ECG Background

Reference: Chapter 6 of Bioelectrical Signal Processing in Cardiac and Neurological Applications (Biomedical Engineering) by Leif Sörnmo and Pablo Laguna

1. Electrical Activity of the Heart

Sinoatrial (SA) Node: Located in the right atrium, the SA node fires the initial electrical impulse. This triggers atrial contraction and starts the cardiac cycle.
Atrial Contraction: Following the SA node impulse, both atria contract to push blood into the ventricles. This is reflected as the P wave on an ECG.
Atrioventricular (AV) Node: The impulse travels to the AV node, a relay station between the atria and ventricles. Here, there's a brief delay to allow the ventricles time to fill with blood.
Bundle of His and Bundle Branches: The delayed impulse moves from the AV node down the Bundle of His and into the right and left bundle branches toward the ventricles.
Purkinje Fibers: The impulse reaches the Purkinje fibers, which spread throughout the ventricles, causing them to contract forcefully. This is represented by the QRS complex in an ECG.
Ventricular Contraction: The ventricles contract, pushing blood into the aorta and pulmonary arteries. The force of this contraction is essential for effective blood circulation.
Ventricular Repolarization: After contraction, the ventricles reset or "repolarize." This phase is shown as the T wave on an ECG.
Resting Phase: Following repolarization, the heart muscle briefly rests before the next cardiac cycle begins. This is the flat line following the T wave and before the next P wave on an ECG.
Return to SA Node: The cycle restarts with a new impulse from the SA node.

2. ECG Recording Techniques

Lead

In the context of an ECG, a "lead" refers to an electrical pathway between two electrodes. The lead captures the heart's electrical activity from a specific angle, allowing for a multi-dimensional view of what's going on in the heart. The way leads are configured can differ and are broadly categorized into unipolar and bipolar types.

Unipolar Lead

Unipolar leads measure electrical activity between a single active electrode and a reference (also called the central terminal), often an average of multiple other electrodes. They provide information from one perspective and are generally used in modern 12-lead ECGs for the limb leads and precordial (chest) leads.

Bipolar Lead

Bipolar leads use two active electrodes to measure the electrical activity between them. This type of lead gives a more localized view of electrical activity and is often used in the original limb leads I, II, and III.

Einthoven's Triangle

Einthoven's Triangle is a model that describes the relationship between three bipolar limb leads (I, II, III). These leads are placed on the right arm, left arm, and left leg, forming a triangle around the heart. The model is instrumental in understanding how different leads capture electrical activity from various angles.

einthovens_triangle

Standard 12-lead ECG

The standard 12-lead ECG is the most commonly used ECG configuration. It includes three bipolar limb leads (I, II, III), three unipolar limb leads (aVR, aVL, aVF), and six unipolar precordial or chest leads (V1-V6). These 12 leads provide a comprehensive view of the heart's electrical activity from different angles.

Orthogonal Leads

Orthogonal leads are leads placed at right angles to each other, often used in vectorcardiography. They are typically labeled as X, Y, and Z and give a three-dimensional representation of the heart's electrical activity. While not commonly used in standard clinical settings, they can provide additional data for specific diagnostic needs.

Synthesized Leads

Synthesized leads are generated using data from other existing leads, usually in specialized ECG setups. For example, a 12-lead ECG system might synthesize additional leads like V3R or V4R for a more detailed view of certain heart regions. These aren't directly measured but are calculated from the available data for enhanced diagnostic value.

3. Heart Rhythms

3.1 Sinus Rhythm

Sinus rhythm refers to the normal heart rhythm generated by the sinoatrial (SA) node, which is the heart's natural pacemaker. The term "sinus" indicates that the electrical activity of the heart is originating from the SA node, as it should in a healthy individual.

Heart Rate and BPM

The normal rate for a sinus rhythm can vary but is generally between 60 and 100 beats per minute (BPM). If the heart rate falls below 60 BPM, it's called bradycardia. Conversely, if it goes above 100 BPM, it's termed tachycardia.

Autonomic Influence

The heart rate isn't fixed; it varies due to the influence of the autonomic nervous system. Increased parasympathetic activity will slow the heart rate down, while increased sympathetic activity will speed it up. This constant push-pull between the two arms of the autonomic nervous system results in small fluctuations in heart rate, commonly referred to as heart rate variability (HRV).

Clinical Implications of Decreased HRV

A decrease in heart rate variability could signify underlying cardiac or autonomic system issues. It's not just about the rate itself but the ability of the heart to adapt to different conditions that is a marker for health.

