Artifacting

What You Will Learn

Heart rate variability recordings are vulnerable to artifacts that can dramatically distort your measurements and compromise clinical decisions. In this unit, you will learn to identify artifacts in raw signals and interbeat interval (IBI) data, understand why prevention far surpasses correction, and master two primary correction approaches for handling artifacts when they occur. Whether you are conducting publication-quality research or performing routine clinical assessments, this knowledge will ensure your data accurately reflects your client's physiology.

This unit covers artifact recognition across four categories: ectopic beats, movement, environmental interference, and hardware limitations. A dedicated section explains why a real premature beat becomes an artifact only inside a normal-to-normal interval series, how to tell true ectopy from a detection error, and when correction turns into invention. You will develop practical skills for each correction approach—including the use of Kubios noise detection, beat correction, and deviation threshold settings—and learn when to apply them. By the end, you will be able to make informed decisions about data quality that support valid clinical conclusions.

HRV recordings are vulnerable to diverse artifacts that can produce interbeat interval (IBI) values that are too long or too short. The IBI, also called the NN (normal-to-normal) interval when artifacts have been removed, represents the time between consecutive R-spikes in the ECG waveform. Artifacts are false values produced by the client's body (ectopic beats), the client's actions (movement), the environment (line current), and hardware limitations (light leakage).

Artifacts in HRV data — Artifacts represent false values that do not reflect actual cardiac activity and can significantly distort HRV measurements.

The clinical stakes are high: a single artifactual IBI value in a 2-minute epoch can markedly distort time-domain indices like SDNN, the standard deviation of normal-to-normal intervals (Berntson et al., 1997). This sensitivity to outliers (i.e., extreme values) makes artifact detection and correction essential for valid HRV assessment. Consider that SDNN reflects overall autonomic function—a distorted value could lead you to overestimate or underestimate your client's regulatory capacity.

Effect of multiple artifacts in a recording segment — Multiple artifacts in a recording segment can produce dramatically distorted HRV metrics, emphasizing the importance of data quality.

Client assessment and training are only as good as your data. As the familiar saying goes, "garbage in, garbage out." This principle applies whether you are working with a veteran managing PTSD symptoms, an athlete optimizing performance, or a cardiac rehabilitation patient rebuilding autonomic resilience.

Dilbert cartoon about data quality — The importance of data quality cannot be overstated. © Scott Adams.

We recommend the "close the barn door before the horse escapes" strategy to ensure data integrity. Prevent artifacts before you record data so that you don't have to remove them later. This proactive approach—proper sensor placement, controlled recording conditions, and client preparation—saves time and produces more reliable results than even the most sophisticated correction algorithms. Please review our BVP Hardware and ECG Hardware units for a review of artifacts and strategies to mitigate them.

Prevention is always preferable to correction. A closed barn door represents proper sensor placement and recording conditions, while the escaping horse represents artifacts that contaminate your data after the fact.

Manually inspect the raw signal and IBI values to ensure that artifacts do not contaminate your measurements. You can do this within programs like Kubios Scientific. Both provide visualization tools that display the ECG or PPG waveforms and computed IBI values, allowing you to spot discrepancies between what the sensor recorded and what the software calculated.

Kubios HRV review screen — Kubios Scientific provides comprehensive tools for reviewing raw signals and identifying artifacts. The review screen allows you to examine both the ECG waveform and IBI values simultaneously.

The relationship between artifacts in the raw signal and false IBI values can be challenging for beginners to grasp. For ECG recording, the software analyzes the raw signal to detect R-spikes—the tall, sharp peaks that mark ventricular depolarization—and measures the time intervals between them. When artifacts distort this process, the resulting IBI values no longer reflect true cardiac timing.

ECG signal with R-spike detection — Kubios Scientific software detects valid R-spikes in the raw ECG signal, designated by a red plus sign, and calculates the time interval between consecutive peaks.

