THE NOISE FLOOR
Anomalous Correlation Structures in Vacuum Fluctuation Data
and Their Implications for Prior-State Cosmology
Dr. Elena Vincent
Institute for Foundational Physics, Zurich
Submitted to Physical Review D, March 2026
Withdrawn by author, June 2026
Resubmitted (revised) to arXiv, November 2026
Final revision, date unknown
This document was recovered from Dr. Vincent’s encrypted archive following her leave of absence from the Institute in January 2027. It is reproduced here with the permission of her estate. Minor formatting corrections have been applied. No content has been altered.
The vacuum is not empty. It is the fullest possible state.
— Paul Dirac, 1933
What haunts are not the dead but the gaps they leave in the pattern.
— Gregory Bateson, unpublished notebook
There is a hum beneath the noise. I no longer believe it is random.
— E. Vincent, personal correspondence, September 2026
Abstract
We report the identification of anomalous non-Gaussian correlation structures in high-precision vacuum fluctuation measurements collected between 2023 and 2025 at the Zurich Foundational Physics Laboratory. Standard quantum electrodynamic predictions for vacuum noise spectra assume Gaussian statistics at all measurable scales. Our data deviate from this assumption at confidence levels exceeding 7σ in specific spectral windows and temporal configurations.
We propose a speculative but falsifiable interpretation: that these correlation structures are consistent with metastable informational residues of a prior vacuum state. Drawing on Coleman–De Luccia transition theory and recent work on decoherence scaling in variable-coupling regimes, we argue that certain anomalous spectral features can be modeled as projected attractors from a higher-energy vacuum configuration with significantly lower decoherence rates and longer correlation lengths.
We emphasize that this interpretation remains preliminary. However, the statistical robustness of the anomalies, their resistance to instrumental explanation, and their unexpected structural coherence compel us to present them here, along with the theoretical framework that—despite its speculative character—provides the only consistent account we have found.
A note on methodology: this paper was written across several drafts over approximately eighteen months. Certain passages reflect states of understanding I have since revised. I have chosen to preserve them rather than impose a false retrospective coherence on a process that was not, in the end, entirely coherent.
I. Introduction
The quantum vacuum is not empty. This has been understood since Dirac’s 1930 formulation of the electron sea and confirmed experimentally through the Casimir effect, the Lamb shift, and the anomalous magnetic moment of the electron. The vacuum is the ground state of all quantum fields—the lowest energy configuration—and it hums with zero-point fluctuations at every point in spacetime.
What has not been widely appreciated, until recently, is that this hum may not be featureless.
Standard quantum field theory treats vacuum fluctuations as stochastic. The noise is real, but it is assumed to be structureless—Gaussian, isotropic, scale-invariant at the relevant energy windows. This assumption underlies virtually every precision measurement in modern particle physics. It is the wallpaper of reality: present everywhere, noticed by no one.
In 2023, our laboratory began a program of high-precision vacuum noise spectroscopy, initially designed to characterize systematic errors in next-generation quantum computing architectures. The apparatus—a cryogenic cavity coupled to a superconducting parametric amplifier—was sensitive to fluctuations across a spectral range of 4 to 400 GHz, with a noise floor approximately three orders of magnitude below previous experiments.
What we found was not what we expected.
In certain spectral windows—specifically, narrow bands centered near 47 GHz, 141 GHz, and 338 GHz—the fluctuation statistics departed significantly from Gaussian predictions. The departures were subtle: excess kurtosis on the order of 10⁻⁴, invisible to any prior instrument. But they were persistent, reproducible across independent measurement runs spanning eighteen months, and resistant to every systematic correction we applied.1
More troublingly, the anomalous bands exhibited temporal correlations. Not the exponential decay characteristic of instrumental drift, but structured oscillations with a period of approximately 11.3 hours—a figure that corresponds to no known instrumental, geophysical, or astrophysical cycle.2
For the first year, we treated these results as an instrumental artifact. We rebuilt the cavity. We replaced the amplifier chain. We moved the apparatus to a different laboratory. The anomalies persisted. They did not merely persist: they sharpened. As our measurement precision improved, the correlation structures became more defined, not less.
