Life Runs on Voltage

Low-Fidelity Biology, Bioelectrical Dissonance, and the Control Layer Missing From Wireless Safety

An RF Safe synthesis and research agenda
By John Coates, Founder, RF Safe
August 2026

The central failure of wireless safety policy is not that it chose the wrong disease. It chose the wrong biological question. Regulators ask how much energy heats tissue. Living systems depend on whether voltage, ion timing, redox signaling, recovery, and collective cellular coordination retain their fidelity.

Executive summary

Life is not controlled by chemistry alone. Every cell maintains a transmembrane voltage. Cells use ion gradients, gap junctions, calcium pulses, mitochondrial redox signals, electrical polarity, chromatin state, and communication with neighboring cells to decide what to build, repair, preserve, or dismantle. These processes are not decorative consequences of biology. Experiments show that changing them can change biological outcomes.

Robert O. Becker’s regeneration research helped establish that injury and repair are accompanied by organized endogenous electrical currents. Later work measured these currents directly and showed that interfering with them could impair regeneration, while applied currents could partially restore regenerative responses in normally non-regenerating tissue.

Michael Levin’s laboratory carried the principle much further. In Xenopus, oncogene-expressing cells formed tumor-like structures in association with membrane depolarization. The altered voltage state appeared before visible tumor morphology. When researchers forced the cells back toward a more normal, hyperpolarized state, tumor-like formation fell—even though the oncogene remained present. In subsequent work, manipulating voltage in distant host tissue also altered tumor incidence, demonstrating that the relevant control state can be distributed across a cellular network rather than confined to the genetically altered cell.

Planarian experiments show the same principle at the scale of anatomy. A transient disruption of gap-junction communication caused genetically normal flatworms to regenerate non-native, species-like head shapes. Those heads later remodeled toward the species-typical form. In a separate study, a brief bioelectric perturbation created a stable, cryptic state in otherwise normal-looking worms, causing later amputations to regenerate two-headed forms at a reproducible rate. Morphology was not simply read from DNA as a fixed construction script. It emerged from an interaction between inherited cellular capacity and the bioelectric state through which cells coordinated their decisions.

These results support a crucial distinction:

• DNA and chromatin provide the evolved biological prior—the molecular capabilities, regulatory constraints, and stable developmental possibilities available to the cell.
• Bioelectric state provides part of the runtime context—where the cell is, what its neighbors are doing, and which region of the available state space should be expressed now.
• The cell is a local inference engine—integrating voltage, calcium, redox state, metabolism, mechanics, extracellular signals, and nuclear architecture to choose its next action.
• Morphology is a collective output—the result of many cells repeatedly coordinating local decisions across time.

RF Safe calls the precision of this process biological fidelity. When the timing, localization, phase, amplitude, recovery, or interpretation of those signals is persistently degraded, biology may continue to operate while becoming noisier, less coordinated, more energetically expensive, and less capable of returning to baseline. That is low-fidelity biology.

Low-fidelity biology is not a new disease label. It is an upstream susceptibility state. It does not assert that one exposure deterministically causes one diagnosis. It predicts that loss of signaling and repair precision changes the probability distribution of downstream outcomes: rare errors become less rare; latent vulnerabilities are expressed more often; recovery takes longer; developmental mistakes become harder to correct; damage is more likely to persist in long-lived cells; and age-associated phenotypes can appear earlier.

RF Safe calls the environmental mismatch that produces this loss of coherence bioelectrical dissonance. Poor nutrition, polluted air and water, circadian disruption, infection, chemical toxicants, sleep loss, and other physiological burdens can all lower biological fidelity. Non-native electromagnetic fields deserve special attention because exposure can be involuntary, continuous, multi-source, waveform-rich, present during sleep, and difficult to escape even when an individual changes personal behavior.

The proposed S4–Mito–Spin framework identifies three plausible transduction and amplification branches:

1. S4 and voltage-gated ion-channel density: time-varying fields may perturb mobile ions and the voltage-sensing machinery that controls channel opening, producing timing errors in calcium and other ionic signals.
2. Mitochondrial density and calcium–redox coupling: mitochondria can convert altered calcium timing and metabolic demand into changes in membrane potential, ATP reserve, reactive species, and recovery.
3. Spin-active redox density: flavins, hemes, iron-sulfur centers, and radical-pair chemistry provide an additional candidate route through which weak magnetic fields may influence reaction probabilities under specific molecular conditions.

These three branches operate through a fourth filter: persistence. Cell longevity, tissue turnover, antioxidant reserve, repair capacity, developmental timing, and opportunity for recovery determine whether a transient perturbation is erased, adapted to, or retained as a lasting phenotype.

This 3+1 architecture explains why biological effects can be real without being uniform. The field is only one input. The receiver matters. Tissue density matters. Differentiation state matters. Genotype matters. Waveform matters. Exposure history matters. Recovery matters.

Several findings already fit that prediction:

• Human cord-blood cells exposed to a pulsed UMTS signal showed a transient oxidative response whose magnitude increased with differentiation state. The response disappeared by three hours and did not produce persistent DNA damage under the tested conditions. That is not “nothing happened.” It is a demonstration that cellular state changes acute transduction gain, while buffering and recovery determine persistence.
• In a randomized, double-blind, sham-controlled human experiment, 3.6 GHz 5G exposure altered sleep-spindle frequency in carriers of one CACNA1C genotype but not the matched comparison genotype. A common regulatory variant in a calcium-channel gene changed the physiological response to the same exposure.
• A 2026 Cell paper identified CYB5B as essential to an engineered electromagnetic-field-responsive gene switch and found that activation depended on rhythmic calcium oscillations rather than generic calcium elevation. That result establishes a direct experimental principle: electromagnetic responsiveness can depend on a specific molecular receiver and on the pattern of calcium signaling.
• The FDA-authorized TheraBionic P1 device uses low-intensity, amplitude-modulated RF as a therapeutic input, and its labeling contraindicates use with calcium-channel blockers. Federal regulation therefore already recognizes that properly structured, non-heating RF can be biologically meaningful. The same fact that makes precision electromagnetic therapy possible makes waveform-blind environmental safety assumptions untenable.

The public-health conclusion is direct. A safety system that measures only time-averaged energy absorption cannot establish protection of biological timing fidelity. It does not measure membrane-voltage stability, calcium-waveform jitter, mitochondrial reserve, redox recovery, chromatin response, gap-junction coordination, genotype-by-exposure interaction, developmental vulnerability, or the cumulative effect of multiple signals.

The National Toxicology Program’s heart schwannomas and brain gliomas, the Ramazzini Institute’s independent heart-schwannoma signal, repeated oxidative and genetic findings, reproductive effects, neurological responses, and low-intensity bioactivity are not fifty unrelated mysteries. They are a demand for a receiver-centered model of electromagnetic biology.

The task now is not to keep arguing about whether non-ionizing radiation can behave like an ionizing photon. It is to test how time-structured fields interact with biological control systems that are themselves electrical, rhythmic, redox-sensitive, state-dependent, and developmentally timed.

The standard measures heat. Biology measures fidelity. Public policy must begin protecting what life actually uses.

The unified master map: one trunk, four outcome domains

 

The developmental, cancer, autoimmune, and metabolic pathways developed by RF Safe should not be read as four claims that RF independently causes four named diseases. They are four views of one proposed systems architecture.

The shared trunk is:

time-structured environmental input → receiver and persistence gates → bioelectrical dissonance → low-fidelity cellular computation → tissue-specific, probabilistic outcomes

Each stage answers a different question.

Stage 1: What reaches the organism?

The biologically relevant input is not exhausted by carrier frequency or time-averaged power. It includes modulation, pulse structure, repetition, duty cycle, polarization, peak-to-average ratio, exposure chronology, and the opportunity—or absence of opportunity—for recovery. RF and ELF are not physically interchangeable, but modern RF systems can contain low-frequency envelopes and repetition structures within biological timing domains.

Other fidelity stressors also enter here: air pollution, chemical toxicants, circadian disruption, poor sleep, inadequate nutrition, infection, and acute metabolic or immune demand. The model is multi-hit from the beginning.

Stage 2: What receives and amplifies the input?

The 3+1 density gate determines whether a field becomes a meaningful biological perturbation:

• S4 voltage sensors and excitable-membrane density provide a proposed ionic timing gate.
• CYB5B, mitochondria, and calcium–redox coupling provide a receiver and amplification layer.
• Spin-active flavin, heme, iron-sulfur, and radical-pair chemistry provide candidate magnetic-field-sensitive reaction pathways.
• The persistence gate—developmental timing, cell longevity, turnover, repair, buffering, and recovery—determines whether the perturbation disappears or becomes history.

