ECG Artifacts Research Articles

0061 Does Combining M1 M2 Reference Influence Amplitude of Slow Waves?

Abstract Introduction Slow wave amplitudes are critical to determining Stage N3 sleep yet ECG artifact frequently interferes with accurate amplitude measurement. This artifact may be lessened by using a combined M1-M2 reference however theoretically this may decrease the amplitude due to shorter inter-electrode distance (predicted 27% loss). The AASM Scoring Manual recommends scoring slow wave activity using F4-M1 channel or alternatively F3-M2, but does not recognize a combined reference. This study measures the differences in slow wave amplitude using contralateral versus combine reference. Methods 12 polysomnograms were randomly selected for analysis of amplitude of slow wave using contralateral and combined reference channels. Six separate EEG channels (F3-M1, F3-M2, F3-M1+M2, F4-M1, F4-M2, and F4-M1+M2) were used to analyze 25 different slow waves from each polysomnogram. Individual slow waves from Stage N3 sleep were analyzed using the Natus Sleepworks Amplitude Measurement Tool if their peak and trough were free EKG artifact. Averages and standard deviations of the waveforms were calculated for each patient and channel. Differences were normalized by dividing by the amplitude of the original wave using the contralateral reference. Results Subjects age ranged from 30–69 yrs, with 6 being females. Mean amplitudes were as follows: F3-M2 was 131.75µV, F3-M1+M2 125.84 µV, F4-M1 130.57 µV, and F4-M1+M2 128.22µV. The overall average difference of F4-M1 to F4-M1+M2 was 0.92% and the average difference of F3-M2 to F3-M1+M2 was 3.52% with the average standard deviation of 8.47%. Conclusion This study shows the average loss in amplitude of converting F4-M1 to F4-M1+M2 was less than 1% and 3.5% for F3-M2 to F3-M1+M2. Combining M1M2 reference may be a valuable alternative to reduce EKG artifact. Support None

Engineering approaches for ECG artefact removal from EEG: a review

Engineering approaches for ECG artefact removal from EEG: a review

Removal of ECG artifacts from EMG signals with different artifact magnitudes by template subtraction

AbstractThis work investigates the performance of eventsynchronous noise cancelling algorithms to separate ECG artifacts from thoracic EMG recordings. In focus are the precise detection of heart beats, the exact time alignment of the QRS segments and the construction of a template QRS signal. Among the utilized methods is a modified structural intensity (SI) approach based on the observation of extrema in low pass filtered versions of the second derivative as a detector. An adoption of this approach is also used to obtain the exact time alignment of QRS segments. Artificial test signals are constructed based on addition of ECG and sEMG with different magnitudes. We show that approaches based on the observation of extrema in different scales yield superior results.

Automatic ECG Artefact Removal from EEG Signals

Abstract Electroencephalography (EEG) signals are frequently contaminated by ocular, muscle, and cardiac artefacts whose removal normally requires manual inspection or the use of reference channels (EOG, EMG, ECG). We present a novel, fully automatic method for the detection and removal of ECG artefacts that works without a reference ECG channel. Independent Component Analysis (ICA) is applied to the measured data and the independent components are examined for the presence of QRS waveforms using an adaptive threshold-based QRS detection algorithm. Detected peaks are subsequently classified by a rule-based classifier as ECG or non-ECG components. Components manifesting ECG activity are marked for removal, and then the artefact-free signal is reconstructed by removing these components before performing the inverse ICA. The performance of the proposed method is evaluated on a number of EEG datasets and compared to results reported in the literature. The average sensitivity of our ECG artefact removal method is above 99 %, which is better than known literature results.