IBIs are stored in a time-stamped table of consecutive values, similar to an Excel spreadsheet column. The ECG waveform circled above at 29.171 milliseconds corresponds to an IBI value of 818.359 milliseconds at the bottom of the table below. Understanding this correspondence is essential for tracing artifacts from their source in the raw signal to their impact on computed metrics.

CardioPro IBI table — IBI values are stored in a time-stamped table, allowing you to identify specific beats and their corresponding intervals.

Artifacts can cause the software to "see" extra beats, resulting in abnormally short IBIs. They can also cause it to miss beats, resulting in abnormally long IBIs. These outliers stand out when you review a table of IBI values. For example, we used the color red to identify a short IBI of 274.9 milliseconds produced by an extra beat—far below the typical resting range of 600-1200 milliseconds.

Short IBI caused by extra beat — An extra beat after 00:00:32 produces an abnormally short IBI value (274.9 ms) that stands out from typical values.

Conversely, software missed a beat before 00:00:31, resulting in a long IBI of 1334.4 milliseconds. When software fails to detect an R-spike, it measures the interval from one detected beat to the next, effectively doubling (or more) the true interval.

Long IBI caused by missed beat — A missed beat before 00:00:31 results in an abnormally long IBI value (1334.4 ms) because the software measures from one detected beat to the next detected beat, skipping the undetected beat. This exceeds the normal adult resting range (600-1200 ms) and the previous IBI by more than 20%.

Short and long values stand out compared to typical adult values. Developing an eye for these outliers is a fundamental skill for any clinician working with HRV data.

Typical IBI values — Typical resting IBI values fall within a normal adult range of 600-1200 ms and don't change more than 20% across successive beats. This makes artifactual values easy to identify.

Artifacts are false IBI values produced by ectopic beats (initiated outside of the S-A node), movement, environmental interference, or hardware limitations. A single artifact can dramatically distort HRV metrics like SDNN because these measures are sensitive to outliers. Prevention through proper recording technique is always preferable to correction. Understanding the connection between raw signal distortions and computed IBI values is essential for identifying artifacts effectively.

BCIA Blueprint Coverage

This unit addresses III. HRV Instrumentation: E. Artifacting Strategies from the BCIA certification blueprint. Whether you are preparing for certification or expanding your clinical competencies, this material provides the foundation for defensible data quality practices.

Topics include Artifacting Overview, Plan A (Segment Removal and Detrending) and Plan B (Manual Correction and Detrending). Each approach suits different clinical scenarios, and knowing when to apply each one will improve both your efficiency and your data integrity.

🎧 Listen to the Full Chapter Lecture

Artifacting Overview: Understanding the Process

This section introduces the fundamental principles of artifacting and explains why certain correction approaches are inappropriate. You will learn what artifacting is, when data should be discarded, and why the "splicing" method fails to preserve data integrity.

Artifacting is the process of removing or replacing false values found within a data record. Think of it as quality control for your HRV data: you are ensuring that every value in your dataset represents a genuine cardiac event rather than noise or error. This process directly parallels the artifact rejection procedures familiar to clinicians who practice neurofeedback.

Removing artifacts concept — Artifacting involves identifying and correcting false values to ensure data integrity.

Researchers may discard the first and last minutes of data due to client adjustment to the monitoring environment and performance degradation after a long recording session. The initial adjustment period reflects the client settling into the recording situation—electrode gel stabilizing, the client becoming accustomed to the sensors, and initial anxiety subsiding. Toward the end of extended sessions, fatigue and reduced attention can affect data quality, particularly if the client shifts position or loses focus.

Person adjusting to monitoring — Clients typically need time to adjust to the monitoring environment, and the first minute of data often contains more artifacts. Graphic © Vitaly Titov/Shutterstock.com.

The Splicing Method: What Not to Do

Never remove a part of an epoch (a defined data collection period, typically 5 minutes for standard HRV assessment) and rejoin the remaining sections. This approach destroys the temporal relationship between heartbeats, which contains meaningful physiological information (Combatalade, 2010). When you cut out a section and splice the remaining parts together, you create artificial transitions that never occurred in the client's actual cardiac rhythm.

don't splice — The splicing approach may seem intuitive, but it destroys the temporal integrity of your data.