Stage 3: What is the common upstream failure?

Bioelectrical dissonance is loss of coordination within the control layer. It can appear as Vmem drift, calcium phase jitter, mistimed ionic pulses, altered redox oscillation, incomplete recovery, or degraded gap-junctional coherence.

The result is low-fidelity biology: the system continues to function, but with less precision, reserve, discrimination, and capacity to return to baseline.

Stage 4: Where does the loss of fidelity go?

Cells interpret their genome through a stateful local environment. When that environment becomes noisier, the downstream error is expressed through the machinery dominant in that cell:

• chromatin and transcription;
• DNA repair and checkpoints;
• mitochondrial and metabolic decisions;
• immune self-versus-danger classification;
• differentiation and tissue patterning.

Stage 5: Why are the outcomes different?

Tissue, genotype, age, differentiation state, developmental window, prior stress, cell turnover, and exposure history determine the branch. The same upstream reduction in fidelity can therefore present as altered development in one context, metabolic drift in another, impaired immune resolution in another, or persistence of a genome-maintenance error in a long-lived lineage.

This is the meaning of a meta-disease state. It is not a universal disease caused by one field. It is a loss of upstream biological quality control that changes the probability of many downstream outcomes.

The master map is therefore a hypothesis map, not a clinical risk calculator. Its purpose is to identify shared variables, clarify which links are established and which are proposed, and generate experiments capable of falsifying the integrated model.

1. The body is an electrical decision-making network

The familiar biochemical description of life is true but incomplete. Proteins bind. Enzymes catalyze. Receptors signal. Genes are transcribed. Yet all of this takes place inside a continuously regulated electrical architecture.

Every living cell separates charge across its membrane. The resulting transmembrane potential, or Vmem, is generated by ion channels, pumps, transporters, membrane composition, cell geometry, and metabolic energy. Vmem influences calcium entry, pH, transport, proliferation, migration, differentiation, apoptosis, gene expression, and communication through gap junctions.

At the tissue level, cells do not make decisions in isolation. They exchange ions and small molecules, respond to electrical gradients, sense mechanical constraints, and compare their local state with neighboring cells. A cell’s behavior therefore depends on more than its genome. It depends on the state in which that genome is being interpreted.

This is the decisive lesson of modern bioelectric biology: the genome defines biological capability, but bioelectric state helps select biological behavior.

That distinction dissolves a false choice. We do not need to decide whether DNA or bioelectricity “contains” the organism. DNA sequence, three-dimensional chromatin, epigenetic state, membrane voltage, metabolism, cytoskeleton, extracellular matrix, morphogen gradients, and cell-cell signaling form a coupled control system. Removing any one layer from the explanation produces a caricature.

The low-fidelity biology hypothesis begins with that coupled system.

2. Becker’s foundational insight: repair has an electrical dimension

Robert O. Becker’s work was important because it challenged the assumption that endogenous electrical signals were merely byproducts of injury. In the 1960s, Becker examined bioelectric factors associated with amphibian limb regeneration and reported injury-related electrical patterns along the body axis. His work proposed that organized electrical state helped distinguish regenerative tissue from ordinary wound closure.

Later experiments by Richard Borgens and colleagues strengthened the empirical foundation. They measured outward ionic currents from regenerating newt limb stumps in the range of roughly 10 to 100 microamperes per square centimeter during the early post-amputation period. They then interfered with the sodium-dependent component of the current and found impaired or delayed regeneration. In adult frogs, whose limbs normally do not regenerate, minute applied currents induced partial regenerative growth.

The enduring conclusion is not that one electrical waveform “commands” an entire limb. It is more fundamental: injury, polarity, growth, and repair are accompanied by organized endogenous electrical states, and changing those states can change the outcome.

Becker supplied the question that modern wireless regulation still avoids. If living tissue uses weak electrical signals to coordinate repair, why would a safety standard assume that an external field is biologically relevant only when it heats that tissue?

The answer cannot be supplied by energy absorption alone. It requires biology.

3. Levin’s tumor experiments: voltage can redirect phenotype despite oncogenes

The 2013 work of Brook Chernet and Michael Levin provides one of the clearest demonstrations that membrane voltage is not simply a passive marker of cell state.

The researchers expressed several human oncogenes in Xenopus laevis embryos and tadpoles, including mutant KRAS, Gli1, Xrel3, and mutant p53. The manipulated cells formed tumor-like structures. These structures were associated with a depolarized membrane state, and depolarized regions could be detected before overt morphological abnormalities appeared.

The critical intervention came next. The researchers expressed hyperpolarizing ion channels and transporters to restore a more normal Vmem. Tumor-like structure formation fell significantly even though the oncogene remained present.

That experiment changes the conceptual order of causation. The oncogene was an important driver, but its presence did not make phenotype inevitable. The electrical state of the cell acted as a gate on what that genetic perturbation became.

Follow-up work broadened the principle:

• Manipulating bioelectric state in distant host tissue changed the incidence of oncogene-induced tumor-like structures.
• Gap-junction communication participated in the long-range control process.
• Optogenetic hyperpolarization reduced tumor-like structure incidence, and delayed activation after formation increased normalization.
• Different tools and ions produced similar suppression when they converged on Vmem, indicating that the control variable was the voltage state rather than one privileged channel protein.

These frog experiments are not human cancer trials, and the induced structures are properly described as tumor-like structures. Their importance is mechanistic: they demonstrate that a genetic insult and a bioelectric control state interact. They show that cancer-like phenotype is influenced by how a cell interprets its genetic and environmental context.

This leads to a core RF Safe principle:

A mutation is not interpreted in a vacuum. It is interpreted inside a voltage-regulated, metabolically powered, redox-sensitive cellular network. Lower the fidelity of that network, and the same genome can produce a different outcome.

4. Vmem is both a local state and a collective control variable

Why can voltage exert such broad influence?

Vmem sits near the top of multiple biological decision pathways. It regulates the electrochemical force on ions; alters voltage-gated channel activity; changes calcium dynamics; affects transporters; influences gap-junctional communication; and can alter transcription through metabolite transport and chromatin-regulatory mechanisms.

In the Chernet–Levin work, one implicated pathway involved the voltage-sensitive transporter SLC5A8, intracellular butyrate, and histone deacetylase activity. The important point is not that this is the only route. It is that a membrane-level state can reach the nucleus through specific molecular machinery and alter gene regulation.

This creates a recurrent control loop:

membrane voltage → ion and metabolite movement → calcium/redox/metabolic state → gene and chromatin regulation → channel and transporter expression → membrane voltage

The loop is self-reinforcing. A high-fidelity state can stabilize itself. A perturbed state can also become self-maintaining if compensatory changes are repeatedly written into transcription, chromatin, metabolism, and tissue architecture.

That is why recovery is not a secondary outcome. The ability to return to baseline is itself a measure of biological fidelity.

5. Planaria reveal morphogenetic state beyond anatomy

Planarian flatworms make this issue visible because their regeneration is dramatic and experimentally accessible.

In 2015, Emmons-Bell and colleagues briefly blocked gap-junction communication in genetically wild-type Girardia dorotocephala. After amputation, the worms regenerated distinct, non-native head shapes resembling those of several other extant planarian species. The changes were accompanied by altered brain morphology, stem-cell distribution, and membrane-voltage domains.

The experiment is often described too loosely as the production of “ancestral heads.” The accurate phrase is non-native, species-like head morphologies. The distinction matters because the result is already extraordinary without embellishment: transiently changing communication among cells altered the anatomical solution regenerated by an unchanged genome.

The non-native heads were not equally stable. Over subsequent weeks they remodeled toward the normal G. dorotocephala head shape. The authors described the result using an attractor-landscape framework: different anatomical states can occupy wells of different stability, and a transient perturbation can move the system into an alternative basin before it returns to the more strongly canalized native form.

The 2017 work of Durant and colleagues revealed a complementary result. A brief perturbation of bioelectric gradients created a stable, cryptic pattern memory in worms that later appeared anatomically normal. When recut, they continued to generate a reproducible mixture of normal and two-headed forms across rounds of regeneration. Resetting the electrical state restored the wild-type outcome.

Together, these studies show three things:

1. Visible anatomy is not the whole control state. A normal-looking organism can carry a latent bioelectric pattern that changes future outcomes.
2. A transient exposure can have delayed consequences. The state revealed by the next injury may differ from the state visible immediately after the perturbation.
3. Biological memory has levels of stability. Some imposed states fade; others persist through repeated regeneration.

