From OpenWetWare
Jump to navigationJump to search


Home        Lab Members        Physiological Systems        Monitoring Parameters in the ICU        ECG        HRV        Cardiorespiratory Monitors        Signal Processing        Deliverables        Journal        Abbreviations       



An ECG is a recording of the electrical signals produced by the heart[1]. There are several different types of ECG monitors, their differences mainly being how many leads they operate with, and how long they monitor an individual for. The ECG (Fig 1a) consists of a P wave, QRS complex and T wave which are caused by atrial depolarisation, ventricular depolarisation and ventricular repolarisation respectively[2]. Variations in ECG readings can be used to detect arrhythmias, coronary heart disease, heart attacks and cardiomyopathy[3], although this proves to be challenging. A study conducted by the Heart Rhythm Society (HRS) found that less than 1/4 of 800 physicians distinguished the length of all QT intervals correctly[4].

A typical ECG recording (Friederich,Patrick)[5]

Figure 1a: A typical ECG recording[5]

Abnormal ECG recordings and corresponding disease type [2]

Figure 1b: Abnormal ECG recordings and corresponding disease type[2]

Electrical Activity of the Heart

The first discovery of electrical activity in the heart was made in 1856 by Rudolf Albert von Kolliker and Heinrich Muller[2]. The heart is an involuntary muscle, meaning that cardiac muscle cells contract according to the action potentials (APs) generated by the pacemaker cells. In a normal person, the sinoatrial (SA) node, located in the right atrium, initially depolarises and APs spread throughout the atria via gap junctions between cells, causing them to contract first[6]. The action potentials are then conducted through internodal pathways to the atrioventricular (AV) node, and finally to the bottom of the ventricles through the bundle of His[6]. The relatively slow conduction of the APs to the ventricles ensures that there is a delay between the contraction of the atria and the ventricles and that the ventricles contract from bottom upwards to ensure maximum ejection[2].

The parts/waves of the ECG in relation to the sequence of cardiac excitation, where the yellow areas indicate depolarisation

Figure 2: The parts/waves of the ECG in relation to the sequence of cardiac excitation, where the yellow areas indicate depolarisation[6]

Parts and Features of the ECG

ECG Waves

The ECG (Fig 3a) consists of a P wave, QRS complex and T wave which are caused by atrial depolarisation, ventricular depolarisation and ventricular repolarisation respectively[2]. Atrial repolarisation cannot be seen on the ECG as the current is too small. The length of the PR interval is the time taken for APs to reach the ventricles from the atria, in other words the AV node delay, and the ST segment represents the uniform depolarisation and rapid ejection of the ventricles, which corresponds to the plateau of the ventricular APs[2]. The QT interval measures the duration of depolarisation and repolarisation of the ventricles [6]. Normalising the QT interval with HR gives the corrected QT (QTc), where the formula is: QTc=(measured QT)/√(RR interval)[7]

A typical ECG recording with normal wave amplitudes and intervals
Figure 3a: A typical ECG recording with normal wave amplitudes and intervals[8]

ECG Intervals

Table 1 shows the range of interval values in a healthy patient, which are affected by age and gender. A study by Uygur et al. shows how the ECG changes in children with age, where there is a decrease in HR, but an increase in P duration, PR interval and QRS duration[9]. PR and QT intervals also increase with age in adults, however there is a decrease in QRS duration[10]. Although data from Rijnbeek shows that there is consistently longer QRS duration for boys, Dickinson suggests that this is not significant[11]. In adults, however, Rijnbeek et al. shows that ST segment and Q wave duration are shorter in women[12].

Age Group PR Interval/s QRS Complex/s corrected QT (QTc) Interval/s
Adult[13] 0.12-0.22 <0.12 <0.45 (males), <0.47 (females)
Children[14] 0.11-0.21 0.05-0.07 <0.49 (infants), <0.44
Table 1: Normal ranges of ECG intervals in different age groups

Other ECG Features

Wave amplitude and axis are also commonly analysed features of ECGs. Bachman et al. provides evidence that the R, S and T wave amplitudes decrease with age and a leftward shift of the frontal plane axis[10]. R, S and T wave amplitudes also differ with gender[12].

