What your Garmin knows about your nights — and how SAYA reads it.

1 · Your nervous system, measured

Every evening, your body decides what kind of night this will be: a night of repair — or a night of vigilance. That decision is made by the autonomic nervous system (ANS), the involuntary control circuit with two branches: the sympathetic branch (fight-or-flight: mobilize, accelerate, stay alert) and the parasympathetic branch (rest-and-digest: repair, refuel, recover).

You can't feel which branch is winning. But your wearable can measure it — through two signals:

Heart rate (HR) is the throttle. When the sympathetic branch dominates, your heart beats faster — even at rest, even in bed.

Heart rate variability (HRV) is subtler and more interesting: it's the tiny variation in time between consecutive heartbeats. A heart under parasympathetic control doesn't tick like a metronome — it breathes with you, speeding up and slowing down flexibly. High HRV means your recovery branch is in charge. Suppressed HRV means your body is still braced for action.

Garmin watches record both around the clock. SAYA's job is to read these signals at the moments that matter most — the evening and the first hours of sleep — and turn them into decisions you can act on tomorrow evening.

2 · The evening resting heart rate — tonight's forecast

Most apps look at your resting heart rate in the morning, summarizing a night that's already over. SAYA adds a measurement that looks forward: the evening resting heart rate (eRHR) — how low your heart rate settles in the final hour before sleep.

Why it matters: elevated heart rate in the evening and night is consistently linked to worse sleep and worse health outcomes [1], and your body's pre-sleep state predicts the night remarkably well — in one 2025 study, pre-sleep heart-rate-variability measures predicted how fast people fell asleep and distinguished chronic insomnia with striking accuracy [2].

How SAYA measures it: from 70 minutes to 10 minutes before your usual sleep onset, SAYA takes your Garmin's heart-rate stream, applies a 5-minute rolling average, and records the lowest settled value. No button to press — it happens automatically, every evening. The value is then compared with your personal 7-day baseline, not a population table.

The best part: unlike your morning RHR, the evening value is actionable the same evening. Slow, paced breathing measurably raises vagal (parasympathetic) activity and lowers heart rate within minutes [3] — so a "running hot" reading isn't a verdict, it's a prompt.

3 · The HRV-Incline — the stress you take to bed

Here is a pattern almost nobody looks at, and it's SAYA's signature metric.

During healthy sleep, HRV rises across the first part of the night: as deep sleep takes over, the parasympathetic branch takes command and your beat-to-beat variability climbs toward a plateau [4]. That climb is recovery happening in real time.

The question is: where does the climb start?

If you go to bed calm, your HRV starts near its plateau — little climbing needed. If you carry stress into bed — a late interval session, a heated email, two glasses of wine — your HRV starts deep below the plateau. Your body spends the first hours of the night climbing out of a hole instead of repairing. We call the size of that hole the HRV-Incline: SAYA fits the initial slope of your night's HRV curve, finds the plateau, and computes the area of the deficit between them.

A small climb is normal and healthy. A large one is a fingerprint of pre-sleep stress: late intense exercise measurably suppresses early-night HRV and elevates heart rate [7], alcohol does the same dose-dependently [5], and evening stress fragments the night well beyond sleep onset [9].

4 · Chronotypes — measured, not guessed

Your chronotype is your body clock's preferred timing — when your biology wants to sleep, eat and perform. Fighting it has a name, social jetlag, and a cost: sleeping against your chronotype is associated with worse sleep quality and worse mood [11].

Most apps determine chronotype with a questionnaire. Questionnaires are decent — the short rMEQ correlates strongly with the full research instrument [10] — but they measure what you think about mornings. Your watch can measure what your body does.

How SAYA does it: chronobiology's standard yardstick is your mid-sleep point — the clock-time halfway through your sleep, corrected for sleep debt (the MCTQ method pioneered by Till Roenneberg's group). SAYA computes it from your actual recorded nights and places you on a seven-type spectrum from extreme early to extreme late — with an explicit confidence level that grows with your data (learning < 7 nights, high ≥ 14).

