Common Artifacts in ECG: Causes and Identification

ECG artifacts are unwanted signals on an electrocardiogram tracing that do not originate from the heart’s electrical activity. They range from gentle undulations that obscure the baseline to rapid, chaotic deflections that can look indistinguishable from life-threatening arrhythmias. Some are caused by the patient’s own body, such as breathing or shivering, while others stem from equipment problems, electrode placement errors, or even the digital filters built into the ECG machine itself. Recognizing these imposters is a core clinical skill because mistaking an artifact for a real cardiac event can lead to unnecessary and invasive treatment.

Baseline Wander

Baseline wander is the slow, rolling drift of the ECG tracing up and down over several seconds. Instead of a flat line between heartbeats, you see a gradual wave that makes it hard to judge whether the ST segment or other features are truly elevated or depressed. The most common cause is simply breathing: each inhalation and exhalation shifts the chest wall and the electrodes sitting on it, introducing a low-frequency oscillation into the signal.¹ Changes in the electrical contact between the skin and the electrode and larger body movements contribute as well.²

In practice, baseline wander tends to be most pronounced when a patient is breathing heavily, is anxious, or is lying in an uncomfortable position that causes them to shift repeatedly. It can also appear when gel on the electrodes has dried out or when adhesive patches have partially peeled away, because both situations increase the impedance between the electrode and the skin. For visual interpretation, the drift needs to be removed, since a cardiologist evaluating ST-segment changes, say during a stress test or in a suspected heart attack, cannot make a reliable call if the entire baseline is rolling like an ocean swell.

Powerline Interference

The thin, regular buzz that appears as a uniform sawtooth pattern superimposed on the ECG is powerline interference, sometimes called 50 or 60 Hz noise depending on the local electrical grid. Modern ECG amplifiers are designed to reject this kind of interference, but residual contamination still creeps in frequently.³ The problem gets worse when electrode cables run near power cords, when fluorescent lights are overhead, or when the patient is connected to other electrical devices like infusion pumps or warming blankets. Poorly grounded outlets and long, unshielded lead wires amplify the effect.

Because the interference sits at a fixed frequency, it can usually be spotted easily: the tracing looks “fuzzy” with a regular, fine oscillation that persists at the same rate regardless of heart rate. Notch filters built into most machines can suppress it, but aggressive filtering can also clip real cardiac signals whose frequency content overlaps with the powerline frequency, so the tradeoff is not always clean.

Muscle and Movement Artifacts

Skeletal muscles generate their own electrical signals, and those signals are picked up by ECG electrodes just as readily as the heart’s. When a patient shivers from cold, shakes during a fever, or simply tenses their arms or legs, the tracing can become erratic. The resulting pattern is an irregular, high-frequency disturbance that “bombards” the baseline with apparently random spikes and deflections.⁴

The clinical danger here is real. Muscle tremors can produce deflections that closely mimic serious cardiac rhythms, including ventricular tachycardia and atrial flutter.⁵ Chest compressions during CPR generate large-amplitude oscillations, and even a patient who fidgets or moves a limb during a recording can create sudden baseline irregularities that look like premature contractions.⁴ The key giveaway is that muscle artifact tends to affect all leads simultaneously and does not follow the orderly morphology of a true cardiac rhythm, but that distinction is not always obvious in a fast clinical situation.

Electrode and Skin-Contact Problems

The interface between the electrode and the skin is more complicated than it might seem. It is not just a matter of good adhesion. When friction or movement disturbs the thin electrochemical layer at the electrode surface, the reduction-oxidation equilibrium at that interface breaks down and generates noise.⁶ In other words, motion artifact is not solely about the electrode jiggling on the skin; the chemistry at the contact point itself becomes a source of electrical noise. Dried gel, oily skin, excessive body hair, and poor electrode adhesion all worsen the problem.

Skin preparation before electrode placement can make a meaningful difference. Gently abrading the skin removes dead cells and oils that increase impedance and reduce signal quality. Shaving small patches where electrodes will sit and using fresh, properly gelled electrodes are among the simplest and most effective steps. These measures do not eliminate motion artifact entirely, but they lower the threshold at which minor patient movement corrupts the signal.

Lead Misplacement

An ECG depends on electrodes being in exactly the right anatomical positions. When they are not, the tracing can look abnormal even though the heart is perfectly healthy. Two categories of misplacement cause the most trouble: limb-lead reversal and precordial-lead misplacement.

Limb-Lead Reversal

Swapping two limb electrodes changes the electrical axis the machine calculates, sometimes dramatically. The good news is that most limb-lead reversals produce predictable axis shifts, often accurate to within about five degrees from the expected pattern, which gives a trained reader a reliable clue that the leads are mixed up rather than that the patient has had a sudden cardiac event.⁷ The classic example is a right-arm/left-arm swap, which inverts lead I. A reader who notices that lead I is entirely flipped relative to what is expected should suspect reversal before suspecting pathology.