These are precisely the features a low-fidelity model predicts: hidden state, altered recovery, history dependence, and probabilistic divergence after a later challenge.

6. The cell is the local inference engine

The planarian results do not require a miniature head blueprint hidden in every cell. Nor do they require DNA to be reduced to a passive protein catalog.

The more useful model is local, iterative, and state-dependent.

At a wound edge, a cell has access to local information: voltage, polarity, calcium, metabolites, mechanical stress, morphogens, gap-junction signals, extracellular matrix, and the behavior of neighboring cells. It combines that information with its inherited and acquired molecular machinery. It then chooses a local action—divide, migrate, differentiate, secrete, die, or change connectivity. That action changes the environment seen by the next cell. Morphology emerges from repeated local decisions.

RF Safe’s Cellular Latent Learning Model, or ceLLM, describes this as vector-driven local inference:

• The local bioelectric and biochemical state acts as a query or conditioning vector.
• DNA sequence, epigenetic state, and three-dimensional chromatin architecture constrain the possible outputs.
• The cell selects a context-appropriate action from that evolved possibility space.
• Neighboring cells exchange updated state, producing collective convergence toward a tissue-level attractor.

This framework does not make body geometry the primary memory store. Geometry is the runtime environment produced by the network. Nor does it replace DNA with an ethereal bioelectric blueprint. It treats the genome and chromatin as an active, physically organized prior whose expression is conditioned by the cell’s electrical and metabolic context.

That is a stronger view of DNA than the phrase “protein code” allows.

7. DNA is sequence, structure, and accessible probability

Modern genomics already shows that the nucleus is not a one-dimensional instruction tape. Chromosomes fold into loops, domains, compartments, and contacts that bring regulatory elements together or keep them apart. Cohesin, CTCF, transcription factors, histone modifications, DNA methylation, nuclear positioning, and mechanical state all influence which genomic regions are accessible and which programs can run.

Three-dimensional genome architecture has causal roles in gene regulation, differentiation, development, and disease. It changes across cell types and states. This supports the empirical core of the ceLLM interpretation: the same DNA sequence can support different outputs because its physical and regulatory configuration changes what is available to be expressed.

ceLLM takes one further, testable step. It proposes that sequence plus three-dimensional chromatin state can be modeled as an evolved probability landscape—a biological prior shaped by development and evolutionary history. Bioelectric, calcium, redox, metabolic, mechanical, and photonic inputs condition sampling from that landscape.

In this interpretation:

• Sequence constrains the library.
• Chromatin topology changes access and coupling.
• Epigenetic marks adjust the local landscape.
• Bioelectric state provides spatial and temporal context.
• Cellular action is the conditional output.

The phrase “geometric weight matrix” is a model, not a claim that chromatin has already been shown to operate identically to an artificial neural network. Its scientific value lies in the predictions it generates: sustained changes in bioelectric and redox state should be accompanied by measurable changes in chromatin contact patterns, transcriptional accessibility, and the stability of future morphological decisions.

That can be tested with voltage imaging, calcium imaging, Hi-C or Micro-C, ATAC-seq, transcriptomics, targeted chromatin perturbation, and repeated regeneration.

A theoretical bridge: the engine, the logbook, and the fidelity gap

Paul Cooney’s “Dual-Histories Ledger” is an ambitious interpretive framework rather than a published Science Advances article. It should not be cited as an established unified theory. Yet one part of its language supplies a useful conceptual bridge to ceLLM.

Cooney distinguishes between a global “engine” that retains the full reversible state and a local “logbook” produced when an observer has access to only part of that state. The local record becomes mixed, incomplete, entropic, and history-dependent. The distinction is expressed using standard open-quantum-system concepts such as the partial trace and relative entropy, then extended speculatively across gravity, string theory, biology, and aging.

ceLLM begins at the opposite end. It does not start with gravity or a universal quantum ledger. It begins with the cell as an embodied observer whose access is local. A cell never sees the entire organism. It reads membrane voltage, ion timing, metabolites, neighboring-cell signals, cytoskeletal tension, extracellular matrix, and the accessible configuration of its own genome. From that partial state it must decide what to do next.

The frameworks align most usefully at three points:

The evolved prior and the local record

In ceLLM, DNA sequence and chromatin architecture provide an evolved prior: a structured landscape of biological possibilities. The local cell state acts as the query. Each response then becomes part of the cell’s history through transcription, epigenetic marking, organelle remodeling, altered channel expression, accumulated damage, and adaptive compensation.

The cell therefore contains both a relatively stable inherited model and an accumulating local logbook.

Distinguishability is a biological resource

A cell must distinguish self from danger, anterior from posterior, growth from repair, calcium signal from calcium noise, and transient stress from a condition requiring durable adaptation. High-fidelity biology preserves these distinctions. Bioelectrical dissonance compresses or corrupts them.

This is where “entropic waste” becomes more than a synonym for damage. It is any input that reduces the system’s ability to make correct distinctions at the right time and place. The relevant loss is not only thermodynamic efficiency; it is degraded inference.

Adaptive patches become the biological ledger

When a stress recurs, the cell compensates. It changes antioxidant defenses, channel expression, metabolism, chromatin accessibility, inflammatory thresholds, and repair priorities. A successful patch preserves function. But repeated patching can move the local state progressively away from its former high-reserve baseline.

That provides a disciplined way to describe aging and chronic vulnerability:

repeated demand → compensation → adaptive record → reduced reserve → greater dependence on the patch → earlier threshold crossing

The value of the Ledger analogy is not that it proves ceLLM’s physics. It is that it sharpens a testable biological proposition: local records of prior stress should be measurable, state-dependent, and predictive of altered response to the next challenge.

 

Figure note: This poster presents an authorial synthesis, not an authentic Science Advances research article. Its engine/logbook distinction is used here as a conceptual analogy, not as evidence for ceLLM.

8. Calcium is a timing code, not a bucket

Biology does not interpret calcium only by asking how much is present. Cells distinguish signals by amplitude, frequency, duration, localization, rise time, decay time, and phase relationship with other oscillations.

Classic experiments by Dolmetsch and colleagues showed that the frequency of calcium oscillations can determine transcriptional efficiency and gene specificity. Different calcium amplitudes and durations preferentially activate different transcription factors. This is information coding in a literal experimental sense: the waveform changes the output.

Mitochondria are embedded in that code. They take up calcium at specialized contact sites, use it to adjust metabolism, influence ATP production, shape cytosolic calcium, and alter reactive oxygen species. A calcium pulse arriving at the wrong time can therefore be different from the same amount of calcium arriving in the correct rhythm.

The relevant biological variables include:

• pulse frequency;
• peak amplitude;
• phase relationship with mitochondrial and circadian rhythms;
• spatial origin and propagation;
• burst duration;
• inter-pulse interval;
• termination kinetics;
• refractory period;
• return to baseline;
• trial-to-trial jitter.

This is the key conceptual upgrade from concentration to fidelity. A cell can maintain a normal average calcium concentration while losing the information carried by its calcium waveform.

A regulatory system centered on average RF power is poorly matched to a biological system that encodes decisions in temporal pattern.

9. CYB5B supplies a new molecular foothold

In 2026, Kim and colleagues reported an electromagnetic-field-inducible gene switch in Cell. A CRISPR screen identified cytochrome b5 type B, CYB5B, as an essential mediator and likely sensor in their engineered system. Activation depended on rhythmic calcium oscillations rather than a generic increase in intracellular calcium.

The authors used the system to control gene expression in vivo, including spatial and temporal induction. The result does not establish that every environmental field engages CYB5B in the same way. What it does establish is more fundamental and highly consequential:

A defined electromagnetic exposure can be converted into gene regulation through a specific molecular mediator and a calcium-rhythm code without relying on tissue heating.

This finding directly undermines the claim that non-thermal electromagnetic bioactivity lacks a plausible receiver. It also tells researchers what to measure next:

• Does CYB5B respond under telecom-relevant waveforms and intensities?
• Which envelope frequencies, pulse widths, duty cycles, and exposure durations produce calcium oscillations?
• Is the response dependent on mitochondrial position, differentiation state, redox baseline, or genotype?
• Do CYB5B loss-of-function and rescue experiments abolish and restore the response?
• Does repeated activation change chromatin accessibility or future responsiveness?

Those are no longer philosophical questions. They are experimental protocols waiting to be run.

10. The carrier is not the biological clock

Wireless signals must be described at more than one timescale.

A Wi-Fi, cellular, Bluetooth, DECT, or other wireless signal has a high-frequency carrier, but its real exposure pattern also contains frames, bursts, amplitude variation, duty cycles, packet timing, beacon intervals, synchronization signals, and low-frequency envelopes. Those time structures can fall in the hertz-to-kilohertz domain where biological oscillations also operate.

