How Space Weather Showed Up in My Wife's Sleep

For months, my wife kept appearing at my office door at 1 a.m. Then it was 2 a.m. Then again at 1 a.m. Same sentence every time for many nights over 2 months at least: I can't sleep.

I'd be at my desk with an Oura ring on, buried in astrophysics data. And after a while I noticed something I couldn't un-notice: her worst nights weren't random. They lined up with what I was working on — the days the sun was active, the days a solar storm was headed our way.

I should explain why that even registered. I've tracked my own sleep and heart rate since 2010 — not for an app, just because I'm the kind of person who wears a chest strap to bed and writes custom software to pull data my devices won't give me. By day, I build the pattern-detection systems that find faint signals buried in noisy space data. Finding patterns nobody else sees is, more or less, the job. So when one starts repeating in my own house, I notice — and because I'm an empiricist before anything else, I don't get to believe it or dismiss it until I've tested it.

So I started where it actually started: her sleep. I pulled her nights apart against the space weather and there was something there — but it was maddening to pin down. The index everyone reaches for, Kp, is one blunt, 3-hour planetary number built for power grids and satellites; held against a single human body it's like reading a heartbeat with a seismograph. The signal kept slipping through the gaps.

So I did the thing I always do when the instrument isn't good enough: I built a better one. I wrote software to break the space environment into the pieces that might actually reach a body, at the resolution a body runs on. And the signal sharpened — it wasn't just that she slept badly, it was in the architecture of the night: her sleep stages, her overnight heart rate, moving with what was happening overhead. Then I turned the same tools on myself. We were both sensitive — differently, but unmistakably.

Only then did I go looking to see whether anyone had ever studied this, half-expecting nothing. Instead I walked into a real, century-old field I'd never once heard named: heliobiology. The first modern papers I opened — a Harvard cohort showing heart-rate variability tracking geomagnetic activity, continuous monitoring showing the autonomic system responding even on the calm days — described, almost line for line, what I'd just dug out of my own house: my wife's nights, my own data. I hadn't found something new. I'd independently re-derived 100 years and hundreds of papers of work, and never knew it existed.

That's what made me read all of it instead of poking at it after hours — back to the man who founded the field in 1920s Russia and got 8 years in a gulag for arguing the sun moved human affairs. And the deeper I went, the more serious it got, until I reached the heart: a body of modern cardiology, replicated across decades and continents, with a worked-out physiological mechanism rather than a bare correlation, showing the cardiovascular system itself responding to space weather. Sleep is one thing. The heart is the most consequential system in the body — and there it was, in peer-reviewed cardiology, tracking the sun.

By then I wasn't wondering whether it was real — I knew. What I didn't have was the whole picture. My wife and I were 2 people, and 2 people is a story, not proof — I don't trust an n of 2, even my own. So I went and got the rest of it myself.

Roughly 85 million raw measurements across 6 independent, publicly available datasets using wearables like Fitbit to HRV watches. I initially ran 3 separate causal-inference methods — Granger causality, ARIMAX regression (validated with Ljung-Box), and superposed-epoch analysis. All corrected for testing thousands of driver-and-biometric pairings at once, and trusted only what survived my testing. The formal write-up and white paper is the thing that keeps growing on my desk — but the signals held across all 6 datasets from all methods I used for over 60 pairings.

And here's the part that genuinely stunned me — and the part I would have never dreamed of showing up. In the data, the body's response tracked the earliest drivers in the chain — the upstream solar-wind conditions that themselves run about a week ahead of the geomagnetic storm everyone watches for. So the physiological shift can show up days before the storm "arrives" — not because the body knows the future, but because it's responding to the front edge of the same chain, while the headline event lags behind. How I visualized it in my head was akin to lightning, then thunder. It's a real pattern, not a prediction.

And here's the structure that surprised me most from all of this testing — and it's the opposite of random. Everyone is sensitive. In the data, essentially every body responded across the same broad set of relationships — more than 60 distinct pairings of a body signal and a space-weather driver — in sleep, in heart rhythm, and beyond. What separates one person from the next isn't whether they're affected — it's how hard each force hits, and how long it takes to land. For one person a geomagnetic storm is a freight train 2 days later; for another the same storm barely registers, while a Forbush decrease — the drop in cosmic rays that rides in with a solar storm — is what takes the wind out of them a week later. Same forces, different gain, different timing — and most of those 60+ pairings, for any given person, stay in the data with known patterns. Your profile is as much a map of what to ignore as what to watch for: it tells you exactly which forces move you, and which ones do nothing and even when that occurs. Which is the whole reason a single number on a screen can't help you alone — so I also built a forecast that uses anyone's personal sensitivity to use and detect these patterns. If the driver that hits you hardest is building a few days out and hits a peak with a recovery time of a few days; we can see it and tell you the timing.

Heliobios pairs your wearable with live space weather and learns your profile — which conditions move which of your systems, how strongly, and roughly how much lead time your body tends to show, so you can understand your patterns and plan around them. If you live with migraines, chronic pain, fatigue, or sleep that falls apart for no reason anyone can name, it does the thing you've probably tried to build by hand and couldn't: it lays your rough days over the space-weather days, side by side, until the pattern you've felt for years finally has a picture. And you can log the days that matter — a migraine, a flare, a wrecked night — so it learns your symptoms against the space weather, not just what your wearable can see. Once it knows which conditions tend to run ahead of your bad days, it forecasts those conditions and gives you the heads-up days out — while there's still time to prepare. A population average is useless to you personally. Your own data isn't.

Two lines I won't cross: your raw biometrics never leave your phone — all the scoring happens on-device. And it's a wellness and educational tool, not a medical device — it doesn't diagnose, treat, or predict any health condition. It forecasts the space weather, surfaces your patterns, and helps you plan around them.

I still think about those nights my wife came into my office late, telling myself it was maybe nothing. Space weather was one of the variables the whole time. I just finally built the instrument to see it — and to do something about it.

Heliobios is in invite-only iOS beta. If you've had a run of days nothing in your routine could explain — or you already know you react to geomagnetic storms or CMEs — come find out whether the sun is one of the things moving you.