Once measured, the chronotype anchors everything else: your recommended bedtime window, your wake profile, and the timing advice below.

5 · The evening levers — what actually moves the needle

SAYA tracks your evening because the evening is where the research says the leverage is:

Alcohol is the bluntest lever. It sedates you to sleep, then runs your heart high and your HRV low all night — dose-dependently. In wearable studies, recovery drops roughly 9% after low doses and up to ~39% after heavy ones [5].

Late, heavy meals can disrupt your sleep: eating within ~3 hours of bed is associated with more night-time awakenings, and hard-to-digest meals make it worse [6]. Digestion competes with recovery — SAYA scores your dinner's timing, size and processing level for exactly this reason.

Late intense training is the athlete's classic own-goal. Finishing hard sessions within a few hours of bedtime delays sleep onset, dampens night HRV and raises night heart rate [7] — while the same session earlier in the day is harmless or beneficial. SAYA learns your personal cutoff from your own nights.

Caffeine lingers longer than it feels: even coffee six hours before bed measurably cost sleep in controlled research [8] — so an afternoon espresso may still tax your night.

Stress doesn't just delay sleep — it fragments it. Pre-sleep stress raises wake-after-sleep-onset and explains that "slept 8 hours, woke up tired" feeling [9]. It's also the single strongest pattern in our example athlete's data: evening stress correlated −0.59 with next-morning recovery across 56 nights.

6 · From data to decisions

Everything above condenses into a small set of numbers each morning:

Recovery Score (0–100). A weighted blend of five signals: your Garmin sleep score (10%), your HRV versus your personal 7-day baseline (40%), your evening resting heart rate versus baseline (10%), your HRV-Incline (20%), and your morning resting heart rate versus baseline (20%). If a signal is missing, weights redistribute — and if the required data isn't there, SAYA says so rather than inventing a number.

Personal baselines, not population tables. Every comparison is against your own rolling 7-day history, cleaned of outliers. A 44 bpm evening heart rate is exceptional for one person and alarming for another; only your baseline knows the difference.

Bayesian timing ideals. Your ideal dinner gap and workout cutoff start from research-based priors (about 3.5 and 3 hours before bed) and shift toward what your best nights show — weighting nights with strong sleep and a small HRV-Incline more heavily. Few data points? The prior dominates and SAYA tells you it's still learning. Lots of data? Your pattern wins.

Honest confidence. Chronotype, ideal duration and timings all carry explicit confidence levels based on how many valid nights back them. No fake precision.

SAYA is a wellness product, not a medical device. It doesn't diagnose or treat sleep disorders — if you suspect one, please talk to a professional.

Sources

1. Bottcher et al., 2024 — night-time heart rate, sleep and outcomes. PMC12073130
2. Tanaka et al., 2025, Frontiers in Physiology — pre-sleep HRV predicts sleep onset and quality. fphys.2025.1627287
3. Laborde et al. — slow-paced breathing and vagal activity. PMC8656666
4. Night-time RMSSD dynamics reliability study, 2022 — HRV rises across the first night-half during recovery. PMC9169847
5. Thakare et al., 2019 — alcohol and nocturnal HRV, dose-dependent. PMC7819834
6. Peixoto et al., 2020 — late eating and night-time awakenings. PMC7215804
7. Gao et al., 2025, Nature Communications — evening exercise timing, sleep and autonomic recovery. s41467-025-58271-x
8. Drake et al., 2013 — caffeine taken 6 hours before bed still disrupts sleep. PMC3805807
9. Lowe et al., Frontiers in Neuroscience — pre-sleep stress and sleep fragmentation. PMC10709081
10. Choi et al., 2023 — rMEQ validity for chronotype assessment. PMC11458378
11. Kim et al., 2020 — chronotype misalignment, sleep quality and mood. PMC7445814