ECG Lead Placement

The first ECG was recorded in 1887 by Augustus Waller and was improved by Willem Einthoven, for which he won a Nobel prize in 1924[15]. Basic ECGs are recorded using 3 leads (see Fig 4a) attached to the wrists and ankles for minimal interference from the contracting muscles[2]. Current cardiographs (see Fig 2) and most cardiorespiratory monitors incorporate 12 leads. The augmented limb, or Goldberger's, leads are positioned 30° between leads I, II and III and improve ischaemia/infarction detection[16]. The placement of leads (see Table 2) can significantly change the recorded ECGs, thus inaccuracy in lead placement could cause misdiagnosis of patients[17].

Leads I, II and III are placed at 60° to each other, forming Einthoven's triangle [19]
Figure 4a: Leads I, II and III are placed at 60° to each other, forming Einthoven's triangle[18]

Placement of leads V1-V6 [20]
Figure 4b: Placement of leads V1-V6[19]

Placement of leads table [20]
Table 2: The standard placement of leads, adapted from Vander's Human Physiology[6]

Different ECG leads give different views of the electrical activity of the heart[20] (Fig 4c). For example, the largest deflection for the P wave is seen in lead II[21], whereas the Q wave is normally observed from leads I, aVL, V5 and V6[22]. Thus, ECG lead placement is important in obtaining correct ECG waveforms for the diagnosis of patients. The ECG pattern varies according to the position and polarity of the electrodes on the chest[23]. Despite this, mistakes due to incorrect lead placement or lead reversal are still prevalent among clinicians, which could lead to unnecessary or even harmful treatment for patients or undetected conditions[24][25] Incorrect cable connections occur in up to 4% of 12-lead ECG recordings, with the greatest proportion found in the ICU[26].

ECG leads record different waveforms (A and B show separate times of recording) [47]
Figure 4c: ECG leads record different waveforms (A and B show separate times of recording)[27]

Clinical Importance of ECG

ECGs are used in hospitals for the following reasons [28]:

  • Looking for the cause of chest pain
  • Identifying irregular heartbeats
  • Determining overall health of heart before procedures, for example surgery
  • Tracking heart health after treatment for conditions such as myocardial infarction or endocarditis
  • Used to obtain a baseline of regular heart function to compare with future examinations

Variations in frequency, amplitude and shape are examined to diagnose issues. It is important to note that age, sex and race can also affect the morphology of an ECG [29].


The main function of ECGs is to aid clinicians in identifying cardiorespiratory-related risks and diseases in patients. Abnormal ECG morphologies indicate problems with the electrical pathways in the heart, from which clinicians can infer the patient's condition.

Some examples of changes in the morphology which can indicate a possible medical condition include[30]:

  • Changes in the shape of P wave – mitral or pulmonary stenosis, this is where the mitral or pulmonary valves become narrower meaning less blood is pumped through the heart.
Figure 6.1a: 12-Lead ECG showing morphology typical of pulmonary stenosis[31].
  • Changes in the interval between P and R waves- first degree block atrioventricular block, this is where conduction through the atrioventricular node is delayed, so there is a delay in the contraction of the ventricles.
  • Loss of R wave amplitude- myocardial infarction, which is a heart attack, this occurs when blood through to the heart in greatly reduced causing damage to heart muscle.
Figure 6.1b: Section of a 12-Lead ECG showing morphology typical in myocardial infarction[32]
  • Increased frequency of the entire PQRST segment- tachycardia, the heart is beating to fast generally classified with a resting rate is above 100bpm.

Cardiac-related Diseases

Cardiac events are the main cause of death for postoperative patients and highlight the importance of patient monitoring in the ICU[33]. A greater risk of coronary heart disease (CHD) is linked to major and minor ECG abnormalities in the older population[34].

PR Interval

The PR interval represents the AV node delay. As such, prolongation of the PR interval implies a defect in the conduction of APs from the atria to the ventricles, which can result in varying degrees of heart block[2]. One of the most common cardiac conditions is arrhythmia, which is an irregular heart rhythm. Ectopic beats are spontaneously generated by ventricular cells and can be observed from an early broad QRS complex which is not preceded by a P wave[2].

Figure 6.2.1: Prolonged PR interval and P:QRS ratio is greater than 1 in third-degree heart block[35]
Atrial arrhythmia

Tachycardia is a condition where the HR is above 100bpm[36]. Atrial tachycardia results from the atria contracting ineffectively and having insufficient time for relaxation and blood filling, thus affecting P waves[2]. AV nodal re-entry tachycardia is the most common form of tachycardia and is seen from narrow QRS complexes[2].