An Improved Approach of DWT and ANC Algorithm for Removal of ECG Artifacts

An Improved Approach of DWT and ANC Algorithm for Removal of ECG Artifacts

EKG artefakati u urgentnoj medicini

Artefacts are technical errors on electrocardiogram that may be present in all leads or in some of them or in just one. They may be subtle with no interference with correct interpretation, but they can also be significant enough to make ECG recording absolutely useless and delay diagnose and timely therapy. On the other hand, artefacts can be misinterpreted and that way patient can be exposed to unnecessary or risky therapeutic procedures. There are two groups of ECG artefacts: the first group originates from external causes and the second group originates from patient's body. The first group can be caused by electromagnetic interference (alternating current from electrical installation, electrical devices in/on patient, light) or by malfunction of the cables and/or electrodes (damaged cables and electrodes, their movement during transport, wrong connecting, 'offset signals', detached electrodes, gel lacking or inappropriate filters). The second group comes from muscles, skin and adnexa. Muscle contractions are often the cause of artefacts, voluntary movements as much as involuntary. Skin and adnexa can cause ECG artefacts in two ways: through poor contact with electrodes or through simple stretching of the epidermis which generates electric potential (turning of the patient in bed for example). Rare form of artefact caused by radial artery pulsations directly under the electrode was also described; it is so called electromechanical association. Artefacts can present in different ways on ECG recording. The ECG line can be thick, irregular, discontinued, wandering or pseudo isoelectric. Artefacts can imitate any of the ECG waves, extra systoles, tachycardia of narrow or wide complexes, long QT interval, and elevation/depression of the ST segment or T wave changes. Removal of the artefacts requires manipulating with patient, auxiliary equipment and ECG machine itself.

Muscle fasciculation detected by ECG

A 79-year-old man presented to hospital with scald burns to the perineum after a syncopal episode while in a hot bathtub. Admission ECG was misdiagnosed as possible ventricular fibrillation with...

Efficient procedure to remove ECG from sEMG with limited deteriorations: Extraction, quasi-periodic detection and cancellation

Abstract The interpretations of the surface electromyography (sEMG) signals from the trunk region are strongly distorted by the heart activity (ECG), especially in case of low-amplitude EMG responses analyses. Many methods have been investigated to resolve this nontrivial problem, by using advanced data processing on the overall sEMG recorded signal. However, if they reduce ECG artifacts, those cancellation methods also deteriorate noiseless parts of the signal. This work proposes an original ECG cancellation method designed to limit the deterioration of sEMG information. To do that, the proposed techniques does not directly attempt to remove the ECG, but is based on two main steps: the localization of ECG and the cancellation of ECG but only where heart pulses have been detected. The phase of localization efficiently extracts the ECG contribution by combining the discrete wavelet transforms (DWT) and the method of independent component analysis (ICA). And finally, this phase takes advantage of quasi-periodic properties of ECG signals to accurately detect pulses localization with an original algorithm based on the fast Fourier transform (FFT). Intensive simulations were achieved in terms of relative errors, coherence and accuracy for different levels of ECG interference. And the correlation coefficients computed from the paraspinal muscles EMG signals of 12 healthy participants were also used to evaluate the developed method. The results from simulation and real data demonstrate that the proposed method accurately detects pulses positions and efficiently removes the ECG from EMG signals, even when both signals are strongly overlapped, and greatly limits the deterioration of the EMG.

ECG Mishaps, Artifacts and Operator Faults

Electrocardiography is one of the most frequent bedside investigations in clinical practice. Often they are performed and interpreted by less experienced personnel with inadequate expertise.Equipment settings, proper technique and standardizations are prerequisites which may be overlooked, leading to costly and dangerous misinterpretations. This article highlights operational faults in ECG and emphasises that clinical staff must be equipped with necessary knowledge to identify these mistakes. It also tries to classify the common errors arising due to artifacts, electrode misplacement and improper standardization along with providing possible remedial measures. An awareness of these often neglected aspects and incorporating them in clinical discussions would not only enhance ECG interpretation but also improve patient safety.Bangladesh Crit Care J March 2017; 5(1): 63-67

P106: Does training with a modified high-fidelity manikin improve junior residents’ ability to establish transcutaneous pacing in an advanced cardiovascular life support course?

