[ CASE FILE: THE NIGHT SHIFT ]
"You ate at ten. You slept at eleven. Your heart worked until two.
Falling asleep on a full stomach feels fine — pleasant, even. But sleep and digestion want opposite things from your nervous system, and asking for both at once means one of them runs at the wrong time. The meal wins. Your recovery loses a little. And the receipt shows up in the numbers you don't look at."
Section 1: Digestion is sympathetic work
At night your body is supposed to power down — resting heart rate drifts to its lowest, variability rises, core temperature drops. A large, late meal interrupts that descent. Digestion is metabolically expensive: blood flow redirects to the gut, core temperature ticks up from the thermic effect of food, and the heart works a little harder to support the circulation. That's mild sympathetic activation arriving exactly when your physiology is trying to hand the night over to parasympathetic recovery.
The bigger and richer the meal, and the closer to sleep, the longer the overlap.
Section 2: What the numbers actually show — honestly
Here is where most "eat early" content oversells. The effect is real but modest, and it varies by person. Tracker-scale data has put a late meal (within about three hours of bed) at roughly a 3% higher heart rate and a 7% lower HRV overnight. But the best-controlled studies are gentler: some crossover trials feeding people late find sleeping heart rate up by under a beat per minute, and at least one found overnight HRV essentially unchanged while morning cortisol rose instead.
So the honest headline isn't "late eating wrecks your sleep." It's: there's a small, genuine autonomic cost, its size depends on you, the meal, and the timing — and you can only know your version by measuring it. That personal variability is a feature of the problem, not a reason to ignore it.
Section 3: Why "3 hours before bed" is a starting point, not a law
The common rule — finish eating about three hours before bed, four if you're sensitive — is a reasonable default precisely because responses differ. A fast metabolizer with a modest dinner may see nothing. Someone eating a large, high-fat meal at 10 p.m. may see a clear bump in overnight heart rate and a flatter HRV. The rule points you at the right neighbourhood; your own data tells you the exact address.
Section 4: Reading your own late-meal receipt
This is a measurable input, so measure it. With a personal baseline — your normal resting heart rate, variability and breathing overnight — a late, heavy dinner shows up as a departure from your corridor: a higher sleeping pulse, a lower HRV than your normal. ONDA reads those overnight signals from an Apple Watch and holds the corridor, so you can compare an early-dinner night to a late one directly and see whether your body pays the tax and how much — your data, your call.
The firewall: this is a lifestyle input written into your own numbers, not a medical finding. ONDA doesn't diagnose reflux, metabolic conditions or anything else — it shows you how a choice moved your baseline.
Section 5: Practical, not dogmatic
You don't need a rigid cutoff; you need to know your own tolerance. Push your main meal earlier when you can, keep late meals lighter and lower-fat, and if you do eat late, a short exhale-led wind-down before bed nudges the nervous system back toward recovery. Then check the corridor — if your late nights read the same as your early ones, you've earned your flexibility honestly. For the deeper mechanics of an overnight fast, see intermittent fasting and the metabolic switch.
The Hack: Don't obey a blanket "no food after 8" rule — test it. Compare an early-dinner night against a late one in your own overnight heart rate and HRV. If your body pays the tax, you'll see it; if it doesn't, you'll stop worrying. Either way, the answer is in your data, not a chart for the average person.
MECHANISM: digestion = sympathetic load when pulse should fall EFFECT: overnight HR ↑ / HRV ↓ — real but modest and personal RULE: ~3h before bed is a default, not a law METHOD: compare your own nights — PRACTICE, NOT DIAGNOSIS[ SYSTEM_STATUS ]