Precordial-Lead Misplacement

Precordial leads V1 and V2 are frequently placed too high on the chest, and this seemingly minor error can change the ECG enough to suggest disease that is not there. When V1 and V2 sit an intercostal space or two above their correct position, the tracing may show patterns that mimic an incomplete right bundle branch block, anterior T-wave inversions, or ST-segment elevation. These features can falsely suggest acute or old cardiac ischemia, pulmonary embolism, or even a Brugada pattern, a genetic condition with a risk of sudden cardiac death.⁸ A clinician who acts on these false patterns could order unnecessary catheterizations, imaging, or hospital admissions.

When Artifacts Mimic Dangerous Arrhythmias

Perhaps the most consequential artifact problem is when a non-cardiac signal fools a clinician into diagnosing a lethal rhythm. ECG artifacts can mimic a variety of serious conditions, including acute myocardial infarction and ventricular tachycardia.⁹ A physician survey found that artifact mimicking ventricular tachycardia frequently resulted in patients being subjected to unnecessary invasive cardiac procedures, including defibrillation, antiarrhythmic drug infusions, and catheter-based interventions.¹⁰ The recommendation from that study was straightforward: clinicians should include artifact in their differential diagnosis whenever they see a wide complex tachycardia on the monitor.

Tremor-induced artifacts are a particularly common culprit. In one reported case, a patient with Parkinson’s disease had an ECG initially read as atrial fibrillation, prompting a hospital admission. The irregular baseline caused by the patient’s resting tremor was misinterpreted as fibrillatory waves.¹¹ In another case, a patient’s bilateral upper-extremity tremor produced a pattern that the ECG machine itself labeled as atrial flutter. When the patient’s arms were immobilized, the tracing instantly reverted to normal sinus rhythm, and repeated tracings confirmed the pattern was entirely artifact.¹² Stories like these reinforce a basic principle: if the patient looks stable and comfortable but the monitor suggests a crisis, take a breath and recheck the leads before reaching for the crash cart.

Equipment, Devices, and the Hospital Environment

The hospital setting is full of electronic equipment, and much of it can corrupt an ECG signal. Electrical devices used in clinical settings can induce artifacts through several different mechanisms, and newer diagnostic and therapeutic technologies may generate their own artifactual changes.¹³ Operating rooms and intensive care units are the worst offenders, because patients are simultaneously connected to ventilators, electrocautery devices, warming systems, and multiple monitoring platforms. Electrocautery, or “the Bovie,” is notorious for producing large-amplitude bursts that completely overwhelm the ECG signal and may trigger false arrhythmia alarms. These artifacts are nonspecific but can also resemble serious rhythm disturbances.

Loose or frayed lead wires and broken cables also introduce intermittent noise that can be maddeningly difficult to track down. A cable with an internal break may produce normal tracings most of the time but inject spikes or sudden dropouts whenever it shifts position. Beyond loose leads and broken wires, implanted devices like neurostimulators and certain pumps can add their own electrical signals to the mix.⁹

Pacemaker Spikes on the Tracing

Patients with implanted pacemakers present a unique artifact challenge. The pacemaker delivers a brief electrical pulse to stimulate the heart, and that pulse shows up on the ECG as a sharp vertical spike, usually just before the QRS complex or the P wave it is pacing. In older unipolar pacemakers, these spikes were large and easy to see. Contemporary bipolar pacemakers deliver much smaller stimuli, typically in the range of two to four volts with pulse durations frequently at or below 0.25 milliseconds.¹⁴

This creates a problem for digital ECG machines. To reliably capture such a narrow spike, the machine’s front-end sampling rate needs to be at least 10,000 samples per second. If the sampling rate is lower, the spike may be missed entirely, making the paced rhythm look like an intrinsic one and hiding the fact that the patient depends on a device. Conversely, some machines artificially amplify the spike to make it more visible, but doing so can distort the surrounding ECG morphology and confuse automated interpretation algorithms. Guidelines from major cardiology societies have emphasized that manufacturers should maintain high enough sampling rates to catch narrow pacemaker pulses without inflating their amplitude.¹⁴

How Digital Filters Create Their Own Artifacts

Ironically, the very filters designed to clean up an ECG can introduce new distortions. Every ECG machine applies some level of digital filtering to suppress noise, and the settings matter more than most clinicians realize.

A large simulation study evaluating different baseline-wander removal techniques found that while all tested methods were better than leaving the wander unfiltered, none could reconstruct the original ECG without modifying the ST segment to some degree.¹⁵ This means that even the best algorithm for removing drift will slightly alter the portion of the tracing most critical for diagnosing a heart attack. Among the methods tested, a standard Butterworth high-pass filter offered the best balance of speed and accuracy for clinical use, but the researchers were clear that users should always exercise caution when diagnosing ST changes on a filtered tracing.