Introduction: Transcutaneous cardiac pacing (TCP), a skill taught in Advanced Cardiovascular Life Support (ACLS) courses, is recommended to treat unstable bradycardia. Training manikins currently available fail to reproduce key features of TCP and might be suboptimal to teach this procedure.The objective of this study was to measure the impact of a modified high-fidelity manikin on junior residents’ TCP competency during an ACLS course. We hypothesized that the use of this high-fidelity manikin improves junior residents’ performances. Methods: This prospective cohort study was conducted at the Université de Montréal in July 2015 and 2016. First-year residents undergoing their mandatory ACLS course were enrolled. The control group (2015) received the traditional curriculum, which includes hands-on teaching on Advanced Life Support manikins. The intervention group (2016) received a similar curriculum, but used a modified high-fidelity manikin that reproduces key features of TCP (e.g. use of multifunction pads, TCP induced patient twitching, ECG artifacts). Cohorts were tested with a simulation scenario requiring TCP. Performances were graded based on six critical tasks: turns on pacer function, applies multifunction pads, recognizes TCP is ineffective, achieves captures, verifies mechanical capture and prescribes sedation. Our primary outcome was successful use of TCP defined as having completed all tasks. Secondary outcomes were the success rates for each task. These were compared using Pearson’s chi-squared test. We anticipated that the success rate of TCP would increase from 20% to 50%. To obtain a power of more than 90%, 48 participants were needed in both cohorts. Results: A total of 50 residents were recruited in both cohorts. No resident that received the traditional curriculum was able to successfully establish TCP while 18 residents trained on the modified high-fidelity manikin succeeded (0 vs 36%, P<0.001). Furthermore, the latter were more likely to recognize when pacing was inefficient (12 vs 86%, P<0.001), obtain ventricular capture (2 vs 48%, P<0.001), and check for a pulse rate to confirm capture (0 vs 48%, P<0.001). Conclusion: Successful use of TCP is a difficult skill to master for junior residents. A modified high-fidelity manikin during ACLS training significantly improves their ability to establish effective pacing.

ECG artifact cancellation in surface EMG signals by fractional order calculus application

Background and objectiveNew aspects for automatic electrocardiography artifact removal from surface electromyography signals by application of fractional order calculus in combination with linear and nonlinear moving window filters are explored. Surface electromyography recordings of skeletal trunk muscles are commonly contaminated with spike shaped artifacts. This artifact originates from electrical heart activity, recorded by electrocardiography, commonly present in the surface electromyography signals recorded in heart proximity. For appropriate assessment of neuromuscular changes by means of surface electromyography, application of a proper filtering technique of electrocardiography artifact is crucial. MethodsA novel method for automatic artifact cancellation in surface electromyography signals by applying fractional order calculus and nonlinear median filter is introduced. The proposed method is compared with the linear moving average filter, with and without prior application of fractional order calculus. 3D graphs for assessment of window lengths of the filters, crest factors, root mean square differences, and fractional calculus orders (called WFC and WRC graphs) have been introduced. For an appropriate quantitative filtering evaluation, the synthetic electrocardiography signal and analogous semi-synthetic dataset have been generated. The examples of noise removal in 10 able-bodied subjects and in one patient with muscle dystrophy are presented for qualitative analysis. ResultsThe crest factors, correlation coefficients, and root mean square differences of the recorded and semi-synthetic electromyography datasets showed that the most successful method was the median filter in combination with fractional order calculus of the order 0.9. Statistically more significant (p < 0.001) ECG peak reduction was obtained by the median filter application compared to the moving average filter in the cases of low level amplitude of muscle contraction compared to ECG spikes. ConclusionsThe presented results suggest that the novel method combining a median filter and fractional order calculus can be used for automatic filtering of electrocardiography artifacts in the surface electromyography signal envelopes recorded in trunk muscles.

Causes of Prehospital Misinterpretations of ST Elevation Myocardial Infarction

Objectives: To determine the causes of software misinterpretation of ST elevation myocardial infarction (STEMI) compared to clinically identified STEMI to identify opportunities to improve prehospital STEMI identification. Methods: We compared ECGs acquired from July 2011 through June 2012 using the LIFEPAK 15 on adult patients transported by the Los Angeles Fire Department. Cases included patients ≥18 years who received a prehospital ECG. Software interpretation of the ECG (STEMI or not) was compared with data in the regional EMS registry to classify the interpretation as true positive (TP), true negative (TN), false positive (FP), or false negative (FN). For cases where classification was not possible using registry data, 3 blinded cardiologists interpreted the ECG. Each discordance was subsequently reviewed to determine the likely cause of misclassification. The cardiologists independently reviewed a sample of these discordant ECGs and the causes of misclassification were updated in an iterative fashion. Results: Of 44,611 cases, 50% were male (median age 65; inter-quartile range 52–80). Cases were classified as 482 (1.1%) TP, 711 (1.6%) FP, 43371 (97.2%) TN, and 47 (0.11%) FN. Of the 711 classified as FP, 126 (18%) were considered appropriate for, though did not undergo, emergent coronary angiography, because the ECG showed definite (52 cases) or borderline (65 cases) ischemic ST elevation, a STEMI equivalent (5 cases) or ST-elevation due to vasospasm (4 cases). The sensitivity was 92.8% [95% CI 90.6, 94.7%] and the specificity 98.7% [95% CI 98.6, 98.8%]. The leading causes of FP were ECG artifact (20%), early repolarization (16%), probable pericarditis/myocarditis (13%), indeterminate (12%), left ventricular hypertrophy (8%), and right bundle branch block (5%). There were 18 additional reasons for FP interpretation (<4% each). The leading causes of FN were borderline ST-segment elevations less than the algorithm threshold (40%) and tall T waves reducing the ST/T ratio below threshold (15%). There were 11 additional reasons for FN interpretation occurring ≤3 times each. Conclusion: The leading causes of FP automated interpretation of STEMI were ECG artifact and non-ischemic causes of ST-segment elevation. FN were rare and were related to ST-segment elevation or ST/T ratio that did not meet the software algorithm threshold.