This does not mean that a 60 Hz magnetic-field experiment is physically identical to exposure from a Wi-Fi router or a 5G base station. Carrier frequency, electric and magnetic components, polarization, spatial distribution, field strength, near- versus far-field conditions, modulation depth, and coupling all differ.

The scientifically correct bridge is narrower and stronger:

Modern RF carriers can deliver low-frequency envelopes, pulse trains, and repetition structures that overlap the timing domain used by biological control systems. Therefore, testing only carrier frequency and average power can miss the exposure variables most relevant to a timing-sensitive receiver.

The field is not the cell’s clock. The cell is the clock. An external signal becomes biologically relevant only through the way a receiver transduces it. That is why identical average power can produce different results when modulation, duty cycle, tissue state, or genotype changes.

11. S4: a proposed nanoscale gate for timing error

Voltage-gated ion channels contain charged voltage-sensing domains. In many channels, the positively charged S4 helix moves in response to changes in the electric field across the membrane, helping control whether the pore opens or closes.

The ion forced-oscillation model advanced by Panagopoulos and colleagues proposes that polarized, time-varying electromagnetic fields can displace mobile ions near these voltage sensors. Because the separation is nanoscopic, the resulting Coulomb forces could, under the model, perturb gating even when the field does not measurably heat tissue.

The S4 hypothesis is not yet a universally established explanation for all RF bioeffects. It is valuable because it is specific and falsifiable. It identifies:

• a physical target;
• a force pathway;
• a dependence on polarization and time structure;
• a mechanism for nonlinear response;
• a route from weak external fields to ionic timing changes;
• experimental interventions that should block, enhance, or reshape the effect.

Tests should compare channels with altered S4 charge, channel blockers, membrane composition, ion concentration, field polarization, envelope frequency, and pulse structure while directly recording gating currents and calcium waveforms.

If the effect follows those predicted dependencies, the S4 branch gains support. If it does not, the model must be revised. This is how a serious mechanism advances.

12. Mito: the amplifier and recovery engine

Mitochondria do not merely make ATP. They integrate calcium, oxygen, substrates, membrane potential, redox chemistry, apoptosis, innate immune signaling, and cellular stress.

That makes mitochondria a natural amplifier of small timing errors.

If altered channel gating changes the timing of calcium entry, mitochondria may experience a different calcium load and metabolic demand. Depending on cell type and state, the result can include:

• altered mitochondrial membrane potential;
• changed ATP production and reserve;
• increased or redistributed reactive oxygen species;
• altered nitric-oxide chemistry;
• activation of mitochondrial stress responses;
• changes in fission, fusion, transport, or mitophagy;
• altered apoptotic threshold;
• redox-sensitive changes in transcription and chromatin.

One 900 MHz continuous-wave experiment in mouse bone-marrow stromal cells found increased oxidative stress and activation of the mitochondrial unfolded-protein response after exposure, followed by restoration toward baseline by 24 hours. That pattern is instructive. A transient response with recovery is not evidence of universal safety; it is a measurement of biological workload and buffering under one condition.

The decisive endpoint is therefore not only peak ROS. It is recovery cost:

• How much reserve was consumed?
• How long did normalization take?
• Was the next response altered?
• Did repeated exposures arrive before recovery completed?
• Did gene expression, chromatin, or organelle state retain a memory of the challenge?

Low-fidelity biology begins when perturbation outpaces recovery.

13. Spin: a candidate redox-sensitive branch

Some biochemical reactions proceed through radical pairs whose reaction outcomes can depend on electron-spin dynamics. Cryptochromes, flavin chemistry, heme systems, iron-sulfur clusters, and other redox-active structures provide possible molecular settings for magnetic sensitivity.

Research in magnetobiology has established that weak magnetic fields can affect particular radical-pair systems under defined chemical and kinetic conditions. Recent theoretical work has explored how tightly bound flavin–superoxide radical pairs might retain sensitivity at Earth-strength fields under specific recombination dynamics. Experiments in plants and animals continue to define the complexity of cryptochrome-associated magnetic responses.

The responsible conclusion is not that a complete human telecom-to-disease spin pathway has been proved. It is that radical-pair physics supplies a legitimate candidate transduction branch that cannot be evaluated by temperature rise.

Within S4–Mito–Spin, this branch may interact with the others:

• ion-channel timing changes mitochondrial redox state;
• mitochondrial electron transport produces spin-correlated radical intermediates;
• field-sensitive reaction probabilities alter ROS timing or localization;
• redox shifts feed back on channels, calcium handling, transcription, and chromatin.

This is a coupled network, not a three-step cartoon. Different tissues may be dominated by different branches.

14. Density gating: why the same field produces different outcomes

Uniform exposure does not imply uniform biological response. A tissue’s susceptibility depends on the density, coupling, and persistence of the machinery capable of transducing the field.

RF Safe’s 3+1 density-gating architecture organizes that heterogeneity:

Branch 1: S4 and excitable-membrane density

Tissues with high densities of voltage-gated channels, intense ionic throughput, narrow timing tolerances, or sustained electrical excitability may provide greater opportunities for a small gating perturbation to become a measurable physiological shift.

Branch 2: mitochondrial and organelle density

Cells with high ATP demand, dense mitochondria, tight calcium–mitochondrial coupling, limited energetic reserve, or high dependence on synchronized organelle networks may amplify a small ionic disturbance more strongly.

Branch 3: heme, flavin, iron-sulfur, and radical-pair density

Cells rich in redox-active cofactors and electron-transfer machinery may provide more candidate sites for spin-sensitive or redox-sensitive transduction.

The +1 persistence gate

Cell lifespan, turnover, developmental timing, repair capacity, antioxidant reserve, stem-cell status, immune clearance, and time between exposures determine whether the response disappears or becomes durable.

The architecture can be expressed conceptually as:

lasting biological impact ∝ external signal structure × receiver gain × intracellular amplification × persistence ÷ buffering and recovery

This is not a validated dose equation. It is a research map. Its purpose is to specify variables that average SAR omits.

Four downstream lenses on the same low-fidelity state

The following pathway maps use the same upstream trunk and then ask what a loss of timing fidelity would mean inside four different biological control systems. The proposed bridges are biologically grounded but remain research hypotheses. Their function is to organize experiments—not to convert mechanistic plausibility into a claim of individual diagnosis.

Development: when timing is the phenotype

 

Development is uniquely sensitive because sequence matters less than sequence plus timing. Neural progenitors must divide, migrate, differentiate, extend processes, prune connections, and establish network oscillations in the correct order. A signal that would be transient in mature tissue can redirect a trajectory when delivered during a narrow developmental window.

The developmental lens predicts that receiver density and state will vary across progenitor populations. S4/VGIC expression, mitochondrial maturation, CYB5B-dependent calcium behavior, gap-junctional coupling, and chromatin accessibility all change as cells differentiate. A field-induced timing error therefore need not produce one malformation. It may increase variance, delay a transition, alter the ratio of cell fates, or reduce the reserve available for a later challenge.

This is consistent with the cord-blood result: more differentiated cells showed a larger acute ROS response to the same UMTS exposure, while the tested system recovered without persistent damage. The signal was real; its persistence was state-dependent.

The critical developmental measurements are therefore not only gross anatomy or a later diagnosis. They include calcium-waveform fidelity, Vmem domains, progenitor division timing, migration, differentiation ratios, organoid geometry, synaptic maturation, network synchrony, recovery, and morphologic variance.

Cancer: when a rare error survives the control network

 

The cancer lens begins with Levin’s key observation: oncogenic genotype does not operate independently of bioelectric state. Depolarization marked tumor-like structures, and restoring a hyperpolarized state reduced their formation despite continued oncogene expression.

Within the low-fidelity model, cancer is not produced by a single RF switch. It becomes more probable when several protections lose precision together:

• oxidative DNA lesions increase repair demand;
• checkpoint timing becomes strained;
• chromatin state favors an adaptive but unstable program;
• apoptosis is delayed or escaped;
• immune clearance weakens;
• a long-lived cell retains the error;
• chronic exposure reduces the time available for complete recovery.

Density gating predicts that the most informative cells are not necessarily those absorbing the most average energy. They are cells combining excitable membranes, calcium–mitochondrial coupling, redox-active machinery, high functional demand, and persistence. Schwann and glial lineages deserve cell-resolved investigation for exactly that reason.

The NTP and Ramazzini findings anchor the outcome; the S4–Mito–Spin model proposes the missing path that should now be tested rather than assumed.

