Geomagnetic Storms and Bad Days: An Observation, a Hypothesis, and How to Test It on Yourself
2026-11-12Same post, two depths. Switch any time; the research version has the enzyme names, lab markers, and doses. Written from a research and education perspective: a model of what may be going on and what the published literature suggests would address it if the model is right, not medical advice or a treatment plan.
This post is different from the others on this site. It starts with something I noticed, not something the textbooks say, and it ends with a way to find out whether it is true for you rather than a list of things to take.
What I noticed
Among the families I have worked with, several have children on the autism spectrum who also have seizures, anxiety, or impulsive, hard-to-settle days. The parents keep careful logs, because that is what you do when you are trying to find a pattern. Over time I noticed that the worst days clustered, and that the clusters often lined up with something I had been following for other reasons: geomagnetic storms, the days when the Sun sends a burst of charged particles toward Earth and the planet's magnetic field shakes.
When I went back through the logs and laid them against the storm dates from NOAA's Space Weather Prediction Center, the pattern held in the families I checked. More seizures, more anxiety, more impulsivity on storm days and the day after. That is not proof. It is a small number of children, I was the one looking, and I knew what I was looking for. But it was consistent enough that I started telling parents to check the space weather forecast, and enough of them came back and said it matched that I think it is worth writing down and worth testing properly.
What a geomagnetic storm is
The Sun constantly streams charged particles outward (the solar wind). Sometimes it throws out a much larger burst, a coronal mass ejection. When that burst reaches Earth a day or three later, it pushes on the planet's magnetic field, which wobbles and rings. That wobble is a geomagnetic storm. It is measured on the Kp scale from 0 to 9; NOAA calls Kp 5 a minor (G1) storm and Kp 9 an extreme (G5) one. The aurora is the visible side of it. Storms come in clusters that follow the Sun's 11-year cycle, and we have just passed through the peak of the current cycle, which is why there have been so many auroras lately.
The field changes during a storm are tiny compared with a fridge magnet. That is the main reason scientists are skeptical that they affect people. But living things are sensitive to small magnetic signals in ways we are still learning about: birds, turtles, and salmon navigate by the field, and human cells contain tiny magnetic crystals (magnetite) and a light-sensitive protein (cryptochrome) that responds to magnetic fields in other animals.
What the research actually shows
Honest summary: there is a real but modest body of research, most of it finding small associations, some of it finding nothing, and almost none of it in the people I am describing.
- Melatonin. Several studies in electrical workers and in people at high latitudes found that on days of high geomagnetic activity, overnight melatonin production (measured in urine) was lower. Melatonin is the sleep hormone, and it also raises the seizure threshold and calms the stress system. If storms blunt melatonin, worse sleep and a lower seizure threshold the next day would follow.
- Heart rhythm and the autonomic nervous system. A large study of older men in Boston found that heart rate variability, a measure of how well the calming (vagal) side of the nervous system is working, dropped on days of higher geomagnetic activity. Lower HRV means a body that is more on edge. Other studies have found more heart attacks and strokes on storm days; some have not.
- Mood and behavior. A British study in the 1990s found more hospital admissions for depression in the week after storms. Russian and Eastern European groups have published for decades on storms and everything from blood pressure to accidents; the quality varies.
- Seizures. A handful of small studies report more seizures or seizure-related hospital visits on days of higher geomagnetic activity. Others find no link. Nobody has studied autistic children specifically.
So the state of the science is: plausible mechanisms (melatonin, the autonomic nervous system), small effects in some populations, mixed findings, and a gap exactly where my observation sits.
The hypothesis
If the storms matter for these children, I think it runs through two doors that are already open in them. First, melatonin and sleep: many autistic children already make less melatonin and sleep poorly, so a further dip pushes an already-low seizure threshold lower. Second, the autonomic nervous system: many of these children already run with low vagal tone and a jumpy stress response, the same picture described in the pots post here, so a day that lowers HRV further tips them into anxiety and impulsivity. Mast cells, which are sensitive to stress signals and are over-active in many of the same children, are a third candidate.
None of that requires the magnetic field to do anything dramatic. It only requires a small nudge to systems that are already close to their limit, which is the same logic as everything else on this site.
How to test it on yourself or your child
This is the part I can stand behind completely, because it costs nothing and settles the question for one person.
- Keep a one-line daily log for three months: a score from 0 to 3 for the symptom you care about (seizures, anxiety, meltdowns, sleep quality, pain, fatigue), plus sleep hours. Do not look at the space weather while you log, or you will see what you expect.
- At the end, pull the Kp history from NOAA's Space Weather Prediction Center (search "NOAA planetary K index" and download the daily values for your dates) and mark every day with Kp 5 or higher, plus the day after.
