How do wearables measure stress?

Most wearables measure stress by tracking heart rate variability (HRV) and heart rate. When stress or arousal increases, parasympathetic nervous system activity drops, HRV decreases, and heart rate rises. Some devices add additional signals: Samsung, Google Pixel Watch, and Fitbit Sense 2 use electrodermal activity (EDA) to detect sympathetic activation through skin conductance. WHOOP adds respiratory rate and SpO2 using log-transformed RMSSD (lnRMSSD). However, wearables detect physiological arousal, not stress specifically. They cannot distinguish anxiety from excitement, caffeine, or exercise.

Can you compare stress scores between Garmin, WHOOP, and Oura?

No. Stress scores are not comparable across brands. Garmin uses a 0 to 100 scale, WHOOP uses 0 to 3 for stress and 0 to 100% for recovery, Oura uses Low/Medium/High labels, and Polar uses -10 to +10. Each device uses different signals, different algorithms, and calibrates against a different personal baseline window (14-day for WHOOP, first-month for Pixel Watch/Fitbit, 14-day for Oura Resilience, 21-31 days for Oura Cumulative Stress, 28-day for Polar). Track trends within one device over time rather than comparing numbers across brands.

Which wearables use EDA for stress tracking?

Three wearable brands use electrodermal activity (EDA) in their stress calculations. Samsung Galaxy Watch uses spot-check EDA via its BioActive Sensor. Google Pixel Watch 2/3 and Fitbit Sense 2 use continuous EDA (cEDA), the first all-day on-wrist implementation launched in 2022. The original Fitbit Sense (2020) was the first consumer wearable with wrist EDA but only as a 2-minute manual spot-check. Fitbit Charge 5 and Charge 6 have 90-second spot-check EDA. Fitbit Versa 4 and Inspire 3 have no EDA.

Does the Apple Watch have a stress score?

No. As of watchOS 11, Apple Watch does not have a native stress score. It tracks HRV (as SDNN, not RMSSD like other wearables), heart rate, respiratory rate, ECG, and blood oxygen. HRV is sampled approximately every 4 hours by default, increasing to every 2 hours with Irregular Rhythm Notifications and every 15 minutes with AFib History enabled. Third-party apps like Athlytic and HRV4Training can compute RMSSD from raw R-R intervals in HealthKit, while Welltory uses frequency-domain analysis.

What factors affect wearable stress scores the most?

The biggest factors are consistent sleep (7 to 9 hours can improve HRV by 15 to 30% within 4 weeks), alcohol (even 1 drink suppresses HRV; 3+ drinks can drop RMSSD by up to 13ms for 2 to 5 days), regular aerobic exercise (150 min/week significantly improves HRV long-term), chronic psychological stress (sustained sympathetic activation reduces HRV), and illness or inflammation. Caffeine affects some individuals more than others, with sensitive people seeing an 8 to 12% HRV drop.

Why does my wearable show high stress when I feel fine?

Wearables detect physiological arousal, not perceived stress. Your score can spike from caffeine, dehydration, a hot environment, excitement, digestive activity after a large meal, or residual effects from last night's alcohol. Research published in Nature Communications Medicine (2025) confirmed that self-reported stress often does not correlate with physiological markers like cortisol or HRV. Your body may be under autonomic load even when you feel calm.

What is the most accurate wearable for measuring stress?

No single wearable has been validated as the most accurate for real-world stress detection. Research shows multi-signal devices (those using HRV + EDA + skin temperature) outperform single-signal devices in controlled settings, with roughly 82% vs. 77% accuracy. This gives Google Pixel Watch, Fitbit Sense 2, and Samsung Galaxy Watch a theoretical edge over HRV-only devices. Device-specific validation shows WHOOP 4.0 lnRMSSD variability falls within reference protocol range for HRV, while Apple Watch Series 9 underestimates SDNN by a mean of 8.31 ms compared to chest strap reference. Lab accuracy does not transfer cleanly to daily life.

What is Garmin Body Battery and how does it relate to stress?

Garmin Body Battery is a 0 to 100 energy metric that combines stress score, sleep, activity, and recovery data into an overall energy tank. It drains with exertion and stress and recharges with rest. It is not a separate measurement but a derived metric built on top of the stress score. Garmin also offers a separate 3-minute chest-strap HRV Stress Test on Fenix and high-end Forerunner models, which uses ECG-derived R-R intervals from a Garmin chest HRM for more precise training-readiness decisions than wrist-based PPG.

Wearable Stress Scores Compared: Garmin vs WHOOP vs Oura vs Samsung vs Apple Watch (2026)

Q: How accurate are wearable stress scores?

A 2024 JMIR meta-analysis found that multi-signal wearable approaches (HRV + EDA + skin temperature) achieve roughly 82% accuracy, while single-signal HRV-only approaches reach about 77%. However, these numbers come from controlled lab settings. Real-world accuracy is consistently lower because models trained in lab stress paradigms lose meaningful accuracy in daily life. A 2025 arXiv study confirmed that reproducing lab-grade stress detection on consumer wearables remains difficult. Cross-user generalization is also poor, meaning models work best when personalized to an individual.

Ryan - Kygo Health
May 7
9 min read

Updated: May 14

Last Updated: May 14, 2026

Five wearable devices: three smartwatches, one fitness ring, and a blue fitness band on a black background. Representing Kygo Health's blog on how wearable brands track stress.

Your Garmin says 72. Your friend's WHOOP says 1.8. Your partner's Oura says "High." All three claim to measure stress. None of them agree, and none of them are wrong.

That's because wearable stress scores aren't measuring stress. They're measuring physiological arousal: shifts in your autonomic nervous system that happen when your body reacts to something. The trigger could be anxiety, excitement, caffeine, a cold shower, or digestion after a big meal. No consumer device can tell the difference. A 2024 JMIR meta-analysis found multi-signal devices (HRV + EDA + skin temperature) reach roughly 82% accuracy detecting arousal in controlled lab settings. Single-signal HRV-only devices hit about 77%. Real-world accuracy is consistently lower for both.

So what separates these devices? The sensors they use, how they score you, and what actually moves the number. We broke down all 10 and built an interactive Stress Factor Explorer to go deeper on every signal and factor by device.

Which Device Measures What

Not every wearable uses the same sensors or even offers the same features. This is the comparison that matters most.

Device	Daytime Stress	Recovery Score	EDA Sensor	Skin Temp	SpO2 in Stress	Resp Rate in Stress	Exercise vs. Stress	Membership Required
Garmin	Yes	-	-	-	-	-	-(pauses)	-
Samsung Galaxy Watch	Yes	-	Yes	-	-	-	-	-
Google Pixel Watch	Yes	-	Yes	Yes	-	-	Yes	-
Fitbit Sense 2*	Yes	-	Yes (cEDA)	Yes	-	-	Yes	-
WHOOP	Yes	Yes	-	Yes	Yes	Yes	Yes (motion-aware)	Yes (Peak/Life)
Oura Ring	Yes	Yes	-	Yes	-	-	Yes	-
Apple Watch	No native	No native	-	-	-	-	-	-
COROS	Yes	-	-	-	-	-	-	-
Amazfit	Yes	-	-	-	-	-	-	-
Polar	- (overnight only)	Yes	-	-	-	Yes	-	-

Three things stand out.

Only Samsung, Pixel Watch, and Fitbit Sense 2 use continuous EDA alongside HRV, which captures a direct sympathetic nervous system signal that HRV alone misses. (Other Fitbit models like the Charge 5/6 have spot-check EDA only, and models like Versa 4 and Inspire 3 have no EDA at all.
All Fitbits calculate a Stress Management Score, but the inputs vary by model. See our full Pixel Watch and Fitbit stress breakdown for model-by-model details.)
WHOOP is the only device pulling from five signals (HRV, HR, respiratory rate, skin temp, SpO2), but its stress monitor is locked behind a premium membership. And Apple Watch tracks the raw data but has no native stress score at all.

How Each Device Scores Stress

Every brand uses a different scale, different algorithm, and different baseline. This is why a Garmin 60 does not equal a WHOOP 1.5.

Device	Scale	Signals Used	Algorithm / Approach
Garmin	0–100 (0–25 rest, 26–50 low, 51–75 med, 76–100 high)	HRV, HR	Firstbeat Analytics. RMSSD-based. Samples every few minutes. Pauses during exercise. Body Battery (0–100 "energy" tank) drains with stress/exertion and recharges with rest. Optional 3-min chest-strap HRV Stress Test uses ECG-derived R-R intervals for training-readiness decisions. Elevate Gen 5 sensor (2023+) on Fenix 8, FR 965, Venu 4.
Samsung Galaxy Watch	0–100	HRV, HR, EDA	BioActive Sensor (3-in-1: PPG + ECG + BIA). GW7/Ultra added AGEs Index (metabolic aging from processed/sugary foods) + Energy Score. GW8/GW8 Classic (July 2025) added Vascular Load (overnight PPG waveform analysis, requires 3+ nights in 14 days) + Antioxidant Index (skin carotenoid scan, GW8 only). Note: AGEs/Vascular Load/Antioxidant Index reflect chronic metabolic stress, not acute stress score inputs.
Google Pixel Watch	Body Response alerts + 1–100	HRV, HR, EDA, Skin Temp	cEDA (continuous EDA) + ML model across 4 signals. Builds baseline over first month. Distinguishes exercise from stress.
Fitbit Sense 2*	Body Response alerts + 1–100	HRV, HR, EDA, Skin Temp	Same cEDA tech as Pixel Watch. Daily score from 12 metrics across 3 categories: exertion balance, sleep patterns, responsiveness. Other Fitbit models (Charge 5/6, Versa 4, Inspire 3) use fewer signals.
WHOOP	Stress 0–3 / Recovery 0–100%	HRV, HR, Resp Rate, Skin Temp, SpO2	lnRMSSD (log-transformed RMSSD) vs. 14-day personal baseline. HRV measured during deepest sleep for Recovery. Motion-aware Stress Monitor requires Peak/Life membership. Feb 2026 algorithm overhaul improved motion-artifact filtering for HR accuracy.
Oura Ring	Low / Med / High + trend	HRV, HR, Skin Temp, Sleep Metrics	3 layers: Daytime Stress (real-time HRV), Resilience (5 levels, 14-day baseline), Cumulative Stress (31-day scan, 21-31 day baseline, developed with Univ. of Southern Denmark). Finger-based PPG reads arterial signal with less motion noise than wrist.
Apple Watch	No native score	HRV (SDNN, not RMSSD), HR	No native stress feature in watchOS 11. HRV sampled as SDNN every ~4 hrs by default (every 2 hrs with Irregular Rhythm Notifications, every 15 min with AFib History). Third-party apps (Athlytic, Welltory, Livity) compute scores; Athlytic and HRV4Training derive RMSSD from raw R-R intervals.
COROS	0–100	HRV, HR	HRV + HR every 5 min vs. individual baseline. Endurance athlete focus.
Amazfit	0–100	HRV, HR	RMSSD-based HRV + proprietary model. Auto-measures every 5 min.
Polar	ANS Charge (−10 to +10)	HRV, HR, Resp Rate	Nightly Recharge only (overnight, first ~4 hrs of sleep). Measures recovery from prior day. No daytime stress score.

Why Scores Don't Compare Across Brands

This is the most common mistake people make.

Reason	What It Means
Different scales	Garmin: 0–100. WHOOP: 0–3. Oura: Low/Med/High. Polar: −10 to +10. The numbers aren't on the same axis.
Different signals	A 4-signal score (Pixel Watch) incorporates data a 2-signal score (Garmin) can't see. Adding EDA and skin temp changes what gets detected.
Different baselines	Most devices compare against a personal baseline: 14-day for WHOOP, first-month for Pixel Watch/Fitbit, 14-day for Oura Resilience, 21-31 day for Oura Cumulative Stress, 28-day for Polar, 60-day for Athlytic on Apple Watch. Your "normal" is calibrated to your history, not a universal standard.
Different timing	Polar only measures overnight. Garmin pauses during exercise. Oura runs continuous daytime stress but rolls up cumulative stress daily.
Proprietary algorithms	Each algorithm is a black box. Two devices reading the same raw HRV data will produce different scores.

The practical takeaway: pick one device, track trends within it over time, and don't compare your numbers to someone wearing a different brand. The trend matters. The absolute number does not.

What Signals Are Behind Every Stress Score

Every wearable stress score starts with one or more physiological signals. Here's what each one captures and which devices use it.

Signal	What It Measures	Used By
HRV (RMSSD / SDNN)	Parasympathetic nervous system activity via beat-to-beat heart rate variation. Primary driver for most devices. All use RMSSD except Apple Watch, which uses SDNN (total autonomic variability). WHOOP uses lnRMSSD (log-transformed).	All 10 devices
Heart Rate	Sympathetic activation via elevated resting or real-time HR. Secondary driver.	All 10 devices
EDA (Electrodermal Activity)	Sympathetic NS activation via skin conductance. Detects arousal directly, but cannot distinguish positive from negative.	Samsung, Pixel Watch, Fitbit Sense 2 (continuous); Fitbit Charge 5/6 (spot-check)
Skin Temperature	Peripheral vasoconstriction (stress drops skin temp) and circadian patterns. Major confounder: ambient temperature.	Pixel Watch, Fitbit Sense 1/2, WHOOP, Oura Ring
Respiratory Rate	Autonomic state via breathing rate changes. Stress and illness both raise it.	WHOOP, Polar
SpO2 (Blood Oxygen)	Respiratory and cardiovascular recovery status. Drops at altitude, with sleep apnea, or illness.	WHOOP
Sleep Metrics	Sleep architecture as an input to cumulative stress and recovery calculations.	Oura Ring

More signals generally means better accuracy. But no combination on any consumer device can tell whether you're stressed, excited, caffeinated, or cold. This is a physiological limitation of the autonomic nervous system, not an algorithm problem.

What Actually Moves Your Stress Score

Because most devices rely primarily on HRV and heart rate, the factors that move your stress score are largely the same across brands. The difference is whether additional sensors (EDA, skin temperature, SpO2) pick up things HRV-only devices miss.

Factors that improve stress scores (lower stress / better recovery)

Factor	Mechanism	Effect Size
Consistent sleep (7–9 hrs)	Restores parasympathetic dominance, increases vagal tone	15–30% HRV improvement within 4 weeks (Frontiers in Physiology, 2024)
Aerobic exercise (150 min/wk)	Enhanced cardiovascular fitness and vagal tone	Significant long-term HRV increase (PMC8950456)
Meditation / breathwork	Activates parasympathetic NS, slows respiration rate	Acute and chronic HRV improvement (Frontiers in Physiology, 2024)
Healthy body weight	Restores sympathovagal balance, reduces sympathetic activation	Increases parasympathetic activity (Frontiers in Physiology, 2024)
Adequate hydration	Maintains blood volume, reduces cardiac strain	Moderate effect on HRV (PMC9549087)
Cold exposure (controlled)	Triggers vagus nerve via dive reflex	Acute vagal stimulation (Marathon Handbook, 2026)

Factors that worsen stress scores (higher stress / worse recovery)

Factor	Mechanism	Effect Size
Alcohol (even 1 drink)	Suppresses parasympathetic activity directly	RMSSD drops ~2ms/drink; 3+ drinks = up to 13ms drop for 2–5 days (Frontiers 2024)
Sleep deprivation	Shifts autonomic balance toward sympathetic dominance	Significant acute HRV reduction (Frontiers in Physiology, 2024)
Overtraining without recovery	Excessive physical stress suppresses parasympathetic tone	Progressive HRV decline (PMC8950456)
Chronic psychological stress	Sustained sympathetic activation suppresses vagal tone	Sustained RMSSD/SDNN reduction (PMC9974008)
Illness / inflammation	Immune response activates sympathetic NS	Significant HRV drop during acute illness (Frontiers in Physiology, 2024)
Excess caffeine	Overstimulates sympathetic nervous system	8–12% HRV drop in sensitive individuals (PMC11284693)
Sedentary lifestyle	Deconditioned heart works harder at rest	Most common cause of elevated resting HR (Hackensack Meridian Health)

For every factor broken down per device with direction, magnitude, and sources, explore the Stress Factor Explorer.

If you want to go deeper on HRV specifically, we ranked all 44 HRV factors by evidence strength in a separate post.

Track what's behind your stress score. Download Kygo on iOS or Android and start connecting your nutrition to your wearable data.

The EDA Difference: Samsung, Google, and Fitbit

Three devices use electrodermal activity (EDA), a signal that measures sympathetic nervous system activation through changes in skin conductance. This is fundamentally different from HRV. While HRV measures your rest system backing off, EDA measures your fight-or-flight system firing up directly.

EDA Factor	Direction	Mechanism
Emotional arousal (anxiety, fear, anger)	↑ Increases stress reading	Sympathetic NS triggers eccrine sweat glands
Cognitive load / mental effort	↑ Increases	Mental exertion activates sympathetic NS
Sensory stimulation (loud sounds, surprise)	↑ Increases	Startle/orienting response
Ambient heat + humidity (confounder)	↑ Increases	Thermoregulatory sweating, not emotional
Excitement / positive arousal	↑ Increases	EDA cannot distinguish positive from negative arousal
Relaxation / meditation	↓ Decreases	Parasympathetic activation reduces sympathetic drive
Cool ambient temperature	↓ Decreases	Less thermoregulatory sweating

Google Pixel Watch and Fitbit Sense 2 use continuous EDA (cEDA), tracking all day on the wrist. This was the first consumer all-day EDA implementation (2022). The Fitbit Charge 5 and Charge 6 also have EDA, but only as a 90-second spot-check (not continuous). Samsung uses spot-check EDA via its BioActive sensor, with enhanced continuous tracking on the Galaxy Watch 8.

How Accurate Are Wearable Stress Scores?

This is where most wearable marketing falls apart.

Finding	Detail	Source
Multi-signal accuracy	~82% for HRV + EDA + skin temp combined	JMIR Meta-Analysis, 2024
Single-signal accuracy	~77% for HRV alone	JMIR Meta-Analysis, 2024
Lab-to-real-world gap	Models trained in controlled stress tests lose meaningful accuracy in daily life. The field's biggest unsolved problem.	Frontiers Systematic Review, 2024; ScienceDirect Three-Stage Validation, 2025
Arousal, not stress	EDA and HRV cannot distinguish positive from negative arousal. Physiological limit.	Frontiers Systematic Review, 2024
Subjective vs. physiological	Self-reported stress often does not track with cortisol or HRV changes	Nature Comms Medicine, 2025
Cross-user generalization	Real-time prediction works, but models are person-specific. Cross-user generalization is poor.	PMC11230864
TSST reproducibility	Across 171 studies (n=8,452), the Trier Social Stress Test produces robust cortisol + cardiovascular + sympathetic responses, making it the reference protocol for lab stress induction	ScienceDirect TSST Meta-Analysis, 2025
Consumer device reproducibility	Reproducing lab-grade stress detection on consumer wearables remains difficult. Algorithm results don't transfer cleanly across devices.	arXiv, 2025

Device-specific sensor accuracy

Device	HRV Finding	HR Finding	Source
Apple Watch Series 9 / Ultra 2	SDNN underestimated by 8.31 ms vs. Polar H10 chest strap; MAPE 28.88%; failed ±10 ms equivalence	RHR MAPE 5.91%, MAE 3.73 bpm, mean diff -0.1 bpm	Sensors 2024 / Hwang et al. (n=39, 14 days)
Apple Watch Series 6+	N/A (meta did not aggregate HRV)	BPM MAPE 1.16%-6.46%; limits of agreement -3.68 to 2.59 bpm	npj Digital Medicine 2025
WHOOP 4.0	Day-to-day lnRMSSD variability within or below reference protocol range	Strong agreement at rest; degraded with motion (pre-Feb-2026 update)	PMC9505647 (Olympic water polo, n=14); PMC8160717

There is no single gold standard for measuring "stress." Researchers use the Trier Social Stress Test (TSST) for inducing acute stress, salivary cortisol for biochemical measurement, and 12-lead ECG-derived HRV for autonomic measurement. Consumer wearables approximate these with wrist-based PPG sensors, which introduces noise compared to lab-grade equipment.

For a detailed breakdown of how accurate each wearable's raw sensor data actually is across 17 peer-reviewed studies, that comparison goes deeper.

Connecting Stress Scores to Causes

Your stress score tells you something happened. It doesn't tell you what caused it.

If your score spikes on a Tuesday afternoon, it could be a tense meeting, your third coffee, dehydration, or five hours of sleep the night before. The score alone can't differentiate. The only way to connect the output (stress score, HRV, recovery) to the input (nutrition, caffeine, sleep, hydration) is by tracking both sides.

Kygo connects to your wearable (Oura, Apple Watch, Garmin, Fitbit, or WHOOP) and layers food logging, caffeine timing, and meal data on top of your biometric readings. The correlation engine analyzes patterns across days and weeks to surface connections like "Your afternoon stress scores average 15 points higher on days you skip lunch" or "Your recovery drops 12% on nights following 3+ drinks."

Instead of guessing why your HRV dropped, you see the pattern in your own data. Explore every factor by device in the Stress Factor Explorer, or start tracking at www.kygo.app.

Ready to see what's actually moving your numbers? Download Kygo on iOS or Android.

Deep Dives by Device

Want the full breakdown for your specific wearable? Each post covers the unique signals, algorithms, and factors for that device:

Disclaimer: Kygo is a personal data aggregation and insights platform designed for informational purposes only. The information provided by Kygo, including correlations, patterns, and trends identified in your data, does not constitute medical advice, diagnosis, or treatment. Always consult a licensed healthcare provider with any questions regarding medical conditions.

What does your stress score usually look like? Have you noticed patterns between certain habits and your numbers? Share your experience.