Respiratory Phases and Heart Rate

In healthy individuals, the heart rate varies with the phases of respiration. It usually increases during inspiration and decreases during expiration.

Provocation Tests

ECGs can also provide information on the autonomic function through provocation tests. One such test is the orthostatic test, where the subject is tilted from a lying to standing position. Another is the deep breathing test, which involves hyperventilation, and yet another is the Valsalva test, where the person tries to exhale with their nose and mouth closed. These tests can reveal how the heart responds to changes in position or stress, offering further insight into cardiac and autonomic health.

3.2 Premature Beats

Premature beats are unexpected heartbeats that interrupt the normal sinus rhythm. They occur earlier than the next anticipated sinus beat. These beats can originate from various locations within the heart, not just the SA node, and are known as ectopic foci.

Types of Premature Beats

Premature beats can have a normal or abnormal shape and may or may not be accompanied by a P wave. They are categorized based on their origin: - Supraventricular Premature Beat (SVPB): Originates above the ventricles, usually in the atria or the AV node. SVPBs often have an abnormal P wave. - Ventricular Premature Beat (VPB): Originates in the ventricles.

Compensatory Pauses

VPBs are often followed by a compensatory pause, a brief period where the heart skips a beat to get back on rhythm. SVPBs, on the other hand, usually don't have a full compensatory pause. The time between the two normal beats surrounding an SVPB is often shorter than two regular RR intervals.

Variations

Multiform VPBs: Occur when there are multiple ectopic foci in the ventricles, leading to premature beats with different shapes.
Bigeminy and Trigeminy: In bigeminy, a premature beat follows every normal beat. In trigeminy, a premature beat follows every second normal beat.

Interpolated VPBs

When the SA node's rate is slow and a VPB occurs extremely early, the next normal sinus beat may still happen on time. This is because the AV node and ventricles have recovered and are not in a refractory state. Such VPBs are termed "interpolated" and aren't followed by a compensatory pause but are surrounded by two short RR intervals.

premature_beats

(a) A supraventricular premature beat shown with a small, negative P wave, seen after the third regular heartbeat.
(b) A ventricular premature beat followed by a pause to get back on rhythm. The total time of the two adjacent RR intervals nearly equals two normal RR intervals.
(c) Bigeminy: A pattern where every normal beat is followed by a premature beat.
(d) Trigeminy: A pattern where a premature beat follows every second normal beat.
(e) An interpolated ventricular premature beat appearing after the third regular beat, without disrupting the normal rhythm.

3.3 Atrial Arrhythmias

Atrial arrhythmias originate from abnormal electrical activity in the atria, often due to one or multiple ectopic foci. These arrhythmias manifest in two main ways on the ECG:

Abnormal P waves occur when the ectopic focus is far from the SA node, essentially reversing the normal electrical propagation. In these cases, P waves can overlap with QRS complexes, making them hard to distinguish.
Complete absence of P waves occurs in cases of extreme abnormalities.

Two specific types of atrial arrhythmias are atrial tachycardia and atrial flutter/fibrillation: - Atrial Tachycardia: Results from fast pacemaking cells in the atria, causing P waves to often get masked by other waveform components. - Atrial Flutter and Atrial Fibrillation: Both are tachyarrhythmias where the atria beat much faster than the ventricles, usually due to a reentry circuit in the atria (electrical impulse circulates through a closed-loop pathway within the atrial tissue instead of following the normal conduction pathway). They risk causing blood clots, potentially leading to strokes. - Atrial Flutter: More organized, with atria beating around 300 bpm. Not all impulses reach the ventricles due to a refractory AV node, resulting in a ventricular rate around 150 bpm. ECG shows "F waves" that look like a sawtooth. - Atrial Fibrillation: Highly chaotic, with atrial rates between 400-700 bpm, making the atria quiver. The ventricles beat irregularly and rapidly, and ECG shows "f waves" that are multiform and irregular.

atrial_tachyarrhythmias

(a) Atrial flutter
(b) Atrial fibrillation

3.4 Ventricular Arrhythmias

Ventricular arrhythmias mainly result from the reentry mechanism, similar to their atrial counterparts. However, they manifest differently on an ECG and can be much more dangerous. Let's look at the key types:

Ventricular Tachycardia: This kicks in at rates over 120 beats/minute. The beats look like premature ventricular beats, featuring wide QRS complexes and large amplitudes. P waves usually get lost in the mix because the ventricular cycles happen back-to-back. A minimum of three consecutive abnormal beats is needed for this to be considered ventricular tachycardia.
Ventricular Flutter: Here, the rhythm is organized but very rapid, with no clear QRS complexes or T waves showing up on the ECG. It's like atrial flutter but on a bigger, fluctuating scale. This can progress into ventricular fibrillation.
Ventricular Fibrillation: This is chaotic and disorganized, causing the ventricles to quiver rather than pump blood effectively. If not treated immediately, this leads to cardiac arrest and is usually fatal. On an ECG, it looks similar to ventricular flutter but is more chaotic.