- Compare. Average symptom score on storm days and the day after versus all other days. If storm days are clearly worse across several storms, the pattern is real for you. If they are not, it is not, and you have lost nothing. Three months usually catches four to eight storm days during an active part of the solar cycle.
- If it holds, run it forward. Check the three-day forecast (NOAA publishes one) and write down your prediction before the day. A prediction that keeps coming true is the strongest evidence you can get for one person.
If you do this and want to share the result, the People's Data Commons on this site is built for exactly this kind of pooled observation.
If the model is right, what would help
Nothing here is a treatment for storms. It is the ordinary list of things that raise the seizure threshold and steady the autonomic nervous system, pulled forward onto the days when you know the nudge is coming.
- Protect sleep on storm nights. Dark room, consistent bedtime, no screens for an hour before, and, if a clinician already uses melatonin for the child, keep the dose on schedule.
- Magnesium at night. Magnesium is the mineral that calms the nervous system and raises the seizure threshold; magnesium glycinate or malate at bedtime is where most families already start.
- Lower the other loads that day. Fewer transitions, no sugar binge, no late night, no skipped meals (blood sugar dips lower the threshold too; see the 3am post here).
- Morning light. Twenty minutes outside in the morning anchors the melatonin rhythm for the following night.
- Keep the rescue plan close. For children with seizures, whatever the neurologist has prescribed for a cluster should be where it belongs on those days.
What I am not saying
I am not saying storms cause autism, seizures, or anxiety. I am saying that in a small number of children who already have those things, the bad days may cluster with space weather, that there are plausible reasons they might, and that a three-month log will tell you whether it is true in your house. If it turns out to be nothing, I will say so here.
Educational content written from a research perspective, not medical advice. I am a PhD scientist, not a physician, and I do not diagnose, treat, cure, or prevent disease. Seizure management is directed by a neurologist; nothing here replaces that plan.
The observation and its limits
In a small number of autistic children with comorbid epilepsy, anxiety, or impulsivity followed in a research consulting setting, parent-kept symptom logs retrospectively aligned with days of elevated planetary geomagnetic activity (Kp 5 or higher) and the following day. This is an uncontrolled, unblinded, retrospective observation by an observer aware of the hypothesis, in a convenience sample, and is reported here as a hypothesis-generating signal and nothing more. Its value is that it is cheaply testable.
Geomagnetic activity in brief
Coronal mass ejections and high-speed solar wind streams perturb the magnetosphere, producing geomagnetic storms quantified by the planetary K index (Kp, 0 to 9, three-hourly) and the Dst index. NOAA's G scale maps Kp 5 to G1 (minor) through Kp 9 to G5 (extreme). Ground-level field variation during storms is on the order of tens to hundreds of nanotesla against a static field of about 50,000 nT, with associated changes in the extremely low frequency (ELF) spectrum, including Schumann resonance amplitudes and geomagnetic pulsations (Pc1, 0.2 to 5 Hz, overlapping cardiac and neural rhythms). Storm frequency follows the roughly 11-year solar cycle; Solar Cycle 25 reached maximum in 2024 to 2025.
What the literature contains
Melatonin. Burch and colleagues reported reduced overnight urinary 6-sulfatoxymelatonin in electric utility workers on days of elevated geomagnetic activity (Neuroscience Letters 1999; further analyses 2008), with an interaction with occupational magnetic field exposure. Weydahl and colleagues found reduced melatonin at high latitude during disturbed periods (Biomedicine and Pharmacotherapy 2001). Effect sizes are modest and the populations specific; replication outside these groups is limited.
Autonomic function. In the Normative Aging Study, Zilli Vieira and colleagues (Environmental Health Perspectives 2019) found that higher geomagnetic activity in the preceding days was associated with reduced heart rate variability (SDNN, high-frequency power) in older men, independent of air pollution and weather; the same group later associated geomagnetic disturbance with cardiovascular mortality and blood pressure. Earlier work by Stoupel and others reported associations with myocardial infarction, stroke, and arrhythmia admissions, with inconsistent replication. A 2023 systematic review of geomagnetic activity and cardiovascular outcomes concluded that associations are reported more often than not but that heterogeneity and methodological quality limit inference.
Mood and behavior. Kay (British Journal of Psychiatry 1994) reported a 36 percent increase in male hospital admissions for depression in the second week after geomagnetic storms in a Scottish series. Associations with suicide and with traffic accidents have been reported, mainly from Eastern European and Russian literature of variable quality.