Figure 6.2.2a: Multifocal Atrial Tachycardia can be identified with a fast HR and a minimum of 3 different P waveforms [37]

Similarly, atrial flutter is caused by the passing of the excitations in RA in an anticlockwise direction, rather than to the LA and AV node[2]. Atrial flutter is characterised by an ECG with a saw-tooth pattern (Fig 6.2.2b)[2].

Figure 6.2.2b: Atrial flutter has a distinct saw-tooth pattern and varied rhythm[38]

Atrial fibrillation is a form of arrhythmia and is the most common heart rhythm disturbance[39]. ECGs show small, irregular f waves instead of P waves which result from the occasional transmission of APs to the ventricles as the excitation spreads throughout the atria.

Ventricular arrhythmia

Ventricular tachycardia may arise from ischaemia and cardiomyopathy. P waves appear infrequently on the ECG[2]. Ventricular fibrillation (Fig 6.2.3) is the most dangerous cardiac arrhythmia, as the uncoordinated excitations in the ventricle result in no cardiac output and death within minutes. The ECG is irregular and random[2].

Figure 6.2.3: The ECG for ventricular fibrillation has no fixed pattern, P waves or PR intervals as excitations are uncoordinated[40]
ST Segment and ischaemia

Respiratory-related Diseases

Chronic Obstructive Pulmonary Disease (COPD)

ECGs are also used in the monitoring of respiratory diseases, as cardiac complications are common. For COPD patients, the destruction of lung tissue eventually leads to the failure of the right side of the heart[41]. ECG monitoring is important as cardiac failure is the main cause of death for COPD[41]. The RR interval, from which HR can be calculated, is one of the main parameters monitored for COPD exacerbations[41]. Other common ECG changes include late R wave progression, Q wave abnormality and changes seen in arrhythmias[41].

Figure 6.3.1: ECG of an emphysema patient with sagging PR and ST intervals[42]

Covid-19 is another respiratory disease with various CVD implications[43][44][45]. Covid-19 infects the body through angiotensin-converting enzyme 2 (ACE2), which is found mostly in the lungs and moderately in the heart[45]. The severity and morbidity of Covid-19 is higher for patients with existing CV diseases[43]. Multiple drugs used for the treatment of Covid-19, such as hydroxychloroquine (HCQ), azithromycin and antivirals, cause QT prolongation which could result in ventricular arrhythmias and Torsades des Pointes[45][46]. The American Heart Association has released a guideline for ECG monitoring of patients treated with these drugs, while local hospitals have defined recommended clinical pathways for discontinuation of drug therapy as in Fig 6.3.2b[46]. In addition, there is a positive correlation between QT prolongation and ICU admission and intubation[46].

Figure 6.3.2a: ECG of a Covid-19 patient which shows ST elevation (leads I and aVL), ST depression (lead aVR), J-Point elevation and T-wave inversion (leads II, III and aVF)[47]

Figure 6.3.2b: Situation Background Assessment Recommendation (SBAR) tool used in clinical management of Covid-19 patients. QT prolongation is considered to be significant if the patient's QTc interval >0.47s for QRS duration <0.12s or QTc interval >0.5s for QRS duration >0.12s[46]

Relationship with HRV

This section gives a brief discussion of how ECGs and HRV are related. For more in-depth information on HRV, please visit the HRV tab.

RR Interval

The duration between two consecutive R peaks of the QRS complex is measured as the RR interval[48]. RR intervals are normally 0.6-1.2s under resting conditions[49] but change in different environments. HRV is defined as the variability in RR intervals (Figure 2), with age and gender as factors[50]. Longer ECG signals, such as 24h Holter recordings (Figure 7.1b), are observed to determine HRV[51].

Figure 7.1a: Calculation of HRV from the ECG under normal and stress conditions[52]

HRV HolterMonitor.jpg
Figure 7.1b: Holter monitor which records ECGs for up to 24h[53]

HRV Interpretation

The autonomic nervous system (ANS) controls HRV via the sympathetic and parasympathetic nervous system, which increase and decrease HR respectively[54]. A high HRV indicates good health as the body responds to inputs from both systems and can adapt easily to stimuli[55], but may also be caused by atrial fibrillation[56]. HRV is reduced in conditions such as COPD[57] and cardiomyopathy[58].

HRV Analysis

Time and frequency domain analyses of HRV use signal processing toolkits and statistical methods[56][59][60][61]. Difficulties in HRV analysis include the accurate detection of R peaks, irregular rhythms and noise[61]. A signal processing methodology could be developed to overcome challenges in ECG and HRV analysis by collecting large datasets from healthy individuals and patients. Machine learning could be potentially be used to link ECG/HRV features to clinical conditions[62][63].