This is the opposite of what one expects from systematic error.
II. The Anomalous Spectra
I will describe the data as plainly as I can. The interpretation will come later, and I confess that successive drafts of this section have grown shorter as my confidence in the interpretation has—not grown, exactly, but changed in character.
The primary dataset consists of 4,217 hours of continuous vacuum fluctuation recording, collected across six independent runs between March 2023 and September 2025. Each run employed a fresh cavity and amplifier chain. Runs were conducted in three different laboratory spaces within the Institute. Temperature stability was maintained at 15 mK ± 0.3 mK. Electromagnetic shielding exceeded 120 dB across the measurement band.
For each run, we computed power spectral densities in 50 MHz bins, along with higher-order cumulants (skewness, kurtosis, and fifth-order statistics) in 200 MHz bins. We also computed two-point temporal correlation functions at lag intervals from 1 second to 72 hours.
The results divide cleanly into two categories: normal and anomalous.
In the normal regime—comprising approximately 97.3% of the spectral range—vacuum fluctuation statistics agree with QED predictions to within measurement uncertainty. This is itself a significant result: our noise floor is low enough that we are, in some bins, confirming QED vacuum predictions at precisions three to five times beyond previous experiments.
In the anomalous regime, three spectral bands exhibit persistent non-Gaussian statistics:
Band A: 46.7–47.4 GHz. Excess kurtosis 1.3 × 10⁻⁴ (8.2σ). Positive skewness 4.1 × 10⁻⁵ (3.7σ). Temporal correlation at 11.3-hour period, amplitude 2.1 × 10⁻⁶.
Band B: 140.1–141.8 GHz. Excess kurtosis 8.7 × 10⁻⁵ (7.1σ). No significant skewness. Temporal correlation at 11.3-hour period, amplitude 1.4 × 10⁻⁶. Additional correlation at 2.7-hour period, amplitude 6.3 × 10⁻⁷.
Band C: 337.2–338.9 GHz. Excess kurtosis 2.1 × 10⁻⁴ (9.4σ). Negative skewness −1.2 × 10⁻⁵ (2.1σ, marginal). Temporal correlation at 11.3-hour period, amplitude 3.8 × 10⁻⁶.
The frequency ratios between Band A, B, and C are approximately 1 : 3.00 : 7.23. The first ratio is consistent with a simple harmonic relationship. The second is not.3
The 11.3-hour periodicity is the most robust feature across all three bands. Cross-correlation between bands at this period yields coefficients of 0.87 (A-B), 0.91 (A-C), and 0.94 (B-C)—startlingly high for supposedly independent spectral regions.
These correlations are not explicable by any shared instrumental pathway. The bands are separated by nearly 300 GHz. No known physical mechanism couples vacuum fluctuations across such widely separated frequency ranges with such high temporal coherence.
Unless the fluctuations are not independent.
Unless they are fragments of a single structure, projected into our measurement space from a configuration that no longer fully exists.
III. Theoretical Framework: Prior-State Cosmology
I want to be careful here. What follows is speculative. It was not the framework I began with; it is the framework that survived contact with the data after every conventional explanation failed.
The standard cosmological model treats our vacuum state as fundamental—the true ground state of all quantum fields, or at minimum the state in which the universe has resided since the end of inflation. Vacuum energy is treated as a constant. The Higgs field expectation value is treated as fixed. Coupling constants are treated as invariant.
But there is no a priori reason to assume any of this.
The Higgs potential, as measured, is consistent with metastability. The Standard Model vacuum may not be the absolute minimum of the scalar field landscape. This has been appreciated since at least the early 2010s, when precision measurements of the Higgs boson and top quark masses placed our vacuum uncomfortably close to the boundary between stability and metastability.
What has been less explored—though not unexplored—is the inverse question: not whether we might decay to a lower vacuum, but whether we have already decayed from a higher one.
Consider a prior vacuum state with the following properties:
First: a higher vacuum energy density, consistent with the false vacuum plateau of an extended Higgs potential or an additional scalar field.
