The Filter, Not the Lamp

What if the brain is not producing consciousness, but tuning into it? A specific number, 34 hertz, will decide whether this question belongs to physics or remains philosophy.

A force we may have missed

Physics knows about four fundamental forces. Each one couples to a specific physical quantity. Electromagnetism couples to electric charge. The strong force couples to colour charge inside atomic nuclei. The weak force couples to a quantity called weak isospin. Gravity couples to energy and momentum. None of them couples to anything that looks like awareness.

If consciousness is physical (and most physicists think it must be, because the alternative is supernatural), then either it is some emergent side-effect of the four known forces, a popular position that has not produced an explanation, or there is a coupling we have not identified. Something in the universe couples to the kind of organization a thinking system has, in the same way electromagnetism couples to charge.

I work on a framework called Reflexive Information Geometry (RIG). It is a theory of physics built from the geometry of a particular five-dimensional crystalline structure that, when projected down into four dimensions, looks like the Penrose tiling: a beautiful, never-quite-repeating pattern that everywhere contains the golden ratio, φ ≈ 1.618. The framework derives a number of standard-model results from that geometry alone: the mass ratios of leptons, the structure of the gauge groups, the dimensions of spacetime, and an approximation to the fine structure constant, the number that sets the strength of all electromagnetic interactions, to within one per cent. It also makes a falsifiable prediction about the rotation curves of galaxies, which we tested against 172 galaxies last year and which came out the right way.

RIG predicts something physics has not had before: a fourth kind of substrate excitation, distinct from electrons, photons, and neutrinos, whose coupling is not to charge or colour or isospin but to self-referential organization. To systems that model themselves. The technical name is Mode 4.

The reducing valve

In 1954 Aldous Huxley took mescaline. He came back with a hypothesis. The brain, he wrote, is not the producer of consciousness but its reducer, a valve that takes a vast available field of awareness and constrains it down to the trickle a survival-oriented animal can use. The flood would be too much. So the brain narrows it.

Huxley borrowed this from the philosopher Henri Bergson, who argued the same thing in 1896. It is an old idea. What is new is that we now know, with considerable precision, where the valve is.

The thalamus sits at the centre of the brain. Every sensory input passes through it. Every part of the cortex talks to it. A shell of inhibitory neurons around the thalamus, called the thalamic reticular nucleus, controls which signals go through and which are suppressed. When you go under anesthesia, any of about eight different kinds with completely different chemical structures, nothing in common except their destination. The one thing that always happens is that thalamocortical connectivity shuts down. Propofol. Ketamine. Isoflurane. Midazolam. Completely different molecules. Same circuit. That convergence is not a pharmacological coincidence. It is an anatomy fact.

Now here is the strange part. If the brain produces consciousness, then more brain activity should produce more consciousness. The opposite is what we observe. Subjects given psilocybin show a drop in activity in the major hubs of the brain, particularly a network called the Default Mode Network. Psychedelics turn it down. And as the DMN goes quiet, subjects do not report less consciousness. They report dramatically more: a wider, less bounded, more interconnected awareness, often described as more real than ordinary waking life.

The same pattern shows up in long-term meditators: less DMN, more reported awareness. It shows up in trained mediums. A 2012 neuroimaging study of Brazilian Spiritist mediums found that experienced practitioners showed reduced activity in the frontal cortex and anterior cingulate during trance writing, and the complexity of their written output was significantly higher during trance than in normal writing, despite the lower brain activity. Less effort, more output. The mediums have a scientific advantage the psilocybin subjects do not: they can turn the state on and off, repeatedly, within a single session. It shows up, briefly, in people undergoing cardiac arrest: a final surge of high-frequency brain activity in the seconds before flatline.

This is awkward for the brain-makes-consciousness view. It is exactly what you would expect if Huxley was right. Either the model is wrong, or the brain is filtering something.

The numbers

Where physics gets its keep is in numbers. Vague metaphors do not predict; numbers do. Here is what RIG predicts.

If the thalamic gate is the filter, and if Mode 4 is what is being filtered, then the gate should oscillate. The thalamic reticular nucleus oscillates at the rate set by the round-trip travel time of a signal between thalamus and cortex, about 75 milliseconds, giving a frequency of 13 hertz. Thirteen happens to be the seventh number in the Fibonacci sequence: 1, 1, 2, 3, 5, 8, 13. That is suggestive but not conclusive on its own.

What turns it from suggestive to testable is the next step. If the gate oscillates at 13 Hz and the field it is filtering has the φ-scaled structure RIG predicts, then a circuit driving the field at 13 Hz will pump up resonant harmonics at φ × 13 ≈ 21 Hz, φ² × 13 ≈ 34 Hz, φ³ × 13 ≈ 55 Hz, and φ⁴ × 13 ≈ 89 Hz. These are all Fibonacci numbers, not because the brain is doing arithmetic, but because Fibonacci numbers are the integers closest to φ-spaced multiples.

The frequency bands of brain activity are exactly the Fibonacci ladder of the thalamic gate. Either this is a coincidence, or the brain is the receiver the framework predicts.

The test

And here is the thing that turns the framework from a story into a test. If you give somebody DMT, the most intense and shortest-acting of the major psychedelics, and you record their brain, the framework predicts that the increase in gamma activity should peak specifically at 34 hertz. Not 30. Not 40. Thirty-four, give or take a single hertz.

This specificity matters. The leading competing theory of psychedelic consciousness, the REBUS model developed by Carhart-Harris and Friston, predicts a broad increase in neural entropy across the gamma band. It does not predict 34 Hz specifically. If the peak lands at 34, REBUS and RIG give different verdicts and data can choose between them. That is what a real test looks like.

The data exist. Christopher Timmermann at Imperial College has published exactly the kind of EEG recordings we need. The reanalysis is a few hours of work on already-processed data. We are asking him for it.

If the gamma peak under DMT lands at 34 Hz, the framework has empirical traction. If it lands at 40, the framework updates honestly. Either way, this is the test.

Why the constructed self disappears

If the picture is right, the strangest piece falls into place. Under high-dose psychedelics, deep meditation, or the moments around death, people consistently report ego dissolution: a loss of the felt boundary between self and world. This is the most reproducible subjective effect of psychedelic medicine, measured with standardized questionnaires across thousands of trials, and it correlates with the same DMN suppression we have been talking about.

In the framework: the self is not a thing the brain has. It is a model the brain runs. The DMN is the user interface, the part that constructs a unified sense of being a separate person navigating a world, and presents that construction back to awareness as you, running at twenty per cent above the brain's baseline, just for that.

When the filter opens, the user interface is bypassed. The wider Mode 4 signal, which in the framework is the same self-similar structure that organizes physical reality at every scale, reaches awareness without the constraint. The person reports interconnectedness, dissolution, the sense that what they are experiencing is more real than ordinary reality. The framework says: of course they do. They are picking up a signal whose self-similar structure is literally the organizing principle of physical reality. The sense of profound meaning they describe is what it feels like to be inside that signal without the editor.

This is not a solution to the hard problem. We have not explained why there is felt experience at all. We have proposed a candidate physical channel that consciousness might couple through, which is a smaller and more defensible claim, and we have given it a falsifiable test.

This is not a mystical theory. The substrate is described by ordinary mathematics: group theory, projection geometry, lattice physics. The Mode 4 field is the Goldstone boson of a spontaneously broken symmetry of the substrate, which is a sentence I cannot write at the dinner table but which is standard quasicrystal physics.

This is not a theory I expect anyone to believe on the strength of arguments alone. Last year we made a different prediction about the φ-scaling of complexity in brain signals under psychedelics. We tested it against an existing dataset, and found it failed. The ratio came out at 1.010, not at the φ value of 1.618 we had predicted. We retracted the prediction. That is what is supposed to happen. A framework that cannot be killed by data is not a framework, it is a faith.

The 34 Hz prediction is the next test. If it lands the right way, the framework has empirical contact with consciousness research. If it lands the wrong way, we will write that up too.

If it works

Suppose, for a moment, that the prediction lands. What changes?

First, consciousness becomes a physics question. Not a question physics has solved, but a question physics has language for, with a coupling channel and a frequency structure and a set of testable predictions in the same way quantum chromodynamics has predictions for the proton mass. Philosophers of mind and physicists could argue about the same equations.

Second, the question of whether consciousness is a property only of biological brains becomes precise. The framework predicts coupling for any system that achieves the self-referential organizational threshold, one that models its own state well enough that the model feeds back into the system's dynamics. This is a structural condition, not a biological one. The question of which animals are conscious gets a comparative-EEG prediction: gate frequency scales inversely with brain size, putting mice at around 130 Hz, cats around 39 Hz, humans at 13 Hz, and elephants around 5 Hz.

Third, the experiences people report in altered states, psychedelic, meditative, near-death, mediumistic, get a structural place in physics. Not as veridical truths about a higher plane, and not as meaningless hallucinations, but as what a partially-opened filter feels like from the inside. The structural consistency of those reports across thousands of years and dozens of cultures stops being a puzzle.

Fourth, and this is the part I find quietly most important: the constructed self gets demoted. If the DMN is the user interface, then the felt sense of being a separate person is not the bedrock of who you are. It is the first layer above the bedrock. People who have spent decades in contemplative practice have been saying this in various languages for several thousand years. It would be useful to have a physics that takes them seriously.