Autoimmunity: when self-versus-danger classification loses precision

 

Immune function is a classification problem implemented by living cells. The system must distinguish self from non-self, damage from ordinary turnover, pathogen-associated signals from sterile stress, and a threat requiring escalation from one requiring tolerance and resolution.

Calcium timing, membrane potential, mitochondrial state, NF-κB and AP-1 signaling, cytokine thresholds, metabolic switching, antigen presentation, regulatory T-cell function, and cell-death signals all participate in that classification.

The autoimmune lens predicts that bioelectrical dissonance can degrade this decision process without directly “causing autoimmunity.” Repeated sterile danger signals, mitochondrial distress, extracellular ATP, oxidized molecules, cell-free DNA, or mistimed calcium activation could increase classification errors in a genetically or developmentally susceptible system. If tolerance and resolution are already strained, a transient error can become self-reinforcing through epitope spreading, tissue injury, and further danger signaling.

The measurable endpoint is therefore not simply autoantibody presence. It is fidelity across the decision chain: calcium pulse pattern, activation threshold, cytokine timing, regulatory-cell balance, resolution kinetics, mitochondrial danger signals, and response to repeated challenge.

Metabolism: when coordinated pulses become chronic compensation

 

Metabolism is also a timing system. Insulin is pulsatile. Glucose transport, hepatic glucose output, mitochondrial oxidation, circadian feeding signals, adipokines, and muscle demand must be coordinated across organs.

The metabolic lens asks whether mistimed membrane and calcium signals can become mistimed fuel decisions. In pancreatic beta cells, voltage and calcium determine insulin secretion. In skeletal muscle, signaling controls glucose transport and mitochondrial demand. In liver, calcium and hormonal rhythms influence glucose output and lipid handling. In adipose tissue, storage, release, inflammation, and endocrine signaling must remain synchronized.

A low-fidelity state may initially look like successful compensation: more insulin, altered hepatic output, changed substrate use, or increased antioxidant response. If the timing error persists, the compensatory solution itself can become the next burden. Hyperinsulinemia, insulin resistance, ectopic lipid accumulation, and metabolic inflexibility are then understood as progressive loss of coordinated control rather than one broken molecule.

This branch makes a specific prediction: time-resolved measures should detect deteriorating phase relationships before fasting averages become abnormal. Experiments should therefore record pulsatility, glucose variability, mitochondrial reserve, tissue-specific signaling, circadian phase, and recovery after both metabolic and electromagnetic challenge.

The meta-disease synthesis

The four diagrams converge on one proposition:

Non-native EMF is not proposed as four separate disease-specific agents. It is proposed as one possible source of timing noise entering a shared cellular control layer. Tissue architecture and biological history determine where the resulting loss of fidelity becomes visible.

This explains why the outcome maps share S4, CYB5B, calcium timing, mitochondria, redox biology, chromatin, and persistence, yet diverge at the final stage.

It also explains why a successful experimental program must compare outcomes rather than studying each one in isolation. The strongest test is not merely whether one endpoint changes. It is whether the same waveform produces a predictable hierarchy across cell states based on receiver density, timing sensitivity, buffering, and persistence—and whether blocking the proposed upstream gate collapses several downstream responses at once.

15. Durdik’s cord-blood experiment separates response from persistence

Durdik and colleagues exposed human umbilical-cord-blood cells to pulsed GSM and UMTS signals and assessed oxidative, genetic, survival, and functional endpoints.

One hour of UMTS exposure produced a transient increase in reactive oxygen species. The magnitude of the response rose across cellular differentiation states, from less differentiated stem-cell fractions toward progenitors and differentiated lymphocytes. By three hours, the ROS signal had returned toward baseline. The investigators did not detect persistent DNA damage, preleukemic fusion genes, apoptosis, or impaired colony formation under the tested conditions.

This single experiment contains both an effect and a recovery boundary. The transient ROS response demonstrates biological interaction. The absence of lasting damage under those conditions demonstrates successful buffering.

Most importantly for density gating, acute response magnitude tracked differentiation state within the same human sample and exposure system. Cellular lifespan was not the explanation for the short-term gradient. The more differentiated cells possessed different metabolic, mitochondrial, and redox machinery.

That supports a separation between:

• transduction/amplification gain, which can change with differentiation and cellular machinery; and
• persistence, which determines whether the response is erased, retained, or converted into a later phenotype.

A safety program should measure both. Present RF limits measure neither.

16. CACNA1C shows that the receiver’s genotype matters

In 2025, Sousouri and colleagues conducted a randomized, double-blind, sham-controlled study in 34 healthy volunteers genotyped for the common CACNA1C variant rs7304986. Participants received 30-minute exposures at 3.6 GHz and 700 MHz before sleep, followed by high-density EEG.

The investigators found a genotype-by-exposure interaction. The 3.6 GHz exposure accelerated sleep-spindle center frequency in the T/C group, while the matched T/T group did not show the same response.

CACNA1C encodes the pore-forming subunit of the Cav1.2 L-type voltage-gated calcium channel. The studied variant is regulatory rather than a change to the amino-acid sequence of the channel. The finding therefore fits a receiver-density model: two groups can possess structurally similar channel proteins but differ in regulation, expression context, or network behavior sufficiently to change the physiological response.

An earlier observational study in 2,040 participants found that a different CACNA1C allele was associated with both worse subjective sleep quality and self-reported electromagnetic sensitivity. That association does not establish mechanism, but it independently points toward genetic heterogeneity in the population that should be tested rather than averaged away.

The implication for experimental design is immediate. If a response exists mainly in a genetically defined subgroup, an unstratified analysis can dilute it into a null average. The correct conclusion would not be that biology failed to respond. It would be that the study treated different receivers as if they were one receiver.

17. The persistence gate and long-lived target cells

The National Toxicology Program and Ramazzini Institute both reported malignant heart schwannomas in male rats. NTP also reported a brain-glioma signal. The convergence does not by itself prove the density-gating mechanism, but the affected cell lineages make persistence a serious variable to investigate.

Mature myelinating Schwann cells in adult peripheral nerves are highly stable, with little detectable turnover under normal conditions. Mature oligodendrocytes are also long-lived; human white-matter oligodendrocytes have been estimated to exchange at roughly 0.3 percent per year. Mature astrocytes likewise turn over far more slowly than intestinal, epidermal, or hematopoietic cells.

Long-lived cells retain both function and history. They have fewer opportunities to dilute accumulated molecular damage through replacement. If a tissue combines high membrane excitability, calcium–mitochondrial coupling, redox-active machinery, and low turnover, a transient perturbation has more opportunity to become a persistent state.

That gives the tumor findings a coherent research question:

Do heart Schwann cells and susceptible glial populations occupy a high-gain, high-persistence region of the S4–Mito–Spin landscape under chronic RF exposure?

The answer requires cell-type-resolved dosimetry, voltage imaging, calcium dynamics, mitochondrial phenotyping, redox mapping, mutation and chromatin analysis, lineage tracing, and long-term recovery studies. It cannot be inferred from whole-body temperature.

18. What low-fidelity biology means

High-fidelity biology is not a static, noiseless condition. Living systems are dynamic. They use controlled variability, stochastic gene expression, redox pulses, calcium oscillations, and adaptive stress responses. Fidelity means that this variability remains bounded, interpretable, and recoverable.

Low-fidelity biology is a state in which one or more of the following deteriorate:

• signal-to-noise ratio;
• timing precision;
• spatial localization;
• phase coordination;
• discrimination between signals;
• energy reserve;
• error detection;
• DNA and protein repair;
• appropriate transcriptional response;
• tissue-level consensus;
• termination of stress signaling;
• return to baseline.

The system may still look normal. A person may have no immediate symptom. A tissue may pass routine pathology. Yet its reserve can be lower, its recovery slower, and its response to the next challenge less reliable.

This is why low-fidelity biology is an upstream meta-disease state rather than a disease diagnosis. It changes the terrain on which disease emerges.

In that terrain:

• a mutation is more likely to escape correction;
• an inflammatory response is more likely to fail to resolve;
• a developmental transition is more likely to mistime;
• a protein-folding error is more likely to persist;
• a mitochondrial defect is more likely to cross an energetic threshold;
• a latent susceptibility is more likely to become phenotypic;
• a later-life failure can appear earlier;
• a rare event can become less rare.

The model is probabilistic because biology is buffered, heterogeneous, and history-dependent. That is not a weakness. It is why one exposure can produce different endpoints across tissues and individuals while still acting through a common upstream loss of fidelity.