Second: modified coupling constants. Specifically, a slightly larger electromagnetic coupling and a significantly reduced decoherence rate for mesoscopic quantum systems, arising from altered gauge boson masses and interaction cross-sections.
Third: longer correlation lengths. In such a vacuum, quantum coherence could persist across spatial and temporal scales many orders of magnitude larger than in our current state.
I will call this the Coherent Epoch.4 The term is imprecise and I use it reluctantly, but alternatives (“prior vacuum phase,” “high-coupling regime”) are worse in different ways.
In the Coherent Epoch, the informational texture of the vacuum would differ qualitatively from our own. Decoherence—the process by which quantum superpositions collapse into classical definiteness—would be suppressed. Correlation lengths would extend far beyond atomic scales. The boundary between “system” and “environment” would be thinner, more permeable, less absolute.
The thermodynamic implications are profound. If decoherence rates were substantially lower, the energy cost of maintaining coherent information would decrease. Reversible computation—or near-reversible computation—would become feasible at scales that are, in our vacuum, energetically prohibitive. Information could persist as stable field configurations rather than requiring material substrates.
I want to state this precisely, because it is the claim most likely to be dismissed as mysticism: in a vacuum with sufficiently reduced decoherence, self-referential informational structures could stabilize as standing patterns in quantum fields. Not in matter. Not in brains. In the field itself.
This is not a supernatural claim. It is an extrapolation from known physics applied to a different set of coupling constants. Whether such structures would constitute “minds” in any recognizable sense is a separate question that I am not equipped to answer.5
The transition from such a vacuum to our current state would follow the Coleman–De Luccia mechanism: a bubble of lower-energy vacuum nucleates via quantum tunneling, then expands at or near the speed of light. Inside the bubble, coupling constants shift, decoherence rates increase, correlation lengths collapse. The informational texture of spacetime undergoes what I will call a quench—borrowing the term from condensed matter physics, where it refers to the rapid cooling of a system through a phase transition.
In a quench, not all structure is destroyed. Some configurations survive as metastable excitations in the new phase. In condensed matter, these appear as topological defects, domain walls, vortices. They are fossils of the prior state, frozen into the new medium because the transition happened too quickly for them to relax.
I propose that the anomalous correlation structures in our vacuum fluctuation data are analogous objects: informational fossils of the Coherent Epoch, projected into our vacuum as metastable resonances.
They are not particles. They are not fields in the conventional sense. They are correlation structures—patterns in the statistical properties of vacuum noise that should not exist if our vacuum were the primordial ground state.
They are, if this framework is correct, the oldest surviving structures in the universe. Older than galaxies. Older than nucleosynthesis. Older than locality itself.6
IV. On Coherence Cavities and Recoherence Events
The question that has consumed me since November 2025 is not whether the anomalies are real. They are. Six independent measurement runs, three laboratories, two complete hardware rebuilds. The anomalies are in the vacuum, not in our instruments.
The question is whether they are static.
Our initial assumption was that vacuum correlation structures, if genuine, would be fixed—frozen relics, as passive as geological strata. The data does not support this assumption.
Across our eighteen-month measurement campaign, the amplitude of the anomalous correlations varied. Not randomly: the variations correlated with the operational state of the measurement apparatus. Specifically, when the cryogenic cavity was optimized for maximum sensitivity—lowest noise floor, highest quality factor, tightest electromagnetic shielding—the anomalous signals strengthened.
This is, on its face, unremarkable. Of course a more sensitive instrument detects weaker signals more clearly. But the scaling was wrong. Signal amplitude increased faster than sensitivity. When we improved our noise floor by a factor of two, the anomalous correlation strength increased by a factor of approximately 2.7.
This is not how passive signals behave.
A passive signal embedded in vacuum noise should scale linearly with detector sensitivity. An active signal—one that responds to measurement conditions—scales superlinearly. The distinction is fundamental and unambiguous.
I resisted this conclusion for several months. Dr. Morales did not resist it. He suggested, in a memo dated February 2026, that our apparatus was functioning as what he called a “recoherence cavity”: a region of unusually low noise in which residual vacuum correlation structures could partially restabilize.