ECG artefact identification and removal in mHealth systems for continuous patient monitoring.

Continuous patient monitoring systems acquire enormous amounts of data that is either manually analysed by doctors or automatically processed using intelligent algorithms. Sections of data acquired over long period of time can be corrupted with artefacts due to patient movement, sensor placement and interference from other sources. Owing to the large volume of data these artefacts need to be automatically identified so that the analysis systems and doctors are aware of them while making medical diagnosis. Three important factors are explored that must be considered and quantified for the design and evaluation of automatic artefact identification algorithms: signal quality, interpretation quality and computational complexity. The first two are useful to determine the effectiveness of an algorithm, whereas the third is particularly vital in mHealth systems where computational resources are heavily constrained. A series of artefact identification and filtering algorithms are then presented focusing on the electrocardiography data. These algorithms are quantified using the three metrics to demonstrate how different algorithms can be evaluated and compared to select the best ones for a given wireless sensor network.

Assessing ECG signal quality indices to discriminate ECGs with artefacts from pathologically different arrhythmic ECGs

False and non-actionable alarms in critical care can be reduced by developing algorithms which assess the trueness of an arrhythmia alarm from a bedside monitor. Computational approaches that automatically identify artefacts in ECG signals are an important branch of physiological signal processing which tries to address this issue. Signal quality indices (SQIs) derived considering differences between artefacts which occur in ECG signals and normal QRS morphology have the potential to discriminate pathologically different arrhythmic ECG segments as artefacts. Using ECG signals from the PhysioNet/Computing in Cardiology Challenge 2015 training set, we studied previously reported ECG SQIs in the scientific literature to differentiate ECG segments with artefacts from arrhythmic ECG segments. We found that the ability of SQIs to discriminate between ECG artefacts and arrhythmic ECG varies based on arrhythmia type since the pathology of each arrhythmic ECG waveform is different. Therefore, to reduce the risk of SQIs classifying arrhythmic events as noise it is important to validate and test SQIs with databases that include arrhythmias. Arrhythmia specific SQIs may also minimize the risk of misclassifying arrhythmic events as noise.

Corrigendum to "Improved multi-stage neonatal seizure detection using a heuristic classifier and a data-driven post-processor" [Clin Neurophysiol 127 (2016) 3014-3024

The authors regret that some text regarding Fig. 3 inadvertently appeared on p. 3018 in the paragraph after equation (3). The text should have read: To illustrate the difference of MPC between an artifact and an actual seizure, Fig. 4 plots a falsely detected ECG artifact (a, MPC: 0:21) and a truly detected rhythmic seizure (b, MPC: 0:04). In the current work, the MPC of EEG segment and the polygraphic signals ECG, EMG, EOG, and Resp. are measured and used as four synchronization features (Ansari et al., 2015). The authors would like to apologise for any inconvenience caused.

P123. Correlation between EEG and clinical symptoms in depression

The underlying neurophysiological characteristics in Major Depressive Disorders(MDD) have been examined intensively for the past decades. Some studies have indicated pattern of electroencephalography(EEG) activity that distinguishes MDD patients from healthy subjects. However, the correlation between either clinical symptoms such as psychometric scales or sleeping disturbances and EEG activity were seldom investigated to date. Therefore this study aims to examine the correlation between depressive and psychomotor symptoms as well as sleeping disturbances with EEG power values and with current source density (CSD) analysis. 20 patients with MDD (mean age ± SD: 47 ± 12.57; female: 13) were compared with 20 age matched healthy subjects (mean age ± SD: 51 ± 10.5; female: 12). A 10 minutes eyes-closed resting state EEG (31 channels, additional ECG channel, 10/20 system) was applied to all participants. Impedances were kept below 5kΩ. EEG data were pre-processed using Brain Vision Analyzer (Brain Products, Gilching). EEG raw data were manually cleaned from movement artefacts. ECG and ocular artefacts were excluded via independent component analysis. Artefact free EEG was filtered with 0.5Hz high pass and 50Hz low pass and average was set as reference. EEG power and asymmetry indices were calculated from 8 minutes of artefact free data for each frequency and pooled over 6 electrode sites. One minute of artefact free data in each participant were exported for CSD analysis. Voxelwise CSD power was calculated for each frequency with Low Resolution Electromagnetic Tomography (LORETA) and pooled for cortical brodmann areas with relation to executive, attentional, motor and emotional functions. EEG power, asymmetry and CSD power were the subjects of Pearson’s correlation analyses to examine the interconnection with depressive symptoms (BDI, HDRS), psychomotor activity (Motor Agitation and Retardation Scale), psychomotor speed (TMT A), executive functions (TMT B, Stroop), Pittsburgh Sleep Quality Intervention (PSQI) and measurement of motor activity via actigraphy. There were significant positive correlations between TMT A and B and EEG power in beta and gamma bands widespread from central to posterior regions (p). Our results suggested that EEG power in beta and gamma frequency bands over central to posterior regions as well as higher power in delta and beta CSD in frontal regions can possibly predict the decreased performances in neuropsychological tests related to psychomotor speed and divided attention in MDD group. Sleeping disturbances are associated with higher power in beta and gamma EEG frequencies over temporoparietal regions. These findings confirm that depressive symptoms in patients with MDD have neurofunctional expressions in the EEG domain and may suggest further that cognitive and sleep disturbances are associated with activity in distinct neurofunctional areas.

P124. QEEG and CSD power analysis in depression

Introduction Major depressive disorder (MDD) is a chronic mood disorder with significant impairment of daily living. Neurophysiological studies have focused to identify endophenotypes of depression. Frontal asymmetry in EEG power within alpha band has been the most studied topic and a few results indicated the phenomenon of lower left and higher right frontal activity. However, some other studies failed to show homogeneous results. Our aims in this study are to extensively investigate the neurophysiogical pattern in depressive patients by using qEEG power value, transformed current source density (CSD) as well as the neural synchronization. Method 20 patients with MDD (mean age ± SD: 47 ± 12.57 years old, female: 13) were compared with 20 age matched healthy subjects (mean age ± SD: 51 ± 10.5; female: 12). All participants underwent a 10 min resting state EEG recording with closed eyes. 31 electrodes elastic cap (10/20 system) with additional ECG channel were applied to all participants. Impedances were kept below 5 k Ω . EEG raw data were manually cleaned from movement artefacts. ECG and ocular artefacts were excluded via independent component analysis. Filter settings were 0.5 Hz for high pass and 30 Hz for low pass and all channels average was set as reference. EEG power and asymmetry indices were calculated from 8 min of artefact free data for each frequency and pooled over 6 electrode sites. One minute of artefact free data in each participant were exported for CSD analysis. Voxelwise CSD power was calculated for each frequency with Low Resolution Electromagnetic Tomography (LORETA). Lagged phase synchronisation between mean CSD power time series was calculated for frontal and cingular regions. ANOVA with the inner subject factors “pooled electrode locations” and “hemisphere” (6 × 2) and between subjects grouping factor (patients versus controls) were calculated to analyze EEG power and asymmetry for each frequency band. Repetitive measure post hoc t-test was further applied to investigate variables that caused the differences in ANOVA interactions. Group differences of CSD power and synchronisation were analyzed by t-tests. Result We found a main effect of group in the alpha2 (p = .017), delta (p = .043), theta (p = .043) and beta band (p = .007), the MDD patients have higher overall power in these frequencies. For the beta frequency, the increased EEG power of the MDD patients was also significant when analysed separately in post hoc t-test for frontal, central and centroparietal regions (p.Post hoc t-test revealed that in particular the group differences of EEG power in the temporoparietal region (p = .05) was driving this interaction effect. CSD analysis showed no significant difference between groups. Conclusion Depressive patients have shown a non-regional specific higher global EEG power in slow and high EEG frequencies compared to healthy controls. Especially increased overall power in the beta band in frontal, central and centroparietal regions can be seen in relation to difficulty falling asleep, and inner restlessness which is often found in MDD patients. Our result did not show frontal alpha asymmetry in depressive patients as well as in healthy control.

Evolutionary computing based approach for the removal of ECG artifact from the corrupted EEG signal.

Electroencephalogram (EEG) is an important tool for clinical diagnosis of brain-related disorders and problems. However, it is corrupted by various biological artifacts, of which ECG is one among them that reduces the clinical importance of EEG especially for epileptic patients and patients with short neck. To remove the ECG artifact from the measured EEG signal using an evolutionary computing approach based on the concept of Hybrid Adaptive Neuro-Fuzzy Inference System, which helps the Neurologists in the diagnosis and follow-up of encephalopathy. The proposed hybrid learning methods are ANFIS-MA and ANFIS-GA, which uses Memetic Algorithm (MA) and Genetic algorithm (GA) for tuning the antecedent and consequent part of the ANFIS structure individually. The performances of the proposed methods are compared with that of ANFIS and adaptive Recursive Least Squares (RLS) filtering algorithm. The proposed methods are experimentally validated by applying it to the simulated data sets, subjected to non-linearity condition and real polysomonograph data sets. Performance metrics such as sensitivity, specificity and accuracy of the proposed method ANFIS-MA, in terms of correction rate are found to be 93.8%, 100% and 99% respectively, which is better than current state-of-the-art approaches. The evaluation process used and demonstrated effectiveness of the proposed method proves that ANFIS-MA is more effective in suppressing ECG artifacts from the corrupted EEG signals than ANFIS-GA, ANFIS and RLS algorithm.

Ambulatory electrocardiographic monitoring between artifacts and misinterpretation, management errors of commission and errors of omission.

The aim of the study is to contrast the role of conventional ambulatory electrocardiographic monitoring (AEM) artifacts with a less emphasized problem with potentially more serious implications, that is, the failure to recognize, and therefore misinterpret, a genuine arrhythmia episode in the AEM recording. The study material included 500 Holter recordings and 500 recordings from the cardiac telemetry unit. Electrocardiographic (ECG) artifacts were more common in telemetry recordings (5.6%) compared to Holter recordings (4%) for a total of 4.8%. There were 35 examples of misinterpretation of AEM recordings (3.5%). These were significantly more common in telemetry recordings (2.6%) compared to Holter recordings (0.9%). The most common ECG artifacts were examples of pseudo ventricular tachyarrhythmia (VT). The majority of misinterpretation (26 of 35 examples) were fast supraventricular tachyarrhythmias with aberrant QRS (including six examples of atrial flutter with periods of 1:1 atrioventricular conduction) that were misdiagnosed as ventricular VT. Other examples were misinterpretation of arrhythmic episodes consistent with sick sinus syndrome, pacemaker malfunction, and long QT syndrome. Only 5 of 48 examples of AEM artifacts resulted in management errors of commission or errors of omission compared to all 35 examples of misinterpretation. Compared to conventional artifacts in AEM, misinterpretation of nonartifactual arrhythmic episodes consistently resulted in management errors. Misinterpretation was significantly more common with telemetry recordings compared to Holter ECG. This highlights the need for more appropriate training of the entire clinical team in charge of the management of the cardiac telemetry unit.

Two-Dimensional Warping for One-Dimensional Signals—Conceptual Framework and Application to ECG Processing

We propose a novel method for evaluating the similarity between two 1d patterns. Our method, referred to as two-dimensional signal warping (2DSW), extends the basic ideas of known warping techniques such as dynamic time warping and correlation optimized warping. By employing two-dimensional piecewise stretching 2DSW is able to take into account inhomogeneous variations of shapes. We apply 2DSW to ECG recordings to extract beat-to-beat variability in QT intervals (QTV) that is indicative of ventricular repolarization lability and typically characterised by a low signal-to-noise ratio. Simulation studies show high robustness of our approach in presence of typical ECG artefacts. Comparison of short-term ECG recorded in normal subjects versus patients with myocardial infarction (MI) shows significantly increased QTV in patients (normal subject 2.36 ms ± 1.05 ms vs. MI patients 5.94 ms ± 5.23 ms (mean ± std), ). Evaluation of a standard QT database shows that 2DSW allows highly accurate tracking of QRS-onset and T-end. In conclusion, the two-dimensional warping approach introduced here is able to detect subtle changes in noisy quasi-periodic biomedical signals such as ECG and may have diagnostic potential for measuring repolarization lability in MI patients. In more general terms, the proposed method provides a novel means for morphological characterization of 1d signals.