19. Bioelectrical dissonance is a multi-hit model

RF Safe does not propose electromagnetic exposure as the sole explanation for cancer, developmental disorders, infertility, neurodegeneration, or metabolic disease. The model is explicitly multi-hit.

Air pollution can increase oxidative burden. Poor nutrition can reduce antioxidant and metabolic reserve. Sleep loss and mistimed light can disrupt circadian coordination. Infection can increase inflammatory and energetic demand. Chemical toxicants can interfere with enzymes, membranes, endocrine signals, and mitochondria. Genetic variants can alter receiver gain and repair. Age can reduce resilience. Medical interventions can create acute metabolic or immune demands that are handled differently depending on baseline reserve.

These inputs can converge on the same control layers:

• membrane excitability;
• calcium timing;
• mitochondrial function;
• redox balance;
• chromatin regulation;
• immune resolution;
• repair and recovery.

The low-fidelity model therefore rejects single-cause arguments for complex disease. A later event may reveal a decline that began earlier. A challenge may coincide with a visible regression without being the sole origin of the vulnerability. The correct research question is not merely “What happened immediately before the diagnosis?” It is “What reduced the system’s fidelity and recovery reserve across the relevant developmental window?”

This is where non-native EMF differs operationally from many other stressors.

People can often change food, water, sleep schedules, lighting, household products, or some medical choices. Those changes may be difficult or unequal, but the exposure sources are at least identifiable. Ambient RF can cross property lines, enter bedrooms and classrooms, originate from infrastructure outside personal control, vary with traffic, combine across devices, and persist day and night.

The result is not merely another isolated exposure. It can become the background timing environment in which every other challenge is processed.

20. Why repeated biological effects demand a unifying model

Dr. Henry Lai’s 2026 literature compilation reports significant effects in 390 of 438 RF oxidative/free-radical papers, 396 of 550 genetic-effects papers, 192 of 228 gene-expression papers, 396 of 507 neurological papers, and 354 of 415 reproduction/development papers. He also catalogued 260 low-intensity RF studies reporting effects below 0.4 W/kg.

These counts are a literature map, not a pooled risk estimate. They do not weight every design equally. Their importance is convergence. Independent investigators working across species, tissues, endpoints, frequencies, and exposure conditions repeatedly find that RF is biologically interactive below overt heating thresholds.

The large animal bioassays add consequence to that map:

• The U.S. National Toxicology Program found clear evidence of malignant heart schwannomas in male rats exposed to 900 MHz GSM- and CDMA-modulated RF, alongside evidence involving brain gliomas and DNA damage under its test conditions.
• The Ramazzini Institute study found an increase in the same rare heart tumor type in male rats after lifelong far-field 1.8 GHz GSM exposure at much lower whole-body absorption.
• A WHO-commissioned systematic review and its 2026 corrigendum rated the evidence for heart schwannomas and glial-cell tumors in male rats with high certainty.
• Melnick and Moskowitz used standard toxicological modeling to derive cancer-protective whole-body exposure estimates 15 to more than 900 times below the current public limit, depending on daily duration, and male-reproductive estimates 8 to 24 times below it.

The S4–Mito–Spin/low-fidelity framework does not claim to have proved the complete pathway behind each endpoint. It provides a disciplined way to connect the pattern:

1. A time-structured field encounters a heterogeneous biological receiver.
2. Membrane, redox, or spin-sensitive machinery transduces part of the signal.
3. Calcium and mitochondrial networks amplify or buffer the perturbation.
4. Cell type, genotype, differentiation, development, and prior stress determine gain.
5. Longevity, repair, turnover, and recovery determine persistence.
6. Tissue context determines the downstream phenotype.

This is why positive, negative, mixed, and beneficial findings can coexist. A therapeutic effect demonstrates controlled bioactivity. A transient adaptive response demonstrates successful buffering. A null result defines a boundary under one design. An adverse outcome indicates failed buffering or maladaptive persistence. None of those findings justifies pretending that average energy absorption captures the full biology.

21. TheraBionic exposes the regulatory contradiction

The FDA authorized TheraBionic P1 under the Humanitarian Device Exemption pathway for certain patients with advanced hepatocellular carcinoma. The device delivers specific amplitude-modulated RF electromagnetic fields through an intraoral antenna. FDA materials identify calcium-channel blockers as a contraindication.

This does not mean that a therapeutic signal and an environmental wireless signal are interchangeable. They differ in modulation, delivery, target, schedule, and biological purpose.

It proves a more basic point: non-heating, amplitude-modulated RF can be designed to produce a biologically relevant response, and that response can be sufficiently connected to calcium-channel biology to appear in federal device labeling.

The regulatory contradiction is therefore unavoidable.

When RF is medicine, agencies accept that frequency structure and biological receivers matter. When RF is environmental exposure, the public is told that average heat is the controlling fact.

The laws of biology do not change when the transmitter changes its regulatory category.

22. Why the thermal standard cannot protect fidelity

FCC and ICNIRP exposure limits are structured around avoiding established adverse effects associated with excessive energy deposition and tissue heating. Familiar metrics such as specific absorption rate and incident power density average energy across mass, area, and time.

Those metrics cannot answer whether an exposure changes:

• Vmem distributions across a tissue;
• voltage-gated channel timing;
• calcium frequency, phase, localization, or jitter;
• gap-junction conductance;
• mitochondrial membrane potential and ATP reserve;
• ROS and nitric-oxide oscillations;
• transcriptional and chromatin state;
• DNA-damage and repair kinetics;
• sleep-spindle dynamics;
• developmental trajectory;
• genotype-specific susceptibility;
• cumulative multi-source response;
• recovery between repeated exposures.

The 2021 D.C. Circuit decision made the administrative failure explicit. The court held that the FCC had not provided a reasoned explanation for retaining its limits in the face of evidence concerning non-cancer effects below the guidelines. It ordered the agency to address long-term exposure, children, testing procedures, wireless ubiquity and technological change, environmental effects, and the implications of pulsation and modulation.

The court did not decide the biological mechanism. It did something regulators could no longer avoid: it rejected unexplained reliance on the old assurance.

The scientific response should be a new protection objective:

An exposure limit is biologically protective only if it preserves signaling fidelity, recovery capacity, and developmental integrity under realistic chronic waveforms in susceptible receivers—not merely if it prevents measurable heating.

23. A direct experimental program for low-fidelity biology

The framework is useful only if it can fail. RF Safe therefore proposes a research program built around preregistered, waveform-resolved, mechanism-blocking experiments.

A. Characterize the real signal

Report carrier frequency, electric and magnetic field components, polarization, near- or far-field condition, modulation, envelope spectrum, pulse width, repetition rate, duty cycle, peak-to-average ratio, traffic pattern, spatial uniformity, harmonics, and exact exposure chronology.

Record the waveform at the biological sample, not only at the generator.

B. Measure the control variables directly

Use voltage-sensitive dyes and genetically encoded voltage indicators to map Vmem at single-cell and tissue scale. Measure gap-junction coupling, gating currents, calcium frequency and phase, propagation velocity, spatial localization, termination, refractory behavior, and recovery.

Average calcium concentration is not enough.

C. Test the S4 branch

Compare wild-type channels with altered S4 charge or expression. Use pharmacological blockade and genetic knockdown. Manipulate extracellular ion conditions. Test whether responses track field polarization, envelope frequency, and pulse structure as the forced-oscillation model predicts.

D. Test the CYB5B branch

Use CYB5B knockout, knockdown, and rescue. Measure whether telecom-relevant exposures reproduce rhythmic calcium responses in the relevant cell types. Map the dependence on mitochondria, redox state, differentiation, and exposure history.

E. Test the mitochondrial branch

Measure membrane potential, oxygen consumption, ATP reserve, ROS localization, NADH/FAD state, calcium uptake, mitochondrial dynamics, unfolded-protein response, mitophagy, and time to recovery.

Distinguish a successfully resolved stress response from a persistent state change.

F. Test the spin branch

Use isotope substitution, magnetic-field orientation, radical scavengers, controlled oxygen conditions, flavin or cryptochrome perturbation, and reaction-specific spectroscopy to determine whether spin chemistry contributes.

G. Stratify the receiver

Predefine genotype, sex, age, developmental stage, differentiation state, cell type, mitochondrial density, channel expression, redox baseline, circadian phase, and prior exposure. Do not average away subgroup effects.

H. Measure persistence

Follow the system after exposure. Measure minutes, hours, days, and—in regenerative or developmental models—subsequent challenge cycles. Test whether the phenotype reappears after injury even when current anatomy looks normal.