The consequences of splicing are particularly severe for frequency-domain analysis. HRV analysis algorithms expect a continuous time series with natural beat-to-beat relationships. When you splice data together, they interpret the artificial junction as a sudden physiological change, potentially creating spurious power in specific frequency bands that misrepresent your client's autonomic function.

Consider a clinician who notices a 30-second period of movement artifact in the middle of a 5-minute recording. If they simply cut out those 30 seconds and rejoin the remaining data, the resulting time series would show artificial transitions where the splice occurred. Frequency-domain analysis would be particularly affected because the algorithm expects a continuous time series with natural beat-to-beat relationships.

Artifacting is the systematic process of identifying and correcting false IBI values to ensure data integrity. The first and last minutes of recordings often require special attention due to client adjustment and fatigue. The splicing method of removing contaminated sections and rejoining remaining data is never appropriate because it destroys the temporal relationships essential for valid HRV analysis, particularly frequency-domain measures.

Ectopic Beats: When a Real Heartbeat Becomes an Artifact

This section gathers everything you need to know about ectopic beats in one place, before you choose between Plan A and Plan B. You will learn how a real heartbeat becomes an analytic artifact, how to recognize premature ventricular and atrial beats on the strip, how to separate true ectopy from detector error, how much damage a single ectopic beat does to your metrics, and where careful repair turns into invention.

An ECG shows the heart's electrical activity, but HR and HRV throw most of that picture away and keep only the timing between beats (Kligfield et al., 2007; Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology [Task Force], 1996). That simplification works well until the heart adds an early beat. Every recording produces two lists of numbers, and confusing them causes most HRV errors. The RR interval list holds the time between every detected R spike, whatever produced it, while the NN interval list holds only the time between normal sinus beats (Task Force, 1996). Software often labels both lists as RR, so an RR file can quietly contain ectopic beats, missed beats, extra markers, and mistimed markers, even though every standard HRV metric assumes you handed it the second list (Berntson et al., 1997; Lipponen & Tarvainen, 2019; Weinstein et al., 2026).

A PAC or PVC is never an artifact in the clinical record. It becomes an artifact only inside an NN series built to measure sinus variability, and that label never gives you permission to delete the beat from the raw ECG (Peltola, 2012; Task Force, 1996).

One recording therefore asks you to do two jobs at once. The clinical job asks what the heart did, and the measurement job asks how much the sinus node varied. Keep three layers of the same recording and both jobs stay honest: the raw ECG stays untouched and answers clinical questions, an annotation layer records what you or the software called each beat, and a separate NN copy holds the estimated sinus timing you send to your HRV software (Laborde et al., 2017; Peltola, 2012). That habit protects the patient as well as the statistics, because ectopic burden, the count or percentage of ectopic beats in a record, carries clinical meaning of its own (Guichard et al., 2022; Marcus, 2020). Never let cleanup erase the rhythm history of the person in front of you.

Recognizing Ectopic Beats on the Strip

Premature ventricular contractions (PVCs) originate from an ectopic focus in the ventricles rather than the sinoatrial node. The heart's normal rhythm is briefly disrupted: an early ventricular depolarization produces a short interval, followed by a compensatory pause while the cardiac electrical system resets to its normal rhythm. This produces a characteristic short-long pattern that is recognizable once you know what to look for.

ECG strip showing premature ventricular contractions — Premature ventricular contractions (PVCs). In the lower ECG panel, the ectopic beats occur prematurely and show a wide, bizarre QRS complex consistent with ventricular origin, typically 120 ms or longer, rather than normal conduction through the His-Purkinje system. The PVCs show no clearly preceding sinus P wave and are followed by secondary ST-T changes that are discordant, meaning the T wave deflects opposite the main QRS polarity. The rhythm also demonstrates a post-PVC compensatory pause before the next sinus beat.