Hrv ectopicBeat.jpg
Figure 7.3: HRV error due to ectopic beats[64]


  1. NHS, “Electrocardiogram.” (accessed Nov. 14, 2020).
  2. 2.00 2.01 2.02 2.03 2.04 2.05 2.06 2.07 2.08 2.09 2.10 2.11 2.12 2.13 2.14 2.15 N. Herring and D. J. Paterson, Levick’s Introduction to Cardiovascular Physiology, 6th ed., vol. 1. Milton, 2018.
  3. J.-L. Vincent, “Understanding cardiac output,” Crit. Care, vol. 12, no. 4, p. 174, 2008, doi: 10.1186/cc6975.
  4. I. Rowlandson, “How to Detect Long QT in a Heartbeat.” (accessed Nov. 14, 2020).
  5. P. Friederich, “Monitoring of the QT Interval in Perioperative and Critical Care,” 2020. (accessed Nov. 14, 2020).
  6. 6.0 6.1 6.2 6.3 E. P. Widmaier author, Vander’s human physiology : the mechanisms of body function, Fifteenth. 2019.
  7. I. Birnbaum, Y. Birnbaum, and G. N. Levine, “Electrocardiography,” in Cardiology Secrets, Elsevier, 2017, pp. 21–28.
  8. “U Wave • LITFL Medical Blog • ECG Library Basics.” (accessed Jan. 20, 2021).
  9. Ö. Uygur and A. Aydoğdu, “Normal electrocardiogram values of healthy children,” Turkish Arch. Pediatr. Pediatr. Arşivi, vol. 54, no. 2, p. 93, 2019, doi: 10.14744/TURKPEDIATRIARS.2019.04568.
  10. 10.0 10.1 S. Bachman, D. Sparrow, and L. K. Smith, “Effect of aging on the electrocardiogram,” Am. J. Cardiol., 1981, doi: 10.1016/0002-9149(81)90081-3.
  11. D. F. Dickinson, “The normal ECG in childhood and adolescence,” Heart, vol. 91, no. 12, p. 1626, 2005, doi: 10.1136/HRT.2004.057307.
  12. 12.0 12.1 P. R. Rijnbeek et al., “Normal values of the electrocardiogram for ages 16-90 years,” J. Electrocardiol., 2014, doi: 10.1016/j.jelectrocard.2014.07.022.
  13. “Overview of the ECG Waves, Deflections, Intervals, Durations – ECG & ECHO.” (accessed Dec. 21, 2020).
  14. “Paediatric ECG Interpretation • LITFL Medical Blog • ECG Library Basics.” (accessed Dec. 21, 2020).
  15. S. S. Barold, “Willem Einthoven and the birth of clinical electrocardiography a hundred years ago,” Card. Electrophysiol. Rev., vol. 7, no. 1, pp. 99–104, Jan. 2003, doi: 10.1023/A:1023667812925.
  16. “The ECG leads: electrodes, limb leads, chest (precordial) leads, 12-Lead ECG (EKG) – ECG & ECHO.” (accessed Dec. 22, 2020).
  17. R. Rajaganeshan, C. L. Ludlam, D. P. Francis, S. V. Parasramka, and R. Sutton, “Accuracy in ECG lead placement among technicians, nurses, general physicians and cardiologists,” Int. J. Clin. Pract., 2008, doi: 10.1111/j.1742-1241.2007.01390.x
  18. J. Feher, “The Electrocardiogram,” in Quantitative Human Physiology, Elsevier, 2012, pp. 537–546.
  19. “ECG Lead positioning • LITFL Medical Blog • ECG Library Basics.” (accessed Dec. 22, 2020).
  20. A. L. Goldberger, Z. D. Goldberger, and A. Shvilkin, “ECG Leads,” in Goldberger’s Clinical Electrocardiography, 2018.
  21. S. Meek and F. Morris, “Introduction. I—Leads, rate, rhythm, and cardiac axis,” BMJ, 2002, doi: 10.1136/bmj.324.7334.415.
  22. S. Meek and F. Morris, “Introduction . II — Basic terminology QRS complex,” Br. Med. J., 2002.
  23. A. L. Goldberger, Z. D. Goldberger, and A. Shvilkin, “ECG Leads,” in Goldberger’s Clinical Electrocardiography, Elsevier, 2013, pp. 16–25.
  24. R. Rajaganeshan, C. L. Ludlam, D. P. Francis, S. V. Parasramka, and R. Sutton, “Accuracy in ECG lead placement among technicians, nurses, general physicians and cardiologists,” Int. J. Clin. Pract., 2008, doi: 10.1111/j.1742-1241.2007.01390.x.