I. Map the genome’s physical response

Combine transcriptomics with ATAC-seq, Hi-C or Micro-C, methylation, histone marks, and nuclear imaging. Determine whether sustained bioelectric or redox perturbation changes chromatin architecture and whether those changes predict later response.

J. Use positive and negative controls that answer mechanism

Include sham exposure, thermal-matched controls, unmodulated carriers, envelope-only conditions where physically appropriate, alternate pulse structures at equal average power, known channel agonists and antagonists, mitochondrial stress controls, and recovery interventions.

The decisive comparison is not simply “field versus sham.” It is which signal feature, receiver, and pathway makes the response appear or disappear.

24. The planarian fidelity test

The planarian system provides an unusually direct test of the ceLLM attractor prediction.

Hypothesis

A transient gap-junction perturbation places regenerating G. dorotocephala into non-native, species-like morphological attractors of limited stability. If a defined pulsed field adds timing noise to the bioelectric coordination layer, it should alter the probability, variance, stability, or reversion kinetics of those induced morphologies.

Design

• Induce the established species-like head protocol under blinded conditions.
• Randomize fragments to sham, waveform-defined exposure, continuous-wave carrier at matched average power, and additional pulse/envelope conditions.
• Use magnetically characterized shielding and active sham controls appropriate to the tested frequencies. A conventional Faraday cage alone is not sufficient for low-frequency magnetic-field control.
• Map Vmem domains, gap-junction coupling, calcium dynamics, ROS, mitochondrial potential, and ultra-weak photon emission where technically validated.
• Score head morphology blindly and quantify variance, time to native remodeling, survival, growth, and repeat-regeneration outcomes.
• Collect matched tissue for RNA-seq, ATAC-seq, and chromosome-conformation analysis.

Predictions

The low-fidelity model predicts that waveform-defined exposure will not necessarily force one particular head shape. It may increase dispersion among outcomes, weaken maintenance of shallow attractors, change the rate of reversion, or alter the probability that a latent state persists into the next regeneration.

That is the signature of reduced fidelity: not a single deterministic abnormality, but a measurable loss of precision, stability, and recoverability in the collective decision.

25. The endpoints regulators should have been measuring

The next generation of exposure science should measure fidelity rather than waiting for one disease endpoint.

Priority endpoints include:

• Vmem mean, distribution, spatial boundaries, and temporal stability;
• calcium pulse frequency, amplitude, phase, localization, and jitter;
• channel gating-current variance;
• gap-junction communication and tissue-scale coherence;
• mitochondrial membrane potential, ATP reserve, and recovery time;
• ROS/RNS amplitude, localization, oscillation, and termination;
• DNA-damage formation and repair kinetics;
• chromatin accessibility and three-dimensional contact changes;
• transcriptional variance and error-correction response;
• immune activation and resolution;
• senescence, apoptosis, and clonal selection;
• developmental timing and morphologic variance;
• post-exposure recovery and response to a second hit;
• genotype- and cell-state-specific susceptibility.

Safety should mean that normal biological distributions remain stable, recovery remains intact, and vulnerable systems retain adequate margin. It should not mean only that the average temperature stayed low.

26. Source control comes before compensation

Biology can often buffer stress. Antioxidants, folate, sleep, nutrition, exercise, circadian alignment, and metabolic support may improve reserve under particular conditions. That does not make them substitutes for reducing an avoidable environmental input.

In one small rat study, folic acid reduced several kidney changes associated with 900 MHz exposure. The correct public-health lesson is not that people should supplement their way out of exposure. It is that biological reserve can change the outcome.

Compensation is not source control.

The hierarchy should be:

1. Reduce unnecessary exposure and timing noise.
2. Restore recovery windows, especially during sleep and development.
3. Use wired and optical communication where mobility is not required.
4. Support metabolic, circadian, nutritional, and environmental resilience.
5. Develop targeted therapies only after the upstream environment is addressed.

27. Why wired and Li-Fi infrastructure belongs in child protection

The purpose of this argument is not to reject communication technology. It is to move from biologically blind connectivity to biologically informed design.

Fiber and Ethernet can carry enormous data loads without filling indoor space with radiofrequency transmissions. Li-Fi can use modulated visible or infrared light for local wireless data where optical design is appropriate. IEEE 802.11bb-2023 has created a standards pathway for interoperable near-infrared light communication.

Li-Fi is not identical to ordinary ambient light, and any real system should be evaluated for flicker, optical intensity, circadian timing, driver electronics, and incidental RF emissions. But a well-designed optical link can replace a substantial portion of indoor microwave transmission while preserving mobility and bandwidth.

That is especially important in:

• nurseries and bedrooms;
• schools and childcare centers;
• neonatal and pediatric care;
• fertility clinics;
• hospitals and recovery environments;
• workplaces with fixed desks;
• public buildings where wired backhaul already exists.

The best wireless system is not the one that forces every bit through radio. It is the one that uses the least biologically intrusive medium appropriate to the task.

28. The policy program: protect biological fidelity

RF Safe calls for a Clean Ether Act built around a clear objective: reduce involuntary, chronic, close-proximity pulsed RF exposure where safer communication pathways can provide the same public benefit.

The program should:

• require HHS and FDA to restart an independent electronic-product-radiation research program under Public Law 90-602;
• require FCC exposure rules to address the D.C. Circuit remand before further deployment preemption is expanded;
• replace thermal-only assurance with biologically informed limits and uncertainty factors;
• require chronic, developmental, reproductive, waveform-specific, and multi-source testing;
• require disclosure of carrier, modulation, duty cycle, peak power, and realistic transmit behavior;
• fund independent replication with complete conflict-of-interest disclosure and open exposure data;
• create low-RF sleeping, learning, neonatal, pediatric, and reproductive-health environments;
• mandate wired networking and Li-Fi compatibility in public buildings and child-centered facilities;
• preserve wired telephone and broadband options;
• restore meaningful local authority over facility placement when health and environmental evidence is at issue;
• establish post-market biological surveillance and transparent adverse-event reporting;
• evaluate cumulative exposure across devices and infrastructure rather than authorizing each source in isolation.

This is not anti-technology. It is the ordinary maturation of technology: once a system becomes indispensable, its externalized biological costs must be measured and reduced.

29. The scientific proposition in one chain

The entire framework can be stated without mysticism and without a one-disease claim:

1. Living systems use membrane voltage, ion gradients, calcium rhythms, redox signals, mitochondrial state, gap junctions, and chromatin regulation to coordinate cell behavior.
2. Experiments show that manipulating bioelectric state can change regeneration, differentiation, gene expression, tumor-like phenotype, and future morphological outcomes.
3. Electromagnetic fields can produce non-thermal biological responses in defined experimental systems, including calcium-rhythm-dependent gene control and genotype-dependent human neurophysiology.
4. Modern wireless signals contain time structures that average power does not describe.
5. S4 voltage sensing, mitochondrial calcium–redox coupling, CYB5B, and spin-active chemistry provide specific candidate receivers and amplifiers.
6. Tissue density, cellular state, genotype, development, lifespan, repair, and recovery determine whether a response is amplified, buffered, or retained.
7. Repeated perturbation before full recovery can lower the fidelity of cellular computation and collective coordination.
8. That low-fidelity state changes susceptibility across many possible downstream diseases rather than producing one uniform diagnosis.
9. A thermal-only standard cannot establish protection against a control-system failure it does not measure.
10. Public policy must therefore reduce unnecessary exposure and require direct testing of biological fidelity.

Conclusion: preserve the signal life depends on

Becker showed that repair has an electrical dimension. Levin’s laboratory showed that membrane voltage can predict and redirect tumor-like behavior, coordinate effects at a distance, and alter the anatomy regenerated from an unchanged genome. Planaria show that invisible bioelectric state can outlast a visible perturbation and govern future morphology. Calcium biology shows that timing carries information. Mitochondria show how timing becomes energy, redox state, stress, and recovery. CYB5B and CACNA1C show that specific receivers and genetic context matter. Three-dimensional genomics shows that DNA is interpreted through physical regulatory architecture rather than read as a static list.

ceLLM integrates these facts by treating the cell as a local inference engine, chromatin as an evolved physical prior, and bioelectric state as part of the conditioning context. S4–Mito–Spin adds a testable receiver-and-amplifier architecture. Low-fidelity biology describes what happens when that integrated system loses precision faster than it can recover.

This is not a claim that RF alone determines cancer, autism, infertility, neurodegeneration, or any other complex condition. It is the more powerful proposition that an involuntary environmental signal can act upstream—altering the fidelity, reserve, and recovery of the biological systems through which many outcomes are decided.

