At the exposure levels people encounter from commercial 5G networks, no confirmed harmful effects on the human brain have been established in controlled studies. Large reviews of the research, including one covering radio-frequency fields above 6 GHz, have found no evidence that low-level 5G signals are hazardous to human health. That said, the picture is not a blank slate. A growing body of lab and animal research has turned up subtle biological responses, from shifts in barrier permeability in mouse brains to altered neurite growth in cell cultures, that keep the scientific conversation open even as headline safety standards appear to hold.
How 5G Radio Waves Interact With the Head
5G networks operate across a wide range of frequencies. The lower bands, around 700 MHz to 3.5 GHz, behave much like previous-generation cell signals and can penetrate tissue to a modest depth. The higher millimeter-wave bands, roughly 24 GHz and above, are different. These waves have very short penetration depths and are largely absorbed by the skin before reaching deeper structures. Because of that rapid absorption, researchers have stressed the need to investigate potential effects not just in terms of tissue heating but at the cellular level in the skin itself.
Modeling studies that simulate what happens when a 5G phone is held against the head have confirmed that energy absorption drops off steeply as you move inward from the scalp. At millimeter-wave frequencies used in 5G, the maximum specific absorption rate occurs in the outermost tissues and attenuates rapidly toward deeper brain regions.
A recent study using a multi-layer human head model tested 5G phone signals at two power levels and found that peak energy absorption in the brain layer stayed far below international safety limits. After half an hour of simulated exposure, the maximum brain temperature reached only about 37.2 °C, well under the threshold for thermal damage. Moving the antenna from 5 mm to 30 mm away from the head cut energy absorption by roughly half, though temperature changes remained negligible. The takeaway from thermal modeling is fairly consistent: under normal use, 5G phones do not heat the brain to dangerous levels.
Blood-Brain Barrier Findings in Animals
One of the more attention-grabbing lines of research involves the blood-brain barrier, the tightly sealed lining of blood vessels in the brain that keeps most circulating substances out of neural tissue. A Chinese study exposed mice to 5G frequencies of 3.5 GHz and 4.9 GHz for one hour a day over 35 days. Compared with unexposed mice, the exposed animals showed increased barrier permeability in the cerebral cortex. The researchers found that levels of occludin, a protein that helps keep the barrier sealed, were reduced in the exposed groups, with the 4.9 GHz group showing a more pronounced drop. Other tight-junction proteins they measured did not change.
This result is concerning on its face, but context matters. The study involved a specific protocol in mice, not humans, and the exposure parameters do not perfectly replicate how a person uses a phone. Separate research using ultra-wideband electromagnetic pulses at extremely high field strengths, far beyond anything a consumer device produces, found that barrier opening in rats was field-strength dependent and reversed within 24 hours. That work pointed to a decrease in ZO-1, a different tight-junction protein, as part of the mechanism. The reversibility is an important detail: even at these extreme exposures, the barrier appeared to repair itself.
No human study has directly demonstrated 5G-induced blood-brain barrier opening. The animal findings raise a biological question worth pursuing, but they have not been translated into evidence of real-world harm in people.
Cellular and Molecular Changes in Lab Settings
When researchers expose neurons in a dish to radio-frequency fields, they sometimes see changes at the molecular level that do not show up at the whole-organism level. One lab study found that 48 hours of 1,800 MHz exposure reduced neurite outgrowth in both primary mouse hippocampal neurons and a neuronal cell line. Neurites are the branching projections that neurons extend to communicate with each other, and the exposed cells grew shorter ones with fewer branches. The researchers traced this to a decrease in a signaling molecule called Rap1-GTP, and when they artificially boosted Rap1 levels, the neurite growth impairment reversed.
A related study on embryonic neural stem cells exposed to the same frequency at 4 W/kg found that several genes involved in neuronal development were disrupted. Genes that promote neuron formation were turned down while an inhibitory gene was turned up, suggesting a shift in the molecular balance that controls how stem cells mature into neurons.
These cellular findings sound alarming, but they come with major caveats. The exposure intensity of 4 W/kg is twice the international safety limit for the general public. Lab conditions eliminate the shielding that skin, skull, and cerebrospinal fluid provide. And the fact that cells in a dish respond to a stimulus does not mean the intact brain responds the same way.
On the other side of the ledger, a well-controlled study that exposed astrocytic and neuronal cell models to 5G-modulated 700 MHz signals at both low and high absorption rates under carefully maintained temperature conditions found no measurable oxidative stress, cell death, or changes in cell growth. Another study testing combined radio-frequency radiation on three types of neuronal cells found no consistent effect on levels of reactive oxygen species, even when the cells were simultaneously stressed with chemical oxidizers. The cellular picture, in short, is mixed, with negative results often outnumbering positive ones when temperatures are carefully controlled.
Animal Behavior and Memory
Rodent studies that test whether radio-frequency exposure impairs learning or memory have produced inconsistent results. One study delivered near-field radiation at absorption rates between roughly 0.5 and 1 W/kg to rats for two hours daily over a month. In water-maze testing, the exposed rats performed worse on spatial learning and reference memory compared with unexposed controls, and the researchers observed structural changes in the hippocampus, the brain region most involved in forming spatial maps.
A study specifically using the 4.9 GHz 5G band in mice found a different pattern. Spatial memory and anxiety-related behavior were unaffected by the exposure, but the mice did show increased depression-like behavior. The dissociation is interesting because it suggests that different brain circuits may respond differently, or that behavioral outcomes depend heavily on the specific frequency, duration, and intensity of exposure.
A study in juvenile rats exposed to radio-frequency fields found changes in thermal preference and sleep patterns. The exposed pups preferred slightly warmer sleeping environments and spent about 15 percent more time asleep, driven by an increase in slow-wave sleep. Paradoxical sleep, the rodent equivalent of REM sleep, was not affected. Whether this translates to meaningful sleep disruption in humans is unknown, but it hints that developing brains might respond differently from adult ones.
What Human Studies Show
Direct evidence from human experiments is limited but mostly reassuring so far. A triple-blind crossover study exposed volunteers to 26 GHz 5G signals, one of the higher millimeter-wave bands, for about 26 minutes at a field strength within regulatory limits. The researchers recorded brain electrical activity across all major frequency bands and found no detectable changes in any of them. The study’s authors described it as the first controlled investigation of 26 GHz 5G effects on human brain activity.
At lower frequencies more representative of earlier mobile technology, a controlled study testing a generic 3G-type signal at roughly 2 GHz found no effect on attention or reaction time in human subjects under short-term exposure. A literature review examining mobile phone exposure and cognitive function in children and adolescents found a more complicated picture. Data from a large Australian study of adolescents reported that more frequent phone calls were associated with poorer accuracy on working memory and memory tasks, though reaction times were actually faster. The researchers suggested this pattern could reflect an impulsive response style developed through phone use rather than a direct radiation effect on the brain, since phone habits and screen time are tangled up with how teenagers use their attention.
Separating the biological effects of radio-frequency energy from the behavioral effects of phone use itself is one of the hardest challenges in this field. A teenager who makes many phone calls also spends more time multitasking and less time on sustained-focus activities, which could produce the same cognitive profile without any radiation effect at all.
Brain Tumor Risk in Long-Term Phone Users
The question that generates the most public anxiety is whether cell phone use causes brain cancer. The largest and most methodologically rigorous data now available comes from the COSMOS prospective cohort study, which followed over 264,000 participants for a median of about seven years. The study found no increased risk of glioma, meningioma, or acoustic neuroma with increasing cumulative hours of phone call time. Even among the heaviest users, those in the top ten percent of call time, the risk estimate for glioma was statistically indistinguishable from baseline. More than 15 years of phone use was not associated with elevated tumor risk either.
A separate meta-analysis pooling results from five cohort studies reached a similar conclusion overall, finding statistically equivocal results for central nervous system tumors, glioma, and meningioma. One exception was acoustic neuroma, a benign tumor of the nerve connecting the ear to the brain, where the pooled risk ratio was somewhat elevated for long-term users, though the result was not robust enough to be conclusive.
It is worth noting that nearly all the epidemiological data covers 2G, 3G, and 4G frequencies because 5G networks simply have not been widely deployed long enough for cancer studies, which need decades of follow-up, to be completed. The lower-band 5G frequencies overlap substantially with 4G, so the existing data is reasonably relevant for those bands. For the higher millimeter-wave bands, where the physics of tissue penetration is quite different, no long-term human cancer data exists yet.
Neurotransmitter Shifts in Animal Models
A handful of animal studies have looked at whether electromagnetic field exposure shifts the brain’s chemical messaging systems. One older but frequently cited experiment found that after a month of exposure, male rats showed elevated serotonin in the hypothalamus, while both male and female rats had increased levels of a dopamine breakdown product in the striatum. These changes were described as small, and they faded when the exposure period was extended, suggesting the brain may adapt over time.
A more recent review paper cataloging studies on radio-frequency radiation and neurotransmitters noted that research in this area has been growing, but the findings are scattered across different frequencies, exposure durations, and animal models, making it difficult to build a coherent narrative. The field has not converged on a consistent pattern of neurotransmitter disruption at consumer-level exposures.
Electromagnetic Hypersensitivity and the Nocebo Effect
Some people report headaches, fatigue, difficulty concentrating, and other symptoms that they attribute to proximity to wireless devices or cell towers. This condition, sometimes called electromagnetic hypersensitivity, has been studied extensively through provocation experiments where volunteers are exposed to real or sham signals without knowing which is which.
The results are strikingly consistent. A systematic review examining 31 such experiments, testing 725 self-identified hypersensitive participants, found that 24 showed no evidence that participants could detect or react to real electromagnetic fields. Of the seven that initially reported some positive findings, two could not be replicated by the same research groups, three appeared to be statistical artifacts, and the remaining two gave contradictory results. An updated review expanded the dataset to 46 blinded provocation studies involving 1,175 volunteers and again found no robust evidence that electromagnetic fields triggered symptoms. The studies did, however, find support for the nocebo effect: the belief that one is being exposed was enough to produce real symptoms, regardless of whether the signal was actually on.
The symptoms these individuals experience are genuine, but the evidence consistently points away from radio-frequency fields as the cause and toward expectation and anxiety. This distinction matters practically because it means symptom management should focus on the psychological and environmental factors that drive the experience rather than on reducing exposure to signals the body does not appear to detect.
Why the Research Is Hard to Interpret
One reason the 5G-and-brain literature feels so inconclusive is that the quality of exposure measurement varies wildly between studies. A review spanning three quarters of a century of radio-frequency dosimetry research noted that many bioeffects studies still fall short on dosimetric rigor, and called for improved techniques especially at the higher microwave frequencies 5G uses. If two labs expose cells to “the same” frequency but measure absorbed energy differently, or fail to control temperature precisely, their results may diverge for reasons that have nothing to do with biology.
Temperature control turns out to be a critical dividing line. Many of the positive findings in cell studies come from experiments where heating was not strictly accounted for. When researchers have gone to extra lengths to maintain constant temperature throughout exposure, the biological effects tend to disappear. The 5G-modulated 700 MHz study that found no oxidative stress, apoptosis, or proliferation changes in neuronal and glial cells specifically emphasized its isothermal experimental conditions. That detail alone could explain a large share of the discrepancies in the literature.
Animal studies face a different problem: ecological validity. Exposing a mouse’s entire body to a radio-frequency field for hours a day is not a realistic analog for a person holding a phone to their ear for a few minutes. The absorption patterns are different, the ratio of body size to wavelength is different, and the cumulative dose lands on different tissues. Negative findings in animals should therefore not be taken as proof that 5G harms brains, and positive findings should not be taken as proof that it harms human brains specifically.
The WHO has commissioned 12 systematic reviews covering a range of health outcomes related to radio-frequency fields, including cognitive impairment and cancer. A critical commentary on those reviews argued that methodological limitations in the underlying studies mean the reviews themselves cannot offer strong assurance of safety. Whether you read that as a call for concern or simply as a statement about the current limits of the evidence depends on your tolerance for uncertainty, but it underscores that “no confirmed harm” and “proven safe” are not the same claim.
A Surprising Lead in Alzheimer’s Research
Not all the biological effects of radio-frequency exposure point in a worrying direction. A study using a transgenic mouse model of Alzheimer’s disease found that long-term radio-frequency electromagnetic field exposure actually reduced cognitive decline in the affected mice. The mechanism appeared to involve microglia, the brain’s resident immune cells. In the exposed Alzheimer’s mice, markers of microglial activation and proliferation in the hippocampus were significantly reduced compared with unexposed Alzheimer’s mice. Since overactive microglia contribute to the neuroinflammation that drives Alzheimer’s progression, dampening their activity could plausibly slow the disease.
This is a single animal study and far from a therapeutic proposal for humans. But it illustrates something about the 5G-and-brain conversation that often gets lost: biological effects are not synonymous with harmful effects. A stimulus that modulates one cellular process could be damaging in one context and protective in another. The Alzheimer’s finding is a reminder that the question worth asking is not just “does 5G do anything to the brain?” but “what does it do, under what conditions, and does it matter clinically?” For the moment, the honest answer to all three parts is that we are still working on it.