To sum it up, ventricular arrhythmias are severe and often life-threatening, requiring immediate medical intervention. They may share a similar underlying mechanism (reentry) with atrial arrhythmias, but their ECG manifestations and clinical implications are quite different.

ventricular_tachyarrhythmias

(a) Ventricular flutter
(b) Ventricular fibrillation

3.5 Heart Blocks

Heart blocks occur when there's a disruption in the normal electrical pathways of the heart, leading to issues in depolarization and repolarization. These blocks can severely hamper the heart's function. Let's focus mainly on AV node-related blocks, which are categorized based on severity:

Minor AV Block: All electrical impulses make it from the atria to the ventricles, but they're delayed. The heart still functions, just not as efficiently as it should.
Moderate AV Block: Some impulses don't reach the ventricles. This means that not every atrial contraction is followed by a ventricular contraction, which can lead to an irregular heartbeat and potentially serious complications.
Complete AV Block: No impulses from the atria reach the ventricles. On an ECG, you'll see P waves (from the SA node) and QRS complexes (from a ventricular ectopic focus) occurring at totally different, independent rates. This is a severe condition requiring immediate medical attention.

Other types of conduction blocks can also occur, such as in the left or right bundle branches, each with their own sets of complications and treatments.

4. Heartbeat Morphologies

ECGs can reveal a range of abnormal beat shapes, highlighting issues like heart enlargement (hypertrophy) or inflammation of the heart's covering (pericarditis). They're also key for diagnosing arrhythmias. This section zeroes in on two crucial conditions: myocardial ischemia and myocardial infarction. Both can cause rapid, drastic changes in beat morphology.

4.1 Myocardial Ischemia

This condition occurs when there's not enough blood flowing to the heart's cells, usually because coronary arteries are narrowed. The result is a shortage of oxygenated blood, which can lead to:

Angina Pectoris: Chest pain or discomfort often triggered by exercise or stress.
Silent Ischemia: Ischemia without the chest pain, but still dangerous due to the risk of severe arrhythmias like ventricular fibrillation.

ECG Indicators

ST-T Change: In an ECG, ischemia often shows up as a shift in the ST segment and T wave.
A normal ST segment curves up into the T wave, but an ischemic ST segment might be flat or slope downward, sometimes even starting below the baseline (known as ST depression).
The T wave in ischemia might flatten, turn biphasic, or go negative.
T Wave Alternans: This is a subtle, beat-to-beat flip-flop pattern in the T wave. It's a warning sign for lethal arrhythmias and requires special signal processing for detection.

Diagnostic Tests

Exercise Stress Test: Those suspected of having angina or silent ischemia may undergo this test to provoke an ST-T reaction, helping in diagnosis.
Ambulatory ECG Monitoring: In this, a portable ECG device is worn for 24 hours or more to catch episodes of ischemia.

4.2 Myocardial Infarction

Myocardial infarction, commonly known as a heart attack, occurs when there's a sudden, sustained loss of blood supply to an area of the heart. This is usually due to a blocked coronary artery, often as a result of plaque rupture and subsequent blood clot formation. Like silent ischemia, a heart attack also increases the risk of dangerous ventricular fibrillation.

Outcomes and Recovery

Thanks to advances in medical treatment, many patients now survive myocardial infarctions. Although the heart's pumping ability is compromised, enough healthy muscle usually remains to manage the workload, making near-full recovery possible.

ECG Findings

The area affected by the infarction becomes electrically inactive, which disrupts the heart's normal electrical pathways. This alters what's known as the dominant vector, changing the appearance of the ECG waves. Notably:

A decrease in the R wave amplitude.
The appearance of a pathological Q wave, larger than normal.
ST segment elevation or depression.