The public should not have to wait for every downstream disease to be mapped before regulators protect the upstream control layer.

Life is electrical, rhythmic, metabolic, redox-sensitive, and state-dependent. The integrity of those signals is not optional. It is the condition that lets cells cooperate, repair errors, remember form, and return to health.

Invisible does not mean irrelevant. Non-ionizing does not mean non-biological. Legal exposure does not mean biological fidelity has been protected.

The standard measures heat.

Biology measures timing.

Public policy must now protect the signal life depends on.

Primary sources and supporting literature

Bioelectricity, regeneration, Vmem, and cancer-like phenotype

• Becker RO. The bioelectric factors in amphibian-limb regeneration. Journal of Bone and Joint Surgery. 1961. PubMed · DOI
• Becker RO. Search for evidence of axial current flow in peripheral nerves of salamander. Science. 1961. PubMed · DOI
• Becker RO, Murray DG. The electrical control system regulating fracture healing in amphibians. Clinical Orthopaedics and Related Research. 1970. PubMed
• Borgens RB, Vanable JW Jr, Jaffe LF. Bioelectricity and regeneration: large currents leave the stumps of regenerating newt limbs. Proceedings of the National Academy of Sciences. 1977. PubMed
• Borgens RB, Vanable JW Jr, Jaffe LF. Reduction of sodium dependent stump currents disturbs urodele limb regeneration. Journal of Experimental Zoology. 1979. PubMed
• Borgens RB, Vanable JW Jr, Jaffe LF. Bioelectricity and regeneration: initiation of frog limb regeneration by minute currents. Journal of Experimental Zoology. 1977. PubMed
• Adams DS, Masi A, Levin M. H+ pump-dependent changes in membrane voltage are an early mechanism necessary and sufficient to induce Xenopus tail regeneration. Development. 2007. PubMed
• Sundelacruz S, Levin M, Kaplan DL. Membrane potential controls adipogenic and osteogenic differentiation of mesenchymal stem cells. PLoS ONE. 2008. PubMed
• Chernet BT, Levin M. Transmembrane voltage potential is an essential cellular parameter for the detection and control of tumor development in a Xenopus model. Disease Models & Mechanisms. 2013;6:595–607. Full text · PubMed · DOI
• Chernet BT, Levin M. Endogenous voltage potentials and the microenvironment: bioelectric signals that reveal, induce and normalize cancer. Journal of Clinical & Experimental Oncology. 2013. PubMed
• Chernet BT, Levin M. Transmembrane voltage potential of somatic cells controls oncogene-mediated tumorigenesis at long-range. Oncotarget. 2014. PubMed · DOI
• Chernet BT et al. Long-range gap junctional signaling controls oncogene-mediated tumorigenesis in Xenopus laevis embryos. Frontiers in Physiology. 2015. PubMed · DOI
• Chernet BT et al. Optogenetic manipulation of membrane potential to normalize tumor-like structures: a potential strategy for cancer therapeutics. Oncotarget. 2016. PubMed · Full text · DOI

Planarian morphogenesis and bioelectric memory

• Emmons-Bell M et al. Gap junctional blockade stochastically induces different species-specific head anatomies in genetically wild-type Girardia dorotocephala flatworms. International Journal of Molecular Sciences. 2015. PubMed · DOI
• Durant F et al. Long-term, stochastic editing of regenerative anatomy via targeting endogenous bioelectric gradients. Biophysical Journal. 2017. PubMed · ScienceDirect · DOI

Calcium timing, mitochondria, and cellular state

• Dolmetsch RE et al. Signaling to the nucleus by an L-type calcium channel-calmodulin complex through the MAP kinase pathway. Science. 2001. PubMed
• Dolmetsch RE et al. Differential activation of transcription factors induced by Ca2+ response amplitude and duration. Nature. 1997. PubMed
• Dolmetsch RE et al. Calcium oscillations increase the efficiency and specificity of gene expression. Nature. 1998. PubMed
• Modesti L et al. Mitochondrial Ca2+ signaling in health, disease and therapy. Cells. 2021. PubMed
• Durdik M et al. Microwaves from mobile phone induce reactive oxygen species but not DNA damage, preleukemic fusion genes and apoptosis in hematopoietic stem/progenitor cells. Scientific Reports. 2019. Full text · PMC · DOI
• Kim J et al. Electromagnetic field-inducible in vivo gene switch for remote spatiotemporal control of gene expression. Cell. 2026. PubMed · DOI
• Sousouri G et al. 5G radio-frequency-electromagnetic-field effects on the human sleep electroencephalogram: A randomized controlled study in CACNA1C genotyped volunteers. NeuroImage. 2025;317:121340. PubMed · DOI
• Eicher C et al. Reduced subjective sleep quality in people rating themselves as electro-hypersensitive: An observational study. Sleep Medicine. 2024;113:165–171. PubMed · DOI

S4, modulation, and spin-sensitive chemistry

• Panagopoulos DJ et al. A comprehensive mechanism of biological and health effects of anthropogenic extremely low frequency and wireless communication electromagnetic fields. Frontiers in Public Health. 2025. Full text · DOI
• Ramsay J, Kattnig DR. Radical-pair-based magnetoreception amplified by radical scavenging: resilience to spin relaxation. Nature Communications. 2024. Full text
• Bradlaugh AA et al. Essential elements of radical pair magnetosensitivity in Drosophila. Nature. 2023. Full text
• Hammad M et al. Cryptochrome mediated magnetic sensitivity in Arabidopsis occurs independently of light-induced electron transfer to the flavin. Scientific Reports. 2020. Full text
• Xie W et al. 900 MHz radiofrequency field induces mitochondrial unfolded protein response in mouse bone marrow stem cells. Frontiers in Public Health. 2021. PubMed · DOI

Cell longevity and the persistence gate

• Stierli S et al. The regulation of the homeostasis and regeneration of peripheral nerve is distinct from the CNS and independent of a stem cell population. Development. 2018. PubMed
• Tripathi RB et al. Remarkable stability of myelinating oligodendrocytes in mice. Cell Reports. 2017. PubMed · DOI
• Yeung MSY et al. Dynamics of oligodendrocyte generation and myelination in the human brain. Cell. 2014. PubMed · DOI

Three-dimensional genome regulation and the ceLLM hypothesis

• Bonev B, Cavalli G. Organization and function of the 3D genome. Nature Reviews Genetics. 2016. Full text
• Krijger PHL, de Laat W. Regulation of disease-associated gene expression in the 3D genome. Nature Reviews Molecular Cell Biology. 2016. Full text
• Coates J. Topology-Gated Photonic-Redox Control Plane: ceLLM, S4 Flood, and the Geometric Hardware of Cellular Intelligence. RF Safe working manuscript. 2026. PDF

RF bioeffects, animal evidence, risk analysis, and regulation

• Lai H. Effects of exposure to electromagnetic fields: thirty years of research. Updated 2026. Compilation and downloads
• National Toxicology Program. Toxicology and carcinogenesis studies in Hsd Dawley SD rats exposed to whole-body radio frequency radiation at 900 MHz. Technical Report 595. NTP report · NIEHS overview
• Falcioni L et al. Report of final results regarding brain and heart tumors in Sprague-Dawley rats exposed from prenatal life until natural death to mobile phone radiofrequency field representative of a 1.8 GHz GSM base station environmental emission. Environmental Research. 2018. PubMed · DOI
• Mevissen M et al. The effects of radiofrequency electromagnetic fields exposure on cancer in laboratory animal studies: a systematic review. Environment International. 2025. ScienceDirect · DOI
• Mevissen M et al. Corrigendum to “Effects of radiofrequency electromagnetic field exposure on cancer in laboratory animal studies, a systematic review.” Environment International. 2026. PubMed · DOI
• Melnick RL, Moskowitz JM. Exposure limits to radiofrequency EMF do not account for cancer risk or reproductive toxicity assessed from data in experimental animals. Environmental Health. 2026. Full text · DOI
• U.S. Food and Drug Administration. TheraBionic P1—H220001. FDA device page
• Environmental Health Trust et al. v. Federal Communications Commission and United States of America. U.S. Court of Appeals for the D.C. Circuit, 2021. Decision
• IEEE Standards Association. IEEE 802.11bb-2023: Light Communications. IEEE standard description

Editorial note: This paper deliberately distinguishes replicated or directly observed findings from RF Safe’s integrative hypotheses. “ceLLM,” “S4–Mito–Spin,” “density gating,” “bioelectrical dissonance,” and “low-fidelity biology” are proposed frameworks intended to generate experiments. Their value will be determined by whether their predictions survive preregistered, blinded, independently replicated tests.