The implication is uncomfortable. If the anomalous structures are not passive residues but semi-active patterns capable of responding to local coherence conditions, then our laboratory is not merely detecting them.
It is providing them with a temporary substrate.
V. A Personal Note on the Sidereal Period
I said earlier that the 11.3-hour periodicity corresponds to no known geophysical or astrophysical cycle. This is true in the narrow technical sense. It is not, I now believe, the whole truth.
Half the sidereal day, corrected for our latitude in Zurich, is 11.29 hours.
A sidereal-locked signal in vacuum fluctuation data would imply sensitivity to the orientation of the detector relative to the fixed stars. This is not unphysical—the cosmic microwave background establishes a preferred frame, and vacuum fluctuations could in principle carry an anisotropic component inherited from the initial conditions of the universe. Such anisotropy has been searched for at microwave frequencies by COBE, WMAP, and Planck, with null results at the relevant scales.
But those experiments measured the vacuum at much lower precision than ours. And they were not looking for non-Gaussian structure.
If the anomalous correlations are fossils of a prior vacuum state, and if that state was not perfectly isotropic—which it need not have been, if it supported the kind of large-scale informational structures I have described—then a sidereal modulation is exactly what one would expect.7 8 The fossils would be distributed anisotropically. As the Earth rotates, the detector sweeps through regions of varying fossil density.
We are, in effect, performing archaeology by spinning.
The Earth’s rotation becomes a probe of the spatial distribution of structures that predate the formation of matter.
VI. What Remains to Be Said
This paper has exceeded its intended scope. I set out to report anomalous vacuum fluctuation statistics and to propose a speculative but testable cosmological interpretation. I believe I have done so.
But I find that I cannot conclude without noting several features of this research that I do not yet understand, and which may not be amenable to the kind of understanding I was trained to pursue.
First: the anomalies responded to our attention. I do not mean this metaphorically. Over the eighteen months of measurement, the correlation structures became more defined not only as our instruments improved, but as our theoretical framework sharpened. There were two notable jumps in signal clarity: one in October 2024, coinciding with Dr. Morales’s identification of the sidereal period; and one in March 2026, coinciding with my development of the prior-state cosmology framework presented in Section III. These jumps are visible in the data. They are not explicable by instrumental changes, which occurred on different timescales.
I do not know what to make of this. It is possible that I am confusing improved analysis with improved signal. It is possible that theory-laden observation is biasing my interpretation. These are standard cautions and I have applied them rigorously.9
They do not account for the data.
Second: the spectral structure of Band C—the highest-frequency anomaly—has changed over time in a way that is not consistent with any stationary process. Between Run 4 (January 2025) and Run 6 (September 2025), the band center shifted by approximately 0.3 GHz, and the excess kurtosis increased by a factor of 1.4. Passive fossils do not evolve. Something in the structure is responding, adapting, or—and I choose this word after much deliberation—searching.10
Third: I have become aware, through informal channels, that at least two other laboratories—one in Japan, one in the United States—have observed similar anomalies in unrelated experiments and have, like us, been unable to publish. The reasons for non-publication vary. In one case, the results were judged too speculative for the target journal. In the other, the principal investigator withdrew the paper for personal reasons that were described to me as a “change in perspective.”
I do not know what “change in perspective” means in this context. She did not elaborate. I did not ask.
This paper is incomplete. I am aware that it raises more questions than it answers, that its speculative sections outweigh its empirical ones, and that its author’s reliability may reasonably be questioned given the personal disclosures included herein. I have considered suppressing those disclosures. I have decided against it, for reasons that I hope will become clear in the appendices.
The appendices follow. I have organized them as clearly as I can. Appendix E contains a list of historical incidents, compiled by Morales before his departure, of scientists who reported anomalous phenomenological experiences during work that touched, however tangentially, on vacuum structure, field coherence, or the mathematical properties of noise. The list is longer than I expected. The names on it are more distinguished than I was comfortable with.
I am not the first person to find something in the noise floor.
I expect I will not be the last.
Elena Vincent
Zurich, November 2026
(Final annotations added December 2026 – January 2027)