Irregular heartbeats leave visible fingerprints on the waveform a pulse oximeter draws across its screen. That waveform, called a photoplethysmogram (or PPG), is built from tiny fluctuations in how much light passes through your fingertip with each cardiac cycle, and when the rhythm driving those cycles becomes erratic, the shape, timing, and amplitude of each pulse all change in ways that clinicians and algorithms can learn to read. Understanding what those distortions look like, what causes them, and where the technology’s blind spots lie has become increasingly relevant as pulse oximetry moves from hospital monitors onto consumer smartwatches.
How a Pulse Oximeter Builds Its Waveform
A pulse oximeter shines two wavelengths of light, typically red and infrared, through tissue and measures how much light makes it to a detector on the other side. Each heartbeat pushes a small surge of arterial blood into the fingertip, briefly increasing the amount of light absorbed. The device plots these absorption changes over time, producing the PPG waveform you see scrolling across the screen. Medical-grade devices use red and infrared light, while most consumer smartwatches rely on green light instead, which is better at picking up signal through skin in motion but carries different trade-offs for oxygen saturation measurement.1PubMed Central. Photoplethysmogram Analysis and Applications: An Integrative Review
A healthy, normal waveform has a recognizable anatomy. Each pulse rises steeply to a systolic peak as the heart ejects blood, then descends with a small secondary bump partway down called the dicrotic notch. That notch is often taught as the moment the aortic valve snaps shut, but the full story is more interesting: the waveform is actually two superimposed pressure waves, the original forward pulse and a reflection bouncing back from the lower body’s vasculature. How and when that reflected wave arrives shapes the notch and the overall contour of the downslope.2PubMed Central. Arterial flow waveforms, vascular tone, and chronic fatigue: A case report When the heart’s rhythm is regular and the vasculature is healthy, the waveform repeats in a tidy, predictable pattern. Irregular heartbeats break that pattern in several distinct ways.
What Atrial Fibrillation Does to the Waveform
Atrial fibrillation, the most common sustained arrhythmia, turns a steady PPG trace into something visibly chaotic. Because the atria fire erratically, every ventricular contraction arrives at a slightly different moment, with a slightly different volume of blood behind it. The result on the waveform is beat-to-beat variation in both timing and shape. A study comparing PPG morphology in atrial fibrillation patients against people in normal sinus rhythm found significantly higher variability in waveform features during fibrillation, including changes in the steepness of the upstroke, the ratio of key inflection points on the waveform, and the total area under each pulse. Several of these morphological features, along with heart-rate variability measures, reliably differentiated the two groups.3Physiological Measurement. Photoplethysmography wave morphology in patients with atrial fibrillation
In practical terms, if you watch the pulse oximeter screen of someone in atrial fibrillation, you will see pulses that vary in height, width, and spacing from one beat to the next. The dicrotic notch may be hard to identify or shift around unpredictably. The displayed heart rate often fluctuates noticeably between successive readings, because the device is averaging over a rapidly changing interval. These visual cues are what experienced nurses and anesthesiologists look for at the bedside, and they are increasingly what machine-learning algorithms are trained to detect automatically.
Why Some Beats Vanish Entirely
Not every electrical heartbeat produces a detectable pulse at the fingertip. During atrial fibrillation, some ventricular contractions happen so soon after the previous beat that the heart has not had time to fill adequately. These weak contractions may push too little blood into the periphery for the pulse oximeter to register. The gap between the number of electrical heartbeats (counted on an ECG) and the number of peripheral pulses actually felt or measured is called the pulse deficit. It results from a kind of electrical-mechanical disconnection: the heart’s electrical system fires, but the mechanical pump action does not follow through strongly enough to reach the fingertip.4Circulation. Pulse Deficit as a New, Innovative, Non-Invasive Measure to Risk-Stratify Atrial Fibrillation Patients
This is clinically meaningful. A pulse oximeter that shows a rate of 72 beats per minute while the ECG simultaneously counts 90 is revealing a substantial pulse deficit. That discrepancy tells the clinical team something about how effectively the heart is actually pumping, beyond what the electrical tracing alone conveys. It also means the SpOâ‚‚ reading itself may be less reliable during these episodes, since the device depends on pulsatile blood flow to separate arterial signal from background tissue absorption.
The One-Beat Offset During Premature Contractions
Premature ventricular contractions (PVCs) and other ectopic beats create a different kind of waveform puzzle. When a PVC fires, it produces a low-amplitude arterial pulse because the ventricle contracts before it has properly filled. On a bedside monitor that displays both an arterial line tracing and a pulse oximeter waveform simultaneously, you might expect these two to line up. They often do not. A case series of 20 premature contraction events across 10 patients found that the small arterial-line pulse aligned with the premature ECG beat as expected, but the corresponding small pulse on the pulse oximeter waveform was shifted forward and aligned instead with the next normal heartbeat.5A&A Practice. Pulse Oximetry Waveform Represents an Earlier Ventricular Contraction in Relation to the Intraarterial Blood Pressure Tracing and Electrocardiogram on Multichannel Monitors
In other words, the pulse oximeter plethysmograph runs about one full QRS complex behind the arterial waveform. This lag exists because the oximeter measures blood volume changes at the fingertip, which take longer to arrive and are filtered through the compliance of the peripheral vasculature. The offset is usually invisible when every beat is the same size, but a sudden small beat from a PVC makes the mismatch obvious. For clinicians reading multichannel monitors, this is important to know: the odd-looking pulse on the oximeter trace is not necessarily the beat you think it is.
When the Problem Is Not Your Heart
One of the trickiest aspects of PPG waveform analysis is that many things besides arrhythmia can make the tracing look abnormal. Motion artifacts are the biggest offender. Even small finger movements introduce noise that mimics the irregular pulse-to-pulse intervals seen in atrial fibrillation. This overlap is a major challenge for automated AF detection: a normal sinus rhythm signal corrupted by motion can be incorrectly flagged as AF, and conversely, genuine AF can be masked by movement noise.6npj Digital Medicine. Photoplethysmography based atrial fibrillation detection: a review
Cold fingers are another common culprit. When your hands are cold, the blood vessels in your fingertips constrict, dramatically reducing the pulsatile signal the oximeter depends on. One study found that cold exposure reduced the average PPG pulse amplitude by roughly 50 to 60% compared to baseline. Warming the finger back up reversed the effect, improving signal quality by up to four times and bringing SpO₂ estimates much closer to reference values. The researchers recommended maintaining the sensor site near 33°C for reliable readings.7Biomedical Signal Processing and Control. Analysing the effects of cold, normal, and warm digits on transmittance pulse oximetry In a clinical setting, hypothermia can degrade finger PPG signals so severely that the device fails entirely. One experiment showed that during cold exposure, finger PPG signal strength dropped by over 80%, and SpO₂ readings fell below 90% in five out of a group of volunteers, representing outright device failure. An ear-canal sensor, by contrast, maintained its signal far better because the ear is less affected by peripheral vasoconstriction.8PubMed Central. In vivo investigation of ear canal pulse oximetry during hypothermia
Even under conditions of generally poor perfusion (from shock, heavy sedation, or vasopressor use), modern pulse oximeters perform better than their predecessors. A systematic review found that about three-quarters of tested oximeter models still gave accurate readings in poorly perfused patients, with newer devices and earlobe placement outperforming older finger-based models.9PubMed. Accuracy of pulse oximeters in measuring oxygen saturation in patients with poor peripheral perfusion: a systematic review
Skin Pigmentation and Waveform Accuracy
The light a pulse oximeter sends through tissue does not interact identically with every patient’s skin. Melanin absorbs light, and darker skin contains more of it, which increases the total amount of absorption unrelated to blood oxygen. Additionally, the size and distribution of light-scattering particles differ with skin pigmentation, changing the path photons take through tissue. These factors mean that the calibration curves built into most commercial pulse oximeters, which were largely developed on lighter-skinned populations, may not perfectly translate across all skin tones.10PubMed Central. A review of the effect of skin pigmentation on pulse oximeter accuracy This does not necessarily make the waveform shape unreadable, but it can shift the SpOâ‚‚ number the device reports, sometimes overestimating true oxygen saturation in patients with darker skin. Awareness of this bias has grown substantially in clinical practice in recent years, and ongoing engineering efforts aim to develop more inclusive calibration models.
Wearable Devices and Atrial Fibrillation Screening
Consumer smartwatches have turned PPG-based arrhythmia detection into a mass-market technology. These devices continuously sample the PPG waveform from your wrist and use algorithms to flag irregular rhythms. Their performance for detecting atrial fibrillation has been tested extensively. A meta-analysis pooling data from 11 studies on PPG-based smartwatches found a pooled sensitivity of about 97% and specificity of about 97%, meaning they correctly identify the vast majority of both AF and non-AF rhythms.11PubMed Central. Comparison of diagnostic accuracy of electrocardiogram-based versus photoplethysmography-based smartwatches for atrial fibrillation detection: A Systematic Review and Meta-Analysis Interestingly, PPG-based watches in that analysis actually outperformed ECG-based smartwatches, which showed pooled sensitivity around 83% and specificity around 88%. The likely explanation is that PPG continuously monitors in the background, catching more episodes, while ECG-equipped watches require the user to actively touch an electrode and hold still.
One of the landmark real-world demonstrations was the Huawei Heart study, which monitored nearly 188,000 participants using wrist-worn PPG devices over at least 14 days. Among those flagged as having suspected AF, about 87% were confirmed on follow-up ECG, with a positive predictive value above 91%.6npj Digital Medicine. Photoplethysmography based atrial fibrillation detection: a review The WATCH AF trial tested a dedicated PPG algorithm and reported 94% sensitivity, 98% specificity, and 96% overall accuracy for AF detection.12PubMed. The WATCH AF Trial: SmartWATCHes for Detection of Atrial Fibrillation These numbers are impressive for a screening tool, though they come with important caveats: wrist-based PPG signals are noisier than fingertip readings, and as described earlier, motion artifacts can mimic or mask AF. A positive smartwatch notification should always be confirmed with a clinical ECG.
Pulse Rate Variability Is Not the Same as Heart Rate Variability
Many wearable devices report “heart rate variability” (HRV) derived from the PPG signal. Strictly speaking, what they measure is pulse rate variability (PRV), the variation in the time between successive pulses at the wrist or fingertip, which is not always identical to the electrical beat-to-beat variation measured by an ECG. In a healthy person at rest, the two track closely. But when the cardiovascular system is stressed, by arrhythmia, by changes in vascular tone, or by altered blood pressure, PRV and true HRV can diverge. The pulse wave has to travel through arteries that have their own compliance and resistance, and that transit modifies the timing in ways the ECG does not capture.13PubMed. Pulse rate variability in cardiovascular health: a review on its applications and relationship with heart rate variability This does not mean wearable HRV readings are useless, but it does mean they should be interpreted carefully, especially in people with known cardiac conditions. The PRV signal may contain valuable information of its own about vascular health, but treating it as a perfect stand-in for ECG-based HRV can be misleading.
Reading Vascular Health from the Waveform Shape
Beyond heart rhythm, the PPG waveform carries information about the blood vessels themselves. As arteries stiffen with age, high blood pressure, or diabetes, the reflected pulse wave described earlier arrives back at the fingertip sooner, merging with the forward wave and changing the waveform’s shape. This phenomenon can be quantified as an augmentation index, essentially a measure of how much the reflected wave boosts the systolic peak. One study found a strong correlation (r = 0.85) between an augmentation index derived from the PPG and the gold-standard aortic augmentation index measured invasively, with both indices increasing with age and showing significant differences between people with diabetes and healthy controls.14Physiological Measurement. Photoplethysmographic signal waveform index for detection of increased arterial stiffness Gaussian fitting methods have been developed to isolate the forward peak of the PPG waveform more precisely, enabling better estimation of the augmentation index and instantaneous heart rate even in noisy real-world conditions.15PubMed. Investigation on Pulse Wave Forward Peak Detection and Its Applications in Cardiovascular Health
This line of research is significant because it means a simple finger clip, or even a wrist-worn sensor, could potentially screen for arterial stiffness during a routine check, flagging people who might benefit from further cardiovascular workup. The technology is not there yet for clinical decision-making on its own, but the waveform shape carries more health information than the single SpOâ‚‚ number most people associate with pulse oximetry.
Blood Pressure Estimation and Its Limits
A particularly active area of research asks whether the PPG waveform can estimate blood pressure without a cuff. The basic idea exploits pulse transit time (PTT), the delay between the heart’s electrical activity and the pulse’s arrival at the periphery, which shortens as blood pressure rises. Some prototypes have shown promising accuracy. One wrist-worn PTT-based device achieved a mean absolute difference of about 5.4 mmHg for systolic blood pressure and around 2.9 mmHg for diastolic, which met or approached engineering standards for acceptable accuracy.16PubMed Central. Enabling Wearable Pulse Transit Time-Based Blood Pressure Estimation for Medically Underserved Areas and Health Equity: Comprehensive Evaluation Study Another approach combined PTT with the ratio of the PPG signal’s intensity components, capturing both fast and slow blood-pressure fluctuations, and achieved mean errors near zero with standard deviations around 4 to 5 mmHg for systolic, diastolic, and mean arterial pressure.17PubMed. Continuous Cuffless Blood Pressure Estimation Using Pulse Transit Time and Photoplethysmogram Intensity Ratio
The enthusiasm should be tempered, though. An independent analysis of how much blood-pressure information the PPG signal actually contains found that the features extracted from the waveform captured less than 10% of the total variability in blood pressure, and much of what they did capture overlapped with simple heart rate.18Scientific Reports. Examining the challenges of blood pressure estimation via photoplethysmography That suggests the impressive accuracy numbers from controlled studies may not generalize well to the messy, variable conditions of everyday life. Cuffless blood pressure from a wrist sensor remains a goal, not a solved problem.
False Alarms and Alarm Fatigue in Hospitals
In intensive care units and post-anesthesia recovery rooms, pulse oximeter alarms are among the most frequent monitor alerts, and the majority of them are false. This creates alarm fatigue: staff become desensitized and may respond more slowly to genuine emergencies. A study in a pediatric post-anesthesia care unit compared a conventional pulse oximeter against one using more advanced signal-processing (Masimo SET technology) and found that the conventional device generated more than twice as many false alarms. It also missed true desaturation events that the newer device caught: only 59% of genuine low-oxygen episodes were flagged by the conventional oximeter, versus 100% by the advanced one.19Anesthesia & Analgesia. False Alarms and Sensitivity of Conventional Pulse Oximetry Versus the Masimo SETâ„¢ Technology in the Pediatric Postanesthesia Care Unit
More recent work has explored machine-learning classifiers that analyze the PPG waveform itself to distinguish true desaturation events from artifacts. One such system, built on an adaptive boosting algorithm with a reject option, was able to silence false SpOâ‚‚ alarms while ensuring that no genuine life-threatening events were missed.20PubMed Central. Reducing Pulse Oximetry False Alarms Without Missing Life-Threatening Events Robust beat-by-beat artifact detection algorithms now achieve segmentation accuracy above 96% and artifact detection accuracy above 98%, providing a cleaner signal for downstream analysis whether the goal is rhythm detection, oxygen monitoring, or alarm management.21Mathematical Problems in Engineering. A Robust Beat-to-Beat Artifact Detection Algorithm for Pulse Wave
Contactless Monitoring and Sleep Applications
The PPG principle does not strictly require a sensor touching the skin. Camera-based photoplethysmography captures the faint color changes in facial skin caused by each heartbeat using a standard video camera. A system tested during general anesthesia showed high agreement with a conventional bedside monitor, with over 88% of heart-rate correlation coefficients exceeding 0.8.22PubMed Central. Camera-Based Photoplethysmography for Measuring Heartbeat Intervals During General Anesthesia Smartphone-based approaches are also being developed for remote SpOâ‚‚ estimation from facial video, using deep-learning models to extract the subtle signals from standard phone cameras.23Scientific Reports. Innovative approaches in imaging photoplethysmography for remote blood oxygen monitoring
Sleep medicine is another frontier. PPG waveform features change during apnea and hypopnea events because the oxygen desaturation and cardiovascular stress of obstructed breathing alter the pulse shape. Machine-learning models trained on these features have achieved discrimination between apnea and hypopnea with area-under-the-curve values around 0.80 to 0.83 depending on body position, though they struggled more at distinguishing lateral from supine events, with performance falling below 0.65.24PubMed Central. Analysis of differential photoplethysmography signal patterns in apnea and hypopnea – Section: 3. Results The clinical relevance is that a wrist-worn or ring-based sensor could potentially screen for sleep-disordered breathing at home without requiring the full polysomnography setup of a sleep lab.
Newborn Screening With Pulse Oximetry
One of the most established non-rhythm applications of pulse oximetry waveform analysis is newborn screening for critical congenital heart disease. Hospitals routinely measure oxygen saturation in both a newborn’s right hand (pre-ductal) and a foot (post-ductal). A significant difference between the two suggests blood is bypassing the lungs through a heart defect. This screening is moderately sensitive for critical defects that cause low oxygen levels, though it is not designed to catch congenital heart conditions that do not produce detectable hypoxemia.25PubMed Central. Newborn Pulse Oximetry Screening for Detecting Congenital Heart Disease: Experience at a Tertiary Care Center The principle at work is still the same PPG waveform, just being read for a different clinical question: not rhythm irregularity, but an oxygen gradient across two body sites that points to abnormal cardiac plumbing.