The ECG feature on Apple Watch records a single-lead electrocardiogram by measuring the tiny electrical signals your heart produces with each beat. You open the ECG app, hold your fingertip on the Digital Crown for 30 seconds, and the watch generates a rhythm strip that its software classifies as normal sinus rhythm, atrial fibrillation, or inconclusive. The feature received FDA clearance as a Class II medical device, but the technology sits in a particular niche: it is genuinely useful for catching one specific heart-rhythm problem, yet far more limited than the full ECGs recorded in a hospital or clinic.
How the Watch Actually Captures a Heart Rhythm
Every heartbeat begins with an electrical impulse that spreads across the heart muscle in a predictable pattern. That electrical activity is strong enough to be detected at the skin’s surface, which is how traditional ECGs work: electrodes stuck to the chest and limbs pick up voltage differences from multiple angles, producing the familiar 12-lead readout doctors use to diagnose a wide range of cardiac conditions.
The Apple Watch takes a stripped-down approach. It has an electrode built into the crystal back that rests against your wrist, and a second electrode in the Digital Crown on the side of the case. When you touch the crown with the opposite hand, your body completes a circuit: electrical signals travel from the heart through the torso to both arms, and the watch measures the voltage difference between those two contact points. This gives you a single electrical “view” of the heart, roughly equivalent to one of the standard clinical leads known as Lead I.
Because the signal picked up at the wrist is much smaller than what clinical electrodes capture on the chest, the watch relies heavily on signal processing to filter out noise from muscle movement, skin contact shifts, and electrical interference. Research on wearable ECG devices has shown that motion artifacts are one of the biggest challenges, and adaptive filtering techniques using reference signals can substantially clean up the recording.
What the Watch Is Looking For
The Apple Watch ECG algorithm is designed to do one thing well: distinguish normal sinus rhythm from atrial fibrillation. Atrial fibrillation, often shortened to AFib, is the most common sustained heart-rhythm disorder, and it significantly raises the risk of stroke. On an ECG tracing, AFib shows up as an absence of the normal P wave (the small blip that represents the atria firing in an organized way) combined with irregular spacing between the R peaks (the tall spikes marking each ventricular contraction).1PubMed Central. A Review of Atrial Fibrillation Detection Methods as a Service The watch’s algorithm looks for exactly those two features: missing P waves and irregular R-to-R intervals.
A systematic review and meta-analysis pooling data from multiple studies found that the Apple Watch detected AFib with a pooled sensitivity of about 95% and a specificity of about 95% as well, meaning it correctly identifies AFib roughly 19 times out of 20 and correctly identifies normal rhythm at a similar rate.2PubMed Central. Diagnostic Accuracy of Apple Watch Electrocardiogram for Atrial Fibrillation: A Systematic Review and Meta-Analysis Those numbers are impressive for a consumer device strapped to your wrist, and they explain why the feature has drawn serious attention from cardiologists.
But the algorithm can also return “inconclusive,” meaning it could not confidently classify the rhythm. One study examining 242 inconclusive smartwatch ECGs found that about 7% were flagged because the heart rate was too high, 17% because it was too low, and for the remaining 76%, the specific reason was not reported by the device.3JACC: Clinical Electrophysiology. Wearable Technology Inconclusive Single-Lead ECGs From Smart-Devices: Impact of Baseline ECG Anomalies Poor skin contact, excessive wrist movement, or pre-existing ECG abnormalities can all push a recording into the inconclusive bucket.
The Irregular Rhythm Notification Is a Different Feature
People often conflate two separate Apple Watch capabilities. The ECG app requires you to sit still, open the app, and actively record for 30 seconds. The irregular rhythm notification, by contrast, runs passively in the background using the watch’s optical heart sensor, which shines green LED light into the skin and measures blood flow changes to estimate pulse rate and rhythm. This photoplethysmography (PPG) approach does not produce an ECG tracing at all. It monitors pulse patterns over time and sends a notification if it detects irregularity suggestive of AFib.
The distinction matters because the two features have different strengths. The passive PPG check can catch intermittent AFib episodes you might not feel, since many people with AFib are completely asymptomatic. But it cannot generate a waveform a doctor can interpret. When you get an irregular rhythm notification, the recommended next step is to open the ECG app and take a recording, which gives your doctor something concrete to review. Surveys have found that some physicians would consider pursuing further evaluation or even starting anticoagulation based on smartwatch findings alone, though evidence that this approach improves outcomes in asymptomatic patients is still lacking.4PubMed Central. Physician Responses to Apple Watch-Detected Irregular Rhythm Alerts
Why a Single Lead Misses a Lot
A hospital ECG uses 10 electrodes placed on the chest, arms, and legs to generate 12 different electrical perspectives of the heart. Each lead reveals something different: some are better for spotting damage in the front wall of the heart, others for the side or bottom. The Apple Watch’s single lead captures just one of those views. This is more than enough to detect the hallmark irregularity of AFib, but it is fundamentally limited when it comes to conditions that require spatial information from multiple angles.
The biggest gap involves detecting heart attacks. During a heart attack, the ECG typically shows characteristic ST-segment changes, where a specific portion of the waveform shifts upward or downward. A review of consumer ECG devices concluded that while single-channel devices like smartwatches are useful for detecting and monitoring atrial fibrillation, flutter, and ectopic complexes, no consumer ECG device can currently reliably detect ST-segment deviations that would indicate a heart attack or ischemic episode.5PubMed. Review of current ECG consumer electronics (pros and cons) So if you are having chest pain, the Apple Watch ECG is not the tool to rely on. Call emergency services.
Researchers have been exploring workarounds. By holding the watch against different positions on the chest and limbs, it is possible to approximate additional ECG leads. One study showed that smartwatch recordings taken at various body positions could be matched to their corresponding standard 12-lead ECG leads with high accuracy, and all 300 recordings in that study were usable for diagnostic purposes.6PubMed Central. Single-Lead ECG Recordings Including Einthoven and Wilson Leads by a Smartwatch: A New Era of Patient Directed Early ECG Differential Diagnosis of Cardiac Diseases? This multi-position approach is a research technique, not a built-in feature, but it suggests that future consumer devices could eventually offer broader diagnostic capability. Researchers have reviewed progress toward using the Apple Watch to record multiple leads for detecting myocardial infarction, concluding that the data are encouraging but significant limitations remain before the watch could serve as a self-check tool for people experiencing chest pain.7PubMed Central. Using the Apple Watch to Record Multiple-Lead Electrocardiograms in Detecting Myocardial Infarction: Where Are We Now?
How It Compares to Other Consumer ECG Devices
The Apple Watch is the highest-profile consumer ECG device, but it is not the only one. AliveCor’s KardiaMobile is a dedicated, palm-sized ECG recorder that also captures a single-lead tracing when you place your fingers on its electrodes. Withings makes smartwatches with built-in ECG capability as well. How do they stack up?
A head-to-head comparison of the Apple Watch Series 4 and KardiaMobile found that the Apple Watch detected sinus rhythm with 100% accuracy versus 99% for KardiaMobile. For AFib detection, the picture flipped: KardiaMobile scored 100% while the Apple Watch came in at about 91%. The Apple Watch also produced 19 inconclusive readings during AFib, compared to none for KardiaMobile.8PubMed Central. Comparison of Apple Watch vs KardiaMobile: A Tale of Two Devices A broader comparison across Withings, Apple, and KardiaMobile devices found their overall sensitivity and specificity were statistically similar, with no significant differences. KardiaMobile was the most sensitive (best at catching AFib when it was present), while Apple was the most specific (best at correctly identifying normal rhythms). All three devices produced uninterpretable ECGs about 20-24% of the time.9PubMed Central. A comparison of over-the-counter available smartwatches and devices for electrocardiogram based detection of atrial fibrillation
The practical tradeoff is convenience. The Apple Watch is already on your wrist, making it trivially easy to take a recording the moment you feel a flutter or skip. KardiaMobile requires carrying a separate device, but it tends to produce fewer inconclusive readings and has slightly better AFib sensitivity. For someone who already wears an Apple Watch, the ECG feature is a genuine bonus. For someone specifically managing known AFib and wanting the most reliable home recordings, a dedicated device may edge ahead.
The “Worried Well” Problem
Putting a medical sensor on millions of wrists has downstream consequences that extend beyond individual health. One study that tracked patients who sought medical evaluation after receiving an abnormal pulse notification from their Apple Watch found that new clinically actionable cardiovascular diagnoses were made in only about 11% of those patients.10Journal of the American Medical Informatics Association. Clinical evaluation and diagnostic yield following evaluation of abnormal pulse detected using Apple Watch That means roughly nine out of ten people who followed up had nothing clinically wrong, at least nothing new that required treatment. The researchers flagged this as a concern: false positive screening results can drive excessive healthcare resource use and generate anxiety among otherwise healthy people.
This does not mean the feature is useless. AFib often goes undetected for years because it can be intermittent and asymptomatic, and catching it early enough to start blood-thinning medication can prevent strokes. But the math of screening large populations means that even a test with 95% specificity will produce a flood of false alarms when applied to millions of people who have a very low baseline risk. If you are young and have no cardiac risk factors, an occasional irregular rhythm notification is far more likely to be a false alarm than an actual problem. If you are older, have risk factors, or have felt palpitations, the same notification carries more weight.
Children and the Apple Watch ECG
Apple’s ECG feature is FDA-cleared for users 22 years and older. There are currently no data supporting its automated algorithm in children, and the device explicitly states this age restriction.11PubMed. Accuracy of the Apple Watch iECG in Children With and Without Congenital Heart Disease That does not mean the hardware is incapable of recording a useful signal from a child’s wrist. A study assessing the Apple Watch Series 6 in pediatric patients found that the device could provide good-quality single-lead ECGs allowing accurate measurement of key ECG intervals when interpreted manually by a physician.12PLOS Digital Health. Assessment of Apple Watch Series 6 pulse oximetry and electrocardiograms in a pediatric population
The catch is the automated classification. A systematic review of wearable ECG devices in pediatrics found that automated rhythm classification frequently misidentified pediatric arrhythmias. Smartwatch-based ECG recordings showed high diagnostic accuracy when a cardiologist read them manually, but the automated algorithms were unreliable for tachyarrhythmias and conduction abnormalities in children.13PubMed Central. Wearable and Portable Electrocardiographic Devices as Modern Cardiac Telemetry Solutions in Pediatrics: A Systematic Review The underlying issue is straightforward: children have faster resting heart rates and different ECG morphology than adults, and the watch’s algorithm was trained on adult data. A parent strapping an Apple Watch onto a teenager with known palpitations might get a useful recording, but only if a physician reviews the raw tracing rather than relying on the watch’s classification.
Improving AFib Detection Beyond Simple Irregularity
The standard approach to detecting AFib on a single-lead tracing is to look for irregular spacing between heartbeats. But irregular rhythms are not always AFib. Frequent premature beats, a wandering pacemaker, or simple movement artifact can all produce irregular intervals that trick the algorithm into a false alarm or an inconclusive reading. Researchers have been working to make smartwatch detection smarter by looking for patterns within the irregularity.
One approach screens for “regularity within irregularity,” essentially identifying clusters of intervals that follow a repeating pattern rather than the truly chaotic spacing of AFib. By excluding ECGs that contain these regular subpatterns from the AFib bucket, the algorithm can reduce false positives without throwing out recordings that other methods would label inconclusive due to heart rate cutoffs or quality filters.14PubMed Central. Improving Automatic Smartwatch Electrocardiogram Diagnosis of Atrial Fibrillation by Identifying Regularity within Irregularity This kind of refinement matters because the roughly 20% inconclusive rate seen across consumer ECG devices is a real usability problem. Every inconclusive reading is a moment of uncertainty for the person wearing the watch and a potential unnecessary doctor visit.
What to Do with Your Recordings
The Apple Watch stores ECG recordings as PDF waveforms that you can share with your doctor through the Health app. This is genuinely useful in specific scenarios. If you feel palpitations that come and go, capturing a tracing during the episode gives your cardiologist something objective to analyze. In traditional cardiology, catching intermittent arrhythmias requires wearing a Holter monitor for 24 to 48 hours or an event monitor for weeks, and even then, the arrhythmia sometimes fails to show up during the monitoring window. A watch that is already on your wrist closes that gap.
Where the recordings become less helpful is when you take them compulsively, checking your ECG multiple times a day in the absence of symptoms. A normal result does not rule out heart disease, and an inconclusive result does not confirm it. If your doctor orders a 12-lead ECG or other cardiac workup based on symptoms or risk factors, the Apple Watch recording does not substitute for that evaluation. The device functions best as a first filter: it can flag something worth investigating and provide a snapshot of what your rhythm looked like at a specific moment, but the clinical decision-making still belongs to your doctor.
Ongoing Integration Challenges
One underappreciated hurdle is getting smartwatch ECG data smoothly into the medical system. Most electronic health record systems were not designed to accept waveform data from consumer devices. A patient showing their doctor a PDF on a phone screen is functional but clunky. Researchers have noted that it remains unclear how data from these devices can be optimally incorporated into patient care, and what the broader implications are for clinical workflows and health system resources.7PubMed Central. Using the Apple Watch to Record Multiple-Lead Electrocardiograms in Detecting Myocardial Infarction: Where Are We Now? Some cardiology practices have begun accepting Apple Watch ECGs as part of telehealth visits, but standardized processes for reviewing, storing, and acting on consumer-generated cardiac data are still developing. The technology has outpaced the infrastructure designed to use it.