  25. “MisLeading: The clinical implications of misplaced ECG leads - JEMS.” (accessed Dec. 26, 2020).
  26. V. N. Batchvarov, M. Malik, and A. J. Camm, “Incorrect electrode cable connection during electrocardiographic recording,” Europace, vol. 9, no. 11. Oxford Academic, pp. 1081–1090, Nov. 01, 2007, doi: 10.1093/europace/eum198.
  27. K. Ishibashi, M. Takeda, Y. Yamahara, H. Shiraishi, T. Shirayama, and H. Matsubara, “Brugada syndrome associated with J waves in multiple leads and ‘pseudo-epsilon’ wiggle waves in lateral leads: Possible conduction delay in J-wave syndrome,” J. Arrhythmia, vol. 28, no. 4, pp. 228–231, 2012, doi: 10.1016/j.joa.2011.11.002.
  28. “Electrocardiogram | Johns Hopkins Medicine.” Dec. 13, 2020)
  29. “What is an electrocardiogram (ECG)? -Heart Matters magazine.” Dec. 13, 2020)
  30. “ECG interpretation: Characteristics of the normal ECG (P-wave, QRS complex, ST segment, T-wave) –ECG & ECHO.” Dec. 13, 2020)
  31. M. Alpaslan, “pulmonary stenosis and the electrocardiography -ekg -ecg -ankara kardiyoloji -kalp hastalıkları -mete alpaslan” Dec. 13, 2020)
  32. “prwpexamples1.jpg (PNG Image, 559 × 184 pixels).” Dec. 13, 2020).
  33. J. G. Ramsay, “Cardiac management in the ICU,” 1999, doi: 10.1378/chest.115.suppl_2.138S.
  34. R. Auer et al., “Association of major and minor ECG abnormalities with coronary heart disease events,” JAMA - J. Am. Med. Assoc., 2012, doi: 10.1001/jama.2012.434.
  35. “Understanding Atrioventricular Blocks | CareerCert.” (accessed Dec. 27, 2020).
  36. NIH, “Arrhythmia.” (accessed Nov. 14, 2020).
  37. “Multifocal Atrial Tachycardia (MAT) • LITFL • ECG Library Diagnosis.” (accessed Dec. 26, 2020).
  38. “Atrial Flutter training - ACLS Cardiac Rhythms video | ProACLS.” (accessed Dec. 26, 2020).
  39. NHS, “Atrial Fibrillation.” (accessed Nov. 14, 2020).
  40. “ACLS Ventricular Fibrillation and Pulseless Ventricular Tachycardia Guide.” (accessed Dec. 26, 2020).
  41. 41.0 41.1 41.2 41.3 T. I., T. N., T. R., K. M., and K. T., “Continuous remote monitoring of COPD patients—justification and explanation of the requirements and a survey of the available technologies,” Med. Biol. Eng. Comput., 2018.
  42. “ECG in Chronic Obstructive Pulmonary Disease • LITFL.” (accessed Dec. 25, 2020).
  43. 43.0 43.1 T. J. Guzik et al., “COVID-19 and the cardiovascular system: implications for risk assessment, diagnosis, and treatment options,” Cardiovasc. Res., vol. 116, no. 10, pp. 1666–1687, Aug. 2020, doi: 10.1093/cvr/cvaa106.
  44. J. P. Lang, X. Wang, F. A. Moura, H. K. Siddiqi, D. A. Morrow, and E. A. Bohula, “A current review of COVID-19 for the cardiovascular specialist,” Am. Heart J., vol. 226, pp. 29–44, 2020, doi:
  45. 45.0 45.1 45.2 R. B. Azevedo et al., “Covid-19 and the cardiovascular system: a comprehensive review,” J. Hum. Hypertens., 2020, doi: 10.1038/s41371-020-0387-4.
  46. 46.0 46.1 46.2 46.3 S. Jain et al., “Enhanced electrocardiographic monitoring of patients with Coronavirus Disease 2019,” Hear. Rhythm, 2020, doi: 10.1016/j.hrthm.2020.04.047.
  47. M. Rehman, A. Gondal, and N. U. Rehman, “Atypical Manifestation of COVID-19-Induced Myocarditis,” Cureus, vol. 12, no. 6, Jun. 2020, doi: 10.7759/cureus.8685.