In a healthy heart, the ST segment corresponds to the period when the ventricles are depolarized and the myocardial cells are at a uniform electrical state. But if a part of the myocardium is injured and can't fully repolarize, the electrical baseline for those cells is shifted. When this occurs, it manifests as ST elevation on the ECG.

Changes Over Time

As the heart heals, the ECG also changes. In the weeks and months following a myocardial infarction:

Initial ST-T segment elevation usually disappears within about 10 days.
A large Q wave often appears and can persist for 6 months or more.

5. Noise and Artifacts

Accurate ECG analysis relies heavily on good signal quality, which can often be compromised by various types of noise and artifacts. Understanding these factors is crucial before diving into methods designed to mitigate them. Here are some common culprits:

Baseline Wander

This low-frequency activity can disrupt the ECG, making it difficult to interpret. Factors like perspiration, respiration, and body movement can contribute to baseline wander. Because its frequency content is often below 1 Hz, it can usually be tackled effectively through signal processing.

Electrode Motion Artifacts

Changes in skin impedance around the electrode can produce this noise. Unlike baseline wander, these artifacts are more challenging to remove as their frequency content overlaps with the actual ECG signal. This issue is particularly problematic in ambulatory ECG monitoring.

Powerline Interference (50/60 Hz)

Caused by improper grounding or other equipment, this type of noise is often removable via linear or non-linear filtering.

Electromyographic Noise (EMG)

Muscle contractions can introduce EMG noise, especially during exercise or ambulatory monitoring. This noise shares frequency components with the ECG signal, making it hard to remove without distorting the underlying signal.

Respiratory Activity

Breathing not only affects the heart rate but also alters the ECG morphology. For instance, respiration can cause variations in the amplitude of the QRS complex. Although this is usually considered noise, it can be used to estimate the respiratory frequency under certain conditions.

noise and artifacts

(a) Baseline wander
(b) Electrode motion artifact
(c) EMG noise
(d) Respiratory activity

6. Clinical Applications

ECGs serve multiple clinical purposes today, far beyond being a single test. Computer processing has become essential for enhancing the accuracy of these various applications.

6.1 Resting ECG

This test is a go-to diagnostic tool for a wide range of suspected diseases, not just cardiac ones. Conducted with the patient in a resting, supine position, the test usually records a standard 12-lead ECG for 10 seconds. This limited time makes the resting ECG most useful for diagnosing persistent heart issues rather than transient arrhythmias.

Computer Interpretation

Computers usually interpret resting ECGs, extracting various measurements about waveform morphology and rhythm. The quality of resting ECGs allows for even the small P waves to be analyzed, which is crucial for categorizing atrial arrhythmias correctly. However, crafting software that mimics the comprehensive interpretation skills of a human physician is a significant challenge. Though automated systems are highly accurate, their results are usually reviewed by a doctor to confirm the diagnosis.

Database Storage

After interpretation, the ECG data are stored for future reference. This allows for serial ECG analysis, where multiple recordings from the same patient are compared over time to monitor changes, such as those due to myocardial infarction. However, the reliability of such serial comparisons can be compromised by nonphysiological factors, like different electrode placements or changes in the heart's position.

6.2 Intensive Care Monitoring

In intensive care units (ICU) or coronary care units (CCU), continuous ECG monitoring is crucial. This is especially true for patients recovering from myocardial infarction or heart surgery. The central computer in these units receives ECG data from multiple beds.

Objectives

The primary goal is to quickly detect life-threatening arrhythmias like ventricular fibrillation or signs of acute myocardial ischemia, indicated by ST segment changes. Among all ECG applications, ICU/CCU monitoring is the only one that absolutely requires real-time signal processing. Quick detection of a serious event, like cardiac arrest, is essential for immediate life-saving interventions.

Challenges

However, the setting is not ideal. High noise levels and artifacts, often due to muscle activity or position changes, plague the ECG signals. This poor signal quality leads to false alarms, which in turn create distractions and added workload for the ICU staff. This can be especially problematic when such distractions cause real cardiac events to be overlooked.

Signal Quality

Given that a patient may stay in ICU/CCU for an extended period, improving signal quality is imperative. The trend toward continuous monitoring of the full 12-lead ECG—instead of a subset of leads—compounds this challenge. More leads can mean more chances for noise and artifacts, making the need for high-quality signal processing even more critical.

6.3 Ambulatory Monitoring

Ambulatory ECG monitoring, commonly known as Holter monitoring, provides a way to record ECG data while the patient goes about their daily activities. Unlike the resting ECG, which is a brief snapshot, Holter monitoring can span 24 hours or more.