Seizures. A small number of studies report associations between geomagnetic activity and seizure frequency or seizure-related presentations (for example, in epilepsy monitoring unit cohorts and in Russian series), and others report null findings. There are no studies in autistic children. Publication bias in this literature is likely substantial.
Mechanistic candidates. Magnetite nanocrystals are present in human brain tissue (Kirschvink 1992; Maher 2016 distinguished biogenic from pollution-derived particles); cryptochrome proteins mediate radical-pair magnetoreception in birds and are expressed in human retina, though human magnetosensitivity is unproven; ELF electromagnetic fields alter calcium signaling and melatonin synthesis in some in vitro and animal models; pineal and suprachiasmatic function are plausible targets. A 2019 study (Wang and colleagues, eNeuro) reported alpha-band EEG changes in humans in response to rotating Earth-strength magnetic fields in a shielded chamber, suggesting an unconscious response to field direction. None of this establishes a pathway from storm to symptom; all of it makes the question non-trivial.
The hypothesis, stated so it can fail
In individuals with reduced melatonin output, low vagal tone, and a low seizure or anxiety threshold, days of elevated geomagnetic activity (Kp 5 or higher) and the following day are associated with a measurable increase in seizure frequency, anxiety, or dysregulated behavior, mediated by (a) reduced overnight melatonin and consequent sleep disruption and threshold lowering, and (b) reduced parasympathetic tone with sympathetic predominance, with (c) mast cell activation as a possible amplifier in the subset with that phenotype. The hypothesis predicts: effect size larger in those with pre-existing low HRV and poor sleep; attenuation when sleep is protected and melatonin is supplemented; no effect in individuals with normal thresholds.
This is consistent with the general model on this site, in which a small additional load on a system near its limit produces a disproportionate response (the shelf in my pots post here), and does not require a large biological effect of the field.
An n-of-1 design that families can run
- Prospective daily log, 90 days minimum. Primary outcome scored 0 to 3 (seizure count, anxiety, behavioral dysregulation, sleep quality, pain, or fatigue, chosen in advance); sleep duration; any obvious confounder (illness, travel, medication change). The logger stays blind to space weather for the logging period.
- Exposure. Daily maximum Kp from NOAA SWPC archives (GFZ Potsdam publishes the definitive index); exposure days defined a priori as Kp 5 or higher and the subsequent day (lag 0 and lag 1). A secondary analysis with Kp 4 captures sub-storm activity.
- Analysis. Compare mean outcome on exposure versus non-exposure days (Mann-Whitney or a simple permutation test); with enough storms, a case-crossover design using the same weekday in adjacent weeks as control controls for weekly structure. Report the number of storm days observed; fewer than four is uninformative.
- Prospective prediction phase. Using the NOAA 3-day forecast, record a prediction before each forecast storm; score hits and misses. Forward prediction removes the retrospective-fit problem entirely.
- Pooling. De-identified logs can be contributed to the People's Data Commons on this site for aggregate analysis, which is the only route from anecdote to estimate.
What would be reasonable if the hypothesis holds for a given person
Nothing specific to the field; the ordinary measures that raise threshold and vagal tone, concentrated on forecast days:
- Sleep protection (fixed schedule, darkness, screen avoidance) and continuity of any clinician-prescribed melatonin.
- Magnesium at night (glycinate or malate), given its NMDA antagonism and the consistent association of low magnesium with seizure susceptibility and anxiety.
- Avoidance of additional threshold-lowering loads: sleep loss, glycemic swings (the 3am post here), dehydration, illness, sensory overload.
- Morning light exposure to anchor circadian phase.
- Rescue medication availability per the treating neurologist's plan.
- In the mast cell subset, the stabilizing measures described in the histamine and sleep posts.
Genes worth noting
Melatonin synthesis and signaling (AANAT, ASMT, MTNR1A, MTNR1B), circadian clock genes (CLOCK, PER, CRY1 and CRY2, which encode cryptochromes), and autonomic and seizure-threshold loci (SCN1A, GABRG2, and the catecholamine clearance genes COMT and MAOA) define the phenotype the hypothesis targets. The sleep and circadian here report covers the first two groups; the neurotransmitters here report covers the last.
Position
The association literature is suggestive, methodologically uneven, and silent on this population. The mechanisms are plausible and unproven. The appropriate response is not a treatment; it is a prospective log, a forecast, and a prediction register, which any family can run, and which will either produce an effect size or retire the idea. This page will be revised in either direction.
Educational and research content, not medical advice. The author holds a PhD in cellular and molecular nutrition, is not a licensed physician, and does not diagnose, treat, cure, or prevent disease. Seizure and psychiatric management are directed by the treating clinicians.