  48. P. A. Lanfranchi and V. K. Somers, “Cardiovascular Physiology: Autonomic Control in Health and in Sleep Disorders,” in Principles and Practice of Sleep Medicine: Fifth Edition, Elsevier Inc., 2010, pp. 226–236.
  49. “Normal Electrocardiography (ECG) Intervals: Normal Electrocardiography Intervals.” (accessed Jan. 03, 2021).
  50. S. Moodithaya and S. T. Avadhany, “Gender differences in age-related changes in cardiac autonomic nervous function,” J. Aging Res., vol. 2012, 2012, doi: 10.1155/2012/679345.
  51. G. A. Lanza, P. Pedrotti, A. G. Rebuzzi, V. Pasceri, G. Quaranta, and A. Maseri, “Usefulness of the addition of heart rate variability to Holter monitoring in predicting in-hospital cardiac events in patients with unstable angina pectoris,” Am. J. Cardiol., vol. 80, no. 3, pp. 263–267, Aug. 1997, doi: 10.1016/S0002-9149(97)00343-3.
  52. “ECG and HRV of a subject under normal and high cognitive condition. | Download Scientific Diagram.” (accessed Jan. 03, 2021).
  53. “Holter Monitor | Johns Hopkins Medicine.” (accessed Dec. 10, 2020).
  54. M. Campos, “Heart rate variability: A new way to track well-being,” 2017. (accessed Nov. 15, 2020).
  55. M. van Deusen, “Everything You Need to Know About Heart Rate Variability (HRV),” 2019. (accessed Nov. 14, 2020).
  56. 56.0 56.1 F. Shaffer and J. P. Ginsberg, “An Overview of Heart Rate Variability Metrics and Norms,” Front. Public Heal., vol. 5, p. 258, Sep. 2017, doi: 10.3389/fpubh.2017.00258.
  57. C. A. Camillo et al., “Heart rate variability and disease characteristics in patients with COPD,” Lung, vol. 186, no. 6, pp. 393–401, Dec. 2008, doi: 10.1007/s00408-008-9105-7.
  58. L. C. M. Vanderlei, C. M. Pastre, R. A. Hoshi, T. D. de Carvalho, and M. F. de Godoy, “Basic notions of heart rate variability and its clinical applicability,” Brazilian J. Cardiovasc. Surg., vol. 24, no. 2, pp. 205–217, 2009, doi: 10.1590/s0102-76382009000200018.
  59. M. Bolanos, H. Nazeran, and E. Haltiwanger, “Comparison of heart rate variability signal features derived from electrocardiography and photoplethysmography in healthy individuals,” in Annual International Conference of the IEEE Engineering in Medicine and Biology - Proceedings, 2006, vol. 2006, pp. 4289–4294, doi: 10.1109/IEMBS.2006.260607.
  60. S. Sieciński, P. S. Kostka, and E. J. Tkacz, “Heart rate variability analysis on electrocardiograms, seismocardiograms and gyrocardiograms on healthy volunteers,” Sensors (Switzerland), vol. 20, no. 16, pp. 1–16, Aug. 2020, doi: 10.3390/s20164522.
  61. 61.0 61.1 M. J. Reed, C. E. Robertson, and P. S. Addison, “Heart rate variability measurements and the prediction of ventricular arrhythmias,” QJM, vol. 98, no. 2, pp. 87–95, Feb. 2005, doi: 10.1093/qjmed/hci018.
  62. G. T. Taye, E. B. Shim, H.-J. Hwang, and K. M. Lim, “Machine Learning Approach to Predict Ventricular Fibrillation Based on QRS Complex Shape,” Front. Physiol., vol. 10, p. 1193, Sep. 2019, doi: 10.3389/fphys.2019.01193.
  63. E. Agliari, A. Barra, O. A. Barra, A. Fachechi, L. Franceschi Vento, and L. Moretti, “Detecting cardiac pathologies via machine learning on heart-rate variability time series and related markers,” Sci. Rep., vol. 10, no. 1, p. 8845, 2020, doi: 10.1038/s41598-020-64083-4.
  64. D. Nabil and F. Bereksi Reguig, “Ectopic beats detection and correction methods: A review,” Biomed. Signal Process. Control, vol. 18, pp. 228–244, 2015, doi: