A pulse oximeter works by shining two wavelengths of light through your finger (or earlobe, or forehead) and measuring how much of each wavelength gets absorbed by your blood. Oxygenated hemoglobin and deoxygenated hemoglobin absorb red and infrared light differently, and the device exploits that difference to estimate how much of your hemoglobin is carrying oxygen. The number it displays, your SpO2, is typically between 95% and 100% in a healthy person at sea level. The technology is simple enough to clip onto a fingertip, yet the physics, the biology, and the engineering behind it are full of nuances that determine when you can trust the reading and when you should not.
Two Colors of Light, Two Types of Hemoglobin
Hemoglobin is the protein in red blood cells that ferries oxygen from your lungs to your tissues. When hemoglobin is loaded with oxygen, it absorbs relatively little red light (around 660 nm) but absorbs more infrared light (around 940 nm). When hemoglobin has dropped off its oxygen and is heading back to the lungs, the pattern flips: it absorbs much more red light and less infrared. At roughly 800 nm, the two forms absorb equally, a crossover point researchers call the isobestic wavelength.1PubMed. Validation of near-infrared spectroscopy in humans A pulse oximeter contains two tiny LEDs, one red and one infrared, and a photodetector on the opposite side of the tissue. The detector measures how much of each color makes it through.
The underlying science rests on the principle that light traveling through a mixture of absorbers loses intensity in a predictable way. Red and infrared signals that have passed through blood-perfused tissue are used to estimate the oxygen saturation of arterial hemoglobin.2PubMed. The light-tissue interaction of pulse oximetry In a simplified model, if a lot of the red light is absorbed relative to the infrared, the blood contains a high proportion of deoxygenated hemoglobin, and the SpO2 reading will be low. If relatively little red is absorbed compared to infrared, most hemoglobin is oxygenated, and the reading will be high.
Isolating the Arterial Signal
Your finger contains arteries, veins, capillaries, bone, skin, and connective tissue, all of which absorb some light. The clever trick that makes pulse oximetry practical is that only the arterial blood pulsates with each heartbeat. With every pulse, the arteries expand slightly, and the light path through arterial blood lengthens just a fraction. The device isolates this tiny pulsatile component from the constant background absorption of everything else. By looking only at the rhythmic changes in light intensity, the oximeter extracts a signal that comes almost entirely from arterial blood.
The device then computes what engineers call the “ratio of ratios,” comparing the pulsatile red absorption to the pulsatile infrared absorption. That ratio is mapped to an SpO2 value using a calibration curve built from empirical data collected in healthy volunteers whose blood oxygen was simultaneously measured with an arterial blood gas draw.3PubMed Central. Calibration of Contactless Pulse Oximetry – Section: Signal Processing for All Experiments This is why pulse oximeters are not purely theoretical instruments; they are calibrated against real blood samples, and their accuracy depends on how representative that calibration population was.
How an Accidental Discovery Became Standard Equipment
The origin story of pulse oximetry is one of the best accidents in medical technology. In 1972, a Japanese engineer named Takuo Aoyagi was working at Nihon Kohden, trying to measure cardiac output by tracking the washout of an injected dye through an ear sensor. The arterial pulsations in the ear kept interfering with his dye measurements. He tried to cancel the pulse “noise” mathematically by balancing red and infrared signals, and in doing so realized that changes in oxygen saturation broke his cancellation method. The very thing that was ruining his dye experiment turned out to be a way to measure blood oxygen non-invasively.4PubMed. Takuo Aoyagi: discovery of pulse oximetry Aoyagi first presented the idea in 1974 at a biomedical engineering conference in Osaka, and by the mid-1980s, improved commercial versions from multiple manufacturers had entered hospitals worldwide.5PubMed Central. Ninety years of pulse oximetry: history, current status, and outlook Today it is hard to imagine anesthesia, emergency medicine, or critical care without it.
What the Numbers Mean for Your Body
A reading of 95% to 100% is generally considered normal at low elevations. Below 90%, clinicians start to worry. But the relationship between your SpO2 number and the actual amount of oxygen dissolved in your blood is not linear. The oxygen-hemoglobin dissociation curve is S-shaped: at higher saturations, a small drop in the partial pressure of oxygen in your blood barely nudges the SpO2 reading. Below around 90%, though, the curve steepens dramatically. A saturation of 88% corresponds to a partial pressure of roughly 60 mmHg, and even small decreases beyond that point reflect considerable tissue-level oxygen deprivation, even though the number on the screen might not look alarmingly low.6Nigerian Journal of Clinical Practice. Arterial Oxygen Saturation: A Vital Sign? – Section: Role of clinical expertise accompanied by knowledge of physiology
This means a reading of 94% falling to 91% is less alarming, physiologically, than a reading of 91% falling to 88%. The device gives you a number, but knowing where you are on that S-curve changes how urgently you should respond.
The Skin Tone Problem
One of the most important limitations of pulse oximetry came under intense scrutiny during the COVID-19 pandemic: the devices tend to overestimate oxygen saturation in people with darker skin. The light traveling through pigmented skin is absorbed differently, and the calibration curves built from predominantly lighter-skinned volunteers do not fully account for this. A systematic review and meta-analysis found that pulse oximeters overestimate true oxygen saturation in people with high skin pigmentation by a pooled mean of about 1.1%, and by about 1.5% in people identified as Black or African American.7PubMed Central. The accuracy of pulse oximetry in measuring oxygen saturation by levels of skin pigmentation: a systematic review and meta-analysis
A 1-2% overestimation might sound trivial, but at the steep part of the dissociation curve it can mask clinically meaningful hypoxemia. If the true saturation is 88% but the device reads 90%, a clinician might delay supplemental oxygen. A broader systematic review confirmed that the majority of studies reported this overestimation pattern in participants with darker skin tones.8British Journal of Anaesthesia. Effect of skin tone on the accuracy of the estimation of arterial oxygen saturation by pulse oximetry: a systematic review The U.S. FDA has since called on manufacturers to test devices across a more diverse population, and several companies are actively developing multi-wavelength sensors designed to reduce this bias. For now, clinicians caring for darker-skinned patients are advised to interpret SpO2 readings with extra caution, especially when values hover near decision thresholds.
When the Device Gets Fooled by Abnormal Hemoglobin
Standard pulse oximeters use two wavelengths, which is enough to distinguish oxygenated hemoglobin from deoxygenated hemoglobin. But there are other forms of hemoglobin that these two wavelengths cannot properly identify. The two that cause the most trouble are carboxyhemoglobin (hemoglobin bound to carbon monoxide) and methemoglobin (hemoglobin with iron oxidized into a form that cannot carry oxygen effectively).
Carbon monoxide binds to hemoglobin about 200 times more tightly than oxygen does, and carboxyhemoglobin absorbs red light in a pattern similar to oxygenated hemoglobin. So a person with severe carbon monoxide poisoning can have dangerously low functional oxygen while the pulse oximeter cheerfully reads 95% or higher. Studies of carbon monoxide poisoning patients found that SpO2 consistently overestimated true oxygen saturation, and the overestimation grew as carboxyhemoglobin levels increased.9CHEST. Accuracy of Pulse Oximetry in Severe Carbon Monoxide Poisoning This is why firefighters and smoke-inhalation victims need a blood draw, not just a finger clip.
Methemoglobinemia plays a different trick. As methemoglobin levels rise, the pulse oximeter reading drifts toward roughly 85% and then plateaus there, regardless of whether the true saturation is higher or lower.10PubMed. Effects of methemoglobinemia on pulse oximetry and mixed venous oximetry At low methemoglobin levels, the device overestimates; at very high levels, it may underestimate. Either way, the number on the screen stops reflecting reality.11Clinical Chemistry. Laboratory Assessment of Oxygenation in Methemoglobinemia Methemoglobinemia can be caused by certain medications, local anesthetics, and some industrial chemical exposures. Specialized co-oximeters that use eight or more wavelengths can distinguish these hemoglobin variants, but they are laboratory instruments, not fingertip clips.
Cold Hands, Weak Signals
Because pulse oximetry depends on detecting the tiny pulsatile expansion of arteries, anything that reduces blood flow to the measurement site weakens the signal. Cold fingers, low blood pressure, shock, and vasoconstrictor drugs all constrict peripheral blood vessels. In these conditions, the light signal becomes weak and noisy, and some devices simply fail to produce a reading. Others display inaccurate numbers without warning the user.12PubMed Central. In vivo investigation of ear canal pulse oximetry during hypothermia A systematic review confirmed that poor peripheral perfusion is one of the most significant limitations of oximeter performance.13PubMed. Accuracy of pulse oximeters in measuring oxygen saturation in patients with poor peripheral perfusion: a systematic review
The practical takeaway is straightforward: if your hands are cold or you are in a clinical situation involving significant blood loss or circulatory compromise, a fingertip reading may be unreliable. Warming your hands, ensuring the device is snug, and keeping still can all help. In hospital settings, switching to a forehead or earlobe sensor can bypass the perfusion problem. During mild hypothermia, for instance, forehead and ear oximeters maintained their response times while finger oximeters slowed significantly.14PubMed. The desaturation response time of finger pulse oximeters during mild hypothermia
Where You Put the Sensor Matters
Most people encounter pulse oximeters as fingertip devices, but clinical sensors can also go on the earlobe, forehead, toe, or even the ear canal. These sites differ in how quickly they respond to changes in blood oxygen. The finger is farthest from the heart and lungs, and blood takes longer to reach it, so there is an inherent delay. In one study of patients undergoing bronchoscopy, the ear probe detected a drop to 90% saturation about 36 seconds faster than the finger probe.15PubMed. Significant Delay in the Detection of Desaturation between Finger Transmittance and Earlobe Reflectance Oximetry Probes during Fiberoptic Bronchoscopy Half a minute might not sound like much, but in anesthesia or emergency resuscitation, it can be the difference between catching a desaturation early and catching it late.
Forehead sensors have a similar speed advantage because the skin there is supplied by the internal carotid artery, which maintains flow even when peripheral vessels constrict. In intensive care and operating rooms, forehead or ear sensors are often preferred for patients at risk of hemodynamic instability.
Motion Artifacts and the Fidgety Finger
Movement is the everyday nemesis of pulse oximetry. Shivering, tapping your fingers, or simply walking around can create oscillations in the light signal that the device mistakes for arterial pulsations. This produces erratic readings and false alarms. Engineers have developed increasingly sophisticated signal-processing algorithms to separate true pulse signals from motion noise. Techniques based on time-frequency analysis have shown meaningful improvements in accuracy when a finger is bent or pressed against the sensor.16PubMed Central. Reduction of motion artifact in pulse oximetry by smoothed pseudo Wigner-Ville distribution Other approaches use methods like singular spectral analysis to iteratively strip away motion contamination and reconstruct the clean underlying signal.17PubMed. Photoplethysmograph signal reconstruction based on a novel motion artifact detection-reduction approach. Part II: Motion and noise artifact removal
In practice, however, the simplest fix is to hold still. Hospital-grade oximeters with advanced motion-rejection algorithms perform better than cheap consumer models, but even the best device will struggle if you are vigorously shaking the hand it is clipped to.
Nail Polish, Gel Manicures, and Other Optical Obstacles
Because the light has to pass through (or reflect off) your nail bed, anything coating your nails can interfere. A systematic review found that black, blue, brown, and purple nail polish produced small but statistically significant decreases in SpO2 readings.18PubMed Central. Impact of Fingernail Polish on Pulse Oximetry Measurements: A Systematic Review The effects of lighter colors like pink and red were generally negligible. One study found that black polish caused a high proportion of failed readings altogether. Earlier research established that the degree of error correlated with how differently the polish absorbed light at the two key wavelengths (660 nm and 940 nm), explaining why darker and bluer shades cause more trouble: they absorb more red light, mimicking the signature of deoxygenated hemoglobin.19PubMed. The effect of nail polish on pulse oximetry
Gel manicures add another layer of thickness and opacity. One study found that certain gel polish colors, including black, produced wide limits of agreement with reference measurements, suggesting imprecise readings.20Singapore Medical Journal. The effects of gel-based manicure on pulse oximetry – Section: Results If you are heading to the hospital for a procedure, or monitoring yourself at home during an illness, removing polish from at least one finger is a worthwhile precaution. Alternatively, placing the sensor sideways across the finger, so the light passes through the nail bed perpendicular to the polish layer, is a workaround some clinicians use.
Smartwatches and Reflective Oximetry
The pulse oximeters built into smartwatches and fitness trackers work on the same basic principle but use a fundamentally different geometry. A fingertip clip sends light through the finger to a detector on the other side, a setup called transmissive oximetry. A wrist-worn device cannot do that; instead, it shines light into the skin and measures what bounces back. This reflective approach picks up a weaker signal because most of the light scatters or is absorbed in the tissue, and only a fraction returns to the sensor.21Mayo Clinic Proceedings: Digital Health. Accuracy of Smartwatch Pulse Oximetry Measurements in Hospitalized Patients With Coronavirus Disease 2019
Reflective oximetry is inherently noisier, and wrist sensors also contend with more motion, looser contact, and variable skin thickness. Studies comparing smartwatch SpO2 readings against arterial blood gas measurements consistently find wider margins of error than with medical-grade finger clips. These devices are fine for general fitness tracking and spotting broad trends over time, but they should not be relied upon for clinical decision-making. If your smartwatch flags a consistently low reading, the right response is to confirm with a medical-grade device, not to head to the emergency room based on the wrist reading alone.
Home Monitoring and Trending Over Time
The COVID-19 pandemic pushed pulse oximeters from hospitals into millions of homes, and clinicians quickly realized that context matters as much as any single reading. Spot checks can be misleading because of all the factors discussed above: cold fingers, motion, nail polish, poor perfusion, even anxiety-related hyperventilation, which temporarily raises SpO2. What proved more useful for patients isolating at home was tracking the trend. Guidance published during the pandemic recommended that people seek care if their overall trend in oxygen saturation was downward over hours, even if individual readings stayed above a specific threshold.22Annals of the American Thoracic Society. Pulse Oximetry for Monitoring Patients with COVID-19 at Home. Potential Pitfalls and Practical Guidance
A few practical habits improve home oximeter accuracy. Sit quietly for a couple of minutes before measuring. Use the index or middle finger, which tend to have better perfusion than the pinky. Make sure the finger is warm. Remove nail polish from the measurement finger. Take several readings a minute apart and look for consistency rather than trusting a single number. And if you live at high altitude, keep in mind that normal baselines are lower there. At 5,000 feet, a reading of 92% can be perfectly normal. The same guidance noted that thresholds for seeking care may need to be adjusted downward in communities at higher elevations.22Annals of the American Thoracic Society. Pulse Oximetry for Monitoring Patients with COVID-19 at Home. Potential Pitfalls and Practical Guidance
The Hardware Behind the Clip
Inside a pulse oximeter, the LEDs are not perfectly stable. Like all light-emitting diodes, their peak wavelength shifts slightly with temperature and with changes in driving current. One study found that the peak wavelength of a 660 nm red LED increased by about 5.5 nm as temperature rose from 0 to 50°C. A theoretical model predicted the effect on SpO2 accuracy was negligible over the temperature range found in normal clinical use.23PubMed. Temperature dependence of led and its theoretical effect on pulse oximetry Changing the LED’s intensity, however, had a larger potential effect: a tenfold increase in intensity shifted the red LED peak wavelength by about 8 nm, which modeled out to an error of roughly 2.5% SpO2, though only at very low saturations around 50%.24PubMed. The effect of varying LED intensity on pulse oximeter accuracy Manufacturers compensate for these quirks through internal calibration and drive-current regulation. Higher-end devices include temperature compensation circuitry that effectively eliminates the issue.
Pulse Oximetry in Veterinary Medicine
Pulse oximeters are not just for humans. Veterinarians use them routinely during anesthesia in dogs, cats, and horses. The physics is identical, but the practical challenges multiply. Animal tongues, ears, toes, and tails vary wildly in thickness, pigmentation, and hair coverage. A study evaluating multiple oximeter models across species found that accuracy and failure rates differed dramatically from model to model and species to species. In dogs, the devices performed reasonably well, with root-mean-square differences from arterial blood samples around 2-3%. In cats, the errors were larger, ranging from about 6% to 11%, with some models failing to produce readings up to a third of the time.25PubMed Central. An evaluation of pulse oximeters in dogs, cats and horses Horses fell somewhere in between, with highly variable failure rates depending on the model and the sensor placement site. Veterinary teams typically place sensors on the tongue during anesthesia, a site that provides strong perfusion but is obviously not an option in a conscious animal. The ear, vulva, and prepuce are common alternatives in larger species.