Use Cases

This method is particularly useful for patients with transient symptoms like palpitations, light-headedness, or syncope. It's also used for patients at high risk of sudden death after a heart attack, or those whose responses to antiarrhythmic drugs need evaluation. Typically, a 3-lead ECG is used as it's more practical for long-term monitoring than a 12-lead setup.

Data and Storage

Digital storage needs for a 24-hour, 3-lead ECG are substantial. Although modern technology can handle this, data compression remains relevant for database storage and later analysis.

Analysis and Challenges

After recording, the ECG data is computer-analyzed for arrhythmias, but manual checks are essential to filter out artifacts. In the past, much focus was on classifying beat morphologies, particularly Ventricular Premature Beats (VPBs), as they were thought to be a significant risk factor for sudden death. While this is no longer considered as crucial, VPBs are still important for assessing heart rate variability, which has shown promise in predicting mortality rates after myocardial infarction.

Another application of Holter monitoring is in detecting silent ischemia through ST segment analysis. However, the analysis is often complicated by noise and artifacts, especially in identifying atrial arrhythmias, where the P waves can be masked.

6.4 Stress Test

Exercise stress testing is a diagnostic approach that puts the heart under physical stress to assess its ability to cope with increased workload. The procedure is commonly used to diagnose conditions like angina pectoris and to identify undiagnosed episodes of silent ischemia.

Procedure

The test starts at a low level of physical activity, usually on a treadmill or bicycle, and gradually increases in intensity. A standard 12-lead ECG is continuously recorded during the exercise and is subject to real-time signal processing. This processed ECG, often along with other metrics like blood pressure and respiratory rate, is monitored on-screen to provide real-time feedback to the physician.

End Conditions and Assessments

The test is generally terminated when the patient feels fatigued, experiences symptoms like chest pain or shortness of breath, or shows abnormal ECG changes. The subsequent recovery period is also monitored to assess if the ECG returns to its baseline state.

Signal Processing and Challenges

Because the ECG during exercise is subject to noise like baseline wander and EMG noise, accurate signal processing is crucial. Ensemble averaging is a common technique used to improve the ECG signal quality. Exponential averaging or similar recursive algorithms are often used to track exercise-induced changes in the ST segment accurately. Although Ventricular Premature Beats (VPBs) are not the focus in stress tests, they must be identified and excluded from the averaging process, as averaging typically only involves sinus beats.

Interpretation

In a healthy subject, any ST depression observed during the test should largely resolve during the recovery period. In contrast, in a patient with myocardial ischemia, ST depression may persist or even worsen during recovery, indicating a likely problem with blood supply to the heart muscle.

6.5 High-Resolution ECG

High-resolution ECG is a significant advancement over conventional ECG techniques, enabling the detection of much smaller signals in the order of microvolts. This is achieved mainly through signal averaging techniques, which enhance the signal-to-noise ratio.

Acquisition Procedure

The general procedure is similar to a standard resting ECG but extended over a longer time. The idea is to capture enough heartbeats for effective averaging, thereby reducing noise and allowing smaller-amplitude signals to be analyzed.

Timing and Averaging

Unlike evoked potentials where external stimulus timing is known, the exact timing or "fiducial point" for each heartbeat must be accurately determined before ensemble averaging. The need for precise timing is crucial to avoid distortion of low-amplitude, high-frequency ECG components. Additionally, high-resolution ECG assumes a consistent beat-to-beat morphology, unlike signal averaging during stress tests, which has to adapt to gradual changes in the signal.

Sampling Rate

Given that high-resolution ECG often contains high-frequency components, a high sampling rate of at least 1 kHz is typically used. This is in contrast to other ECG applications where lower sampling rates might suffice.

Areas of Focus

In high-resolution ECG, several specific intervals of the cardiac cycle are examined:

Bundle of His: Depolarizes during the PR segment. This segment, often considered silent in standard ECGs, can be analyzed.
Terminal QRS and ST segment: Places where late potentials may be present, particularly interesting in the context of myocardial infarction.
Intra-QRS Potentials: An area of growing interest for researchers.
P Wave: Also targeted for more detailed analysis.

Clinical Importance

Among these, the analysis of late potentials has gained the most attention. Late potentials can be a sign of delayed and fragmented ventricular depolarization, especially in patients with myocardial infarction. The presence of late potentials is often indicative of higher risks of life-threatening arrhythmias and is significant for risk stratification in post-infarction patients.