High-pass filter frequency settings deserve particular attention. When the high-pass cutoff is set higher, as is common on telemetry monitors that default to 0.5 Hz rather than the 0.05 Hz used in diagnostic mode, significant ST-segment elevation can appear where none actually exists. One study found that a 1.0 Hz high-pass filter independently increased the odds of false-positive ST elevation of one millimeter or more by nearly tenfold in lead V1 and roughly fivefold in V2.¹⁶ The distortion was concentrated in the right precordial leads, with V2 showing the largest absolute increase, and limb leads were barely affected. In a clinical scenario, a patient being monitored on telemetry could have their tracing flag for a possible acute coronary event purely because of the filter setting, triggering a code and a catheterization lab activation for a problem that does not exist.

The practical lesson is that telemetry tracings and diagnostic 12-lead ECGs are not interchangeable. When ST-segment changes are the question, a standard diagnostic-mode ECG with lower filter settings should be obtained before making treatment decisions.

Automated Detection and Signal Quality Scoring

As ECG monitoring moves toward continuous, ambulatory, and wearable formats, the volume of data generated far exceeds what a human can review in real time. This has driven development of automated methods that score signal quality on the fly. One approach extracts features from the mathematical structure of the ECG signal and feeds them into a machine-learning classifier, producing a continuous signal quality index that flags segments likely contaminated by artifact.¹⁷ Rather than simply accepting or rejecting a segment, these tools assign a probability that a given stretch of recording is clean, allowing downstream algorithms, or a clinician reviewing the data later, to weight their confidence accordingly.

These methods are especially important for wearable devices used outside clinical settings. A patient wearing a patch monitor at home may generate hours of unusable tracing while exercising, showering, or sleeping in a position that displaces the sensor. Without automated quality scoring, an arrhythmia-detection algorithm might either miss a real event buried in noise or flag dozens of false positives from artifact, both of which erode the clinical value of the recording.

Practical Steps for Reducing Artifacts

Most artifact problems can be minimized with attention to a handful of fundamentals during the recording process. Good skin preparation, including cleaning the skin with alcohol and lightly abrading it, reduces the impedance at the electrode-skin interface and makes the signal more stable.⁶ Using fresh electrodes with adequate gel, shaving excess body hair, and making sure the adhesive is firmly seated are equally basic and equally effective measures.

Patient comfort matters more than it might seem. A tense, cold, or uncomfortable patient will shiver, fidget, or grip the side rails, and all of that ends up on the tracing. Warming the room, offering a blanket, reassuring the patient, and asking them to relax their arms at their sides can quiet the signal immediately. For patients with Parkinson’s or essential tremor, gently supporting or immobilizing the limbs during recording is sometimes the only way to obtain a usable tracing, as demonstrated in the case where arm immobilization instantly resolved a pseudo-atrial flutter pattern.¹²

On the equipment side, keeping cables away from power cords, checking for frayed or damaged wires, and ensuring proper grounding can suppress powerline interference. Knowing what filter settings the machine is using, and switching from monitoring mode to diagnostic mode when ST-segment analysis matters, prevents filter-induced distortion from masquerading as ischemia. And for any tracing that looks alarming, the simplest and most underused intervention remains the same: look at the patient. If they are alert, comfortable, and hemodynamically stable while the monitor shows ventricular tachycardia, the odds strongly favor artifact over arrhythmia.

1
PubMed Central. A hierarchical method for removal of baseline drift from biomedical signals: application in ECG analysis
2
Turkish Journal of Electrical Engineering and Computer Sciences. Upper envelope detection of ECG signals for baseline wander correction: a pilot study
3
PubMed Central. Removal of power-line interference from the ECG: a review of the subtraction procedure – Section: Background
4
PubMed Central. Main artifacts in electrocardiography
5
PubMed Central. Electrocardiogram artifact caused by rigors mimicking narrow complex tachycardia: a case report – Section: Discussion
6
Biomedical Signal Processing and Control. A new method to reduce motion artifact in electrocardiogram based on an innovative skin-electrode impedance model – Section: Conclusion
7
PubMed. Simple diagnosis of limb-lead reversals by predictable changes in QRS axis – Section: CONCLUSIONS
8
PubMed. Misplacing V1 and V2 can have clinical consequences
9
PubMed. Electrocardiographic artifact
10
PubMed. Physician interpretation of electrocardiographic artifact that mimics ventricular tachycardia – Section: CONCLUSIONS
11
PubMed Central. Electrocardiographic artifact potentially misleading to the wrong management
12
PubMed Central. Global Pseudo-Atrial Flutter on Electrocardiogram and the Importance of Clinical Correlation – Section: Abstract
13
Anesthesiology. Equipment-related Electrocardiographic Artifacts: Causes, Characteristics, Consequences, and Correction
14
PubMed Central. Detection and removal of pacing artifacts prior to automated analysis of 12-lead ECG – Section: Discussion
15
PubMed Central. Comparison of Baseline Wander Removal Techniques considering the Preservation of ST Changes in the Ischemic ECG: A Simulation Study – Section: Conclusion
16
PubMed. Quantifying high-pass filter-induced ST-segment distortion – Section: Results
17
PubMed Central. Artefact detection and quality assessment of ambulatory ECG signals – Section: BACKGROUND AND OBJECTIVES