
People standing to the right of objects were invisible. A person standing next to a tree on my right side simply was not there. The tree was reconstructed faithfully. The person, lacking the object-anchored context that might have triggered reconstruction, fell into the gap.
Text was gone. Not distorted, not replaced by something else. Effaced. Where text should have been, there was white. This is why I could not read the hospital forms. This is why I said nothing about it, because the experience of white where text should be did not present itself as absence. It presented itself as blankness, as unimportance, as something to move past. The confabulation did not signal its own failure. It simply did not reconstruct what it had no intact myelinated structure to draw on.
And then there was the Korean.
I could see Korean text. I cannot read Korean but I could see it, because my myelinated visual structure had no accumulated experience of Korean as meaningful. It was pattern without significance, visual texture without semantic weight, and the confabulation reproduced it faithfully because it required no myelinated meaning to reconstruct, only visual form. When my son explained what a piece of Korean text meant, it acquired significance. It entered the domain of language, of meaningful inscription.
and it promptly vanished.
The myelinated structures that would have processed it as meaningful were damaged, and once the text was meaningful, it fell into the gap.
This is not a curiosity. It is a proof. The confabulation was selective precisely according to the boundary between myelinated meaning and unmyelinated pattern. The white matter reconstructs what it knows. It cannot reconstruct what it does not know, and it cannot reconstruct through damaged territory.
The evolutionary ordering of myelination
The confabulation revealed something about the developmental and evolutionary sequence of myelination that I had not anticipated.
The structures reconstructed most faithfully were the oldest, most deeply inscribed, most evolutionarily ancient: natural spaces, the physical environment every vertebrate navigates; faces, the social perceptual capacity that predates language by millions of years; bodies and limbs, the body schema inscribed from the first weeks of life. The structures that were effaced were the most recently acquired: text, abstract symbols, the cultural overlays of a literate modern life.
The confabulation was showing me my own myelination history. It was reconstructing the world in order of the depth of inscription, the oldest layers most faithfully, the newest layers not at all. What I lost was not vision. What I lost was the most recently myelinated vision, the cultural and linguistic layers built on top of the ancient visual foundation. The ancient foundation remained. It filled in what it could. It did so convincingly enough that a hospital emergency department, under pandemic conditions, sent me home.
Two phenomena, one principle
I have been describing my stroke confabulation. But I began with my tinnitus. These are not the same condition. The tinnitus has been present for fifteen years, predating the strokes by eleven years. They arrived separately, through different mechanisms, in different parts of the nervous system. And yet the Myelin Mind thesis accounts for both through the same principle.
Consciousness arises at the chiasm, the encounter between neural flow and myelinated structure. The myelinated white matter is the accumulated biological condition of lived experience. When a neural signal flows through a myelinated pathway it encounters this accumulated structure, and the encounter produces experience. The signal provides the trigger. The white matter provides the content.
What happens when a signal is misrouted? When, through stroke or developmental variation or the accumulated stress of an injured auditory pathway, a neural signal flows into myelinated territory it was not supposed to reach?
The myelinated structure does not know the signal is misrouted. It simply does what it always does: it generates experience consistent with its accumulated history. The signal is wrong. The structure it encounters is real and intact. The experience is therefore coherent, plausible and stable, because it is being generated by genuine white matter. Just not the white matter the signal was meant for.
My confabulation was visual signals, or the absence of visual signals, flowing into intact myelinated visual structure and producing experience consistent with what that structure had accumulated over a lifetime of seeing. My tinnitus is a misrouted signal, most likely originating in the trigeminal or auditory brainstem pathways, flowing into intact myelinated auditory territory and producing exactly the kind of experience that territory is structured to produce.
This is why the tinnitus varies. A broken wire produces a fixed signal. The variability of my tinnitus, the square wave jumping between frequencies, the pulsating throb, the silent days, is the signature of a stable misrouting encountering variable white matter conditions. The myelinated auditory structure the signal flows into is alive and changing, its condition shifting with fatigue, stress, blood flow and inflammation. As the condition of the white matter changes, the experience it generates from the misrouted signal changes with it. The tinnitus has been present for fifteen years not because the signal is fixed but because the myelinated structure it flows into has been stable. On the days it stops, something in that structure has temporarily changed.
The jaw stretch confirms this. The trigeminal nerve, which serves the jaw and the temporomandibular joint, has direct connections to the cochlear nucleus, the first relay station of the auditory pathway. Stretching the jaw modulates trigeminal activity, which temporarily alters the acoustic conditions of the brainstem auditory structures, which changes the environment in which the misrouted signal encounters the myelinated auditory white matter. For thirty minutes, the encounter is disrupted. The tinnitus stops. When the trigeminal effect fades, the misrouting resumes on its accustomed terms, and the familiar sound returns.
The jaw stretch is not fixing the signal. It is temporarily changing the white matter conditions the signal is flowing into.
Confabulation: Synesthesia and hallucination
The same principle extends outward. The synesthete who consistently sees a specific colour when she hears a specific word is not cross-wired in the sense of a manufacturing defect. Her auditory signal is flowing into myelinated visual territory, and that visual territory has an accumulated history, a lifetime of colour experience inscribed in white matter, and it produces experience consistent with that history. The blue she sees when she hears the word phenomenology is not invented. It is generated by real visual white matter, encountering a real signal, producing real experience. The signal is auditory. The territory is visual. The experience is the territory speaking.
This is why synesthetic experience is consistent across years and decades. The same word always produces the same colour because the myelinated visual territory the auditory signal flows into does not change. The white matter is stable. The experience it generates is stable.
Hallucination follows the same logic extended. In waking hallucination, whether induced by psychedelics, fever, extreme fatigue or psychosis, misrouted or internally generated signals flow through intact myelinated territory and produce coherent experience that has no correspondence to the external world but has full correspondence to the accumulated structure of white matter. The hallucination is not noise. It is the white matter speaking without external input to correct it.
In all these cases, confabulation, tinnitus, synesthesia, hallucination, the myelinated structure is the dominant partner. Whatever signal arrives, the white matter will make something coherent from it. It cannot do otherwise. It is the accumulated condition of a lived life, and it speaks in the only language it knows.
A fully awake dream
The sleep article in this series described dreaming as the experiential residue of myelination editing: the white matter reencountering its own accumulated structure in the absence of external input, producing coherent experience from the inside out. What I experienced after my strokes, and what I experience every day with my tinnitus, is the same phenomenon in waking life. A misrouted signal flowing into intact white matter, producing coherent experience that corresponds not to the external world but to the accumulated myelinated history of the nervous system.
A fully awake dream. Plausible, coherent, stable, and generated by the same biological machinery that produces all conscious experience. The signal is wrong. The white matter is right. The experience is real.
I went into a hospital emergency department during a pandemic lockdown, effectively blind on one side, unable to read the forms I was handed, and I was sent home with a migraine diagnosis. The confabulation was so coherent that it did not announce itself as damage. It presented as the world, slightly off, in ways I could not yet name.
The wired mind has no explanation for any of this. The Myelin Mind has nothing else.
Jack Parry is a philosopher, polyglot and biomedical animator at Swinburne University of Technology. He is the author of The Myelin Mind: The Genesis of Meaning.