Three features identify a PVC without any measurement at all. The beat arrives early, its QRS complex, the waveform produced by ventricular depolarization, is wide at roughly 120 ms or more, and no P wave precedes it (Marcus, 2020; Zeppenfeld et al., 2022). Because the ectopic impulse usually leaves the sinoatrial node ticking undisturbed, the next sinus impulse arrives while the ventricles are still refractory and produces nothing visible, which gives you a compensatory pause in which the short interval plus the pause add up to about two normal cycles (Marcus, 2020). Two exceptions break that arithmetic and are worth checking before you correct anything. An interpolated PVC squeezes between two sinus beats with no pause at all, and a fusion beat blends sinus and ventricular activation into a complex that looks intermediate (Marcus, 2020). Confirm that a full compensatory pause is actually present before you assume that a repair can simply redistribute two cycles worth of elapsed time.

Premature atrial complexes (PACs) are initiated by an ectopic pacemaker in the atria rather than the ventricles. Because a PAC usually resets the sinus node and produces no full compensatory pause, there is no simple time-preserving way to remove it through manual artifacting (Clinical EEG Interpretation, 2018; Mond & Haqqani, 2019). PACs are therefore more challenging to correct than the symmetrical short-long pattern that makes an isolated PVC relatively straightforward to address.

ECG strip showing premature atrial contractions — Premature atrial contractions (PACs) interrupt a regular underlying sinus rhythm with beats that arrive early, so the R-R interval suddenly shortens. Each early beat is preceded by an ectopic P wave that looks different from the surrounding sinus P waves and may distort the preceding T wave, indicating that the impulse started in the atria but outside the sinoatrial node. The QRS complexes that follow remain narrow and normal looking, which helps distinguish PACs from premature ventricular beats. After each premature atrial beat, the tracing shows a noncompensatory pause, producing the irregular spacing that raises the apparent rate when PACs are frequent.

A PAC starts in atrial tissue outside the SA node, so its P wave, the deflection representing atrial depolarization, differs in shape or direction from its neighbors, while the QRS that follows usually looks narrow and ordinary (Mond & Haqqani, 2019). A conducted PAC usually resets the sinus node, which pushes the next sinus beat later than scheduled and produces a noncompensatory pause spanning less than two normal cycles (Mond & Haqqani, 2019).

Two variations cause trouble. When the ventricular conduction system has not yet recovered, aberrant conduction gives the PAC a wide QRS that mimics a PVC, and when the atrioventricular node has not recovered, no QRS follows at all, leaving a blocked PAC whose early P wave hides inside the previous T wave (Mond & Haqqani, 2019). The blocked PAC is the one that fools software, because a detector watching only R spikes reads that hidden P wave as a dropped beat or a sinus pause (Mond & Haqqani, 2019; Peltola, 2012).

Telling True Ectopy From Detector Error

A tachogram will show you a short-long pattern immediately, but it will not tell you what caused it (Peltola, 2012; Weinstein et al., 2026). Ectopy, a missed R spike, a double-counted T wave, and a client shifting in the chair can all produce similar spikes, so look at the waveform underneath every candidate before you decide (Laborde et al., 2017; Peltola, 2012). Detector errors leave three arithmetic signatures, and each one calls for a different fix.

A missed beat leaves one interval roughly twice as long as its neighbors, a false extra marker leaves two short intervals that add up to one normal interval, and a marker placed slightly off inside the QRS keeps the beat count correct but the timing wrong (Lipponen & Tarvainen, 2019; Weinstein et al., 2026).

Movement and poor electrode contact look different again, because they usually distort several beats in a row, shift the baseline, and produce shapes that no cardiac event would make (Friesen et al., 1990; Kligfield et al., 2007).

An interval outlier is a question, not a diagnosis. The P wave, the width of the QRS complex, and the length of the pause supply the answer (Marcus, 2020; Mond & Haqqani, 2019).