Improvements in fetal heart rate analysis by the removal of maternal-fetal heart rate ambiguities

Extracting fetal heart signals from Doppler using semi-supervised convolutional neural networks

Cardiotocography (CTG) measurements are critical for assessing fetal wellbeing during monitoring, and accurate assessment requires well-traceable CTG signals. The current FHR calculation algorithm, based on autocorrelation to Doppler ultrasound (DUS) signals, often results in periods of loss owing to its inability to differentiate signals. We hypothesized that classifying DUS signals by type could be a solution and proposed that an artificial intelligence (AI)-based approach could be used for classification. However, limited studies have incorporated the use of AI for DUS signals because of the limited data availability. Therefore, this study focused on evaluating the effectiveness of semi-supervised learning in enhancing classification accuracy, even in limited datasets, for DUS signals. Data comprising fetal heartbeat, artifacts, and two other categories were created from non-stress tests and labor DUS signals. With labeled and unlabeled data totaling 9,600 and 48,000 data points, respectively, the semi-supervised learning model consistently outperformed the supervised learning model, achieving an average classification accuracy of 80.9%. The preliminary findings indicate that applying semi-supervised learning to the development of AI models using DUS signals can achieve high generalization accuracy and reduce the effort. This approach may enhance the quality of fetal monitoring.

Fetal Heart Rate Preprocessing Techniques: A Scoping Review

Monitoring fetal heart rate (FHR) through cardiotocography is crucial for the early diagnosis of fetal distress situations, necessitating prompt obstetrical intervention. However, FHR signals are often marred by various contaminants, making preprocessing techniques essential for accurate analysis. This scoping review, following PRISMA-ScR guidelines, describes the preprocessing methods in original research articles on human FHR (or beat-to-beat intervals) signal preprocessing from PubMed and Web of Science, published from their inception up to May 2021. From the 322 unique articles identified, 54 were included, from which prevalent preprocessing approaches were identified, primarily focusing on the detection and correction of poor signal quality events. Detection usually entailed analyzing deviations from neighboring samples, whereas correction often relied on interpolation techniques. It was also noted that there is a lack of consensus regarding the definition of missing samples, outliers, and artifacts. Trends indicate a surge in research interest in the decade 2011-2021. This review underscores the need for standardizing FHR signal preprocessing techniques to enhance diagnostic accuracy. Future work should focus on applying and evaluating these methods across FHR databases aiming to assess their effectiveness and propose improvements.

Intrapartum cardiotocography with simultaneous maternal heart rate registration improves neonatal outcome

BACKGROUND

Intrapartum cardiotocographic monitoring of fetal heart rate by abdominal external ultrasound transducer without simultaneous maternal heart rate recording has been associated with increased risk of early neonatal death and other asphyxia-related neonatal outcomes. It is unclear, however, whether this increase in risk is independently associated with fetal surveillance method or is attributable to other factors.

OBJECTIVE

This study aimed to compare different fetal surveillance methods and their association with adverse short- and long-term fetal and neonatal outcomes in a large retrospective cohort of spontaneous term deliveries.

STUDY DESIGN

Fetal heart rate and maternal heart rate patterns were recorded by cardiotocography during labor in spontaneous term singleton cephalic vaginal deliveries in the Hospital District of Helsinki and Uusimaa, Finland between October 1, 2005, and September 30, 2023. According to the method of cardiotocography monitoring at birth, the cohort was divided into the following 3 groups: women with ultrasound transducer, women with both ultrasound transducer and maternal heart rate transducer, and women with internal fetal scalp electrode. Umbilical artery pH and base excess values, low 1- and 5-minute Apgar scores, need for intubation and resuscitation, neonatal intensive care unit admission for asphyxia, neonatal encephalopathy, and early neonatal death were used as outcome variables.

RESULTS

Among the 213,798 deliveries that met the inclusion criteria, the monitoring type was external ultrasound transducer in 81,559 (38.1%), both external ultrasound transducer and maternal heart rate recording in 62,268 (29.1%), and fetal scalp electrode in 69,971 (32.7%) cases, respectively. The rates of both neonatal encephalopathy (odds ratio, 1.48; 95% confidence interval, 1.08-2.02) and severe acidemia (umbilical artery pH <7.00 and/or umbilical artery base excess ≤-12.0 mmol/L) (odds ratio, 2.03; 95% confidence interval, 1.65-2.50) were higher in fetuses of women with ultrasound transducer alone compared with those of women with concurrent external fetal and maternal heart rate recording. Monitoring with ultrasound transducer alone was also associated with increased risk of neonatal intubation for resuscitation (odds ratio, 1.22; 95% confidence interval, 1.03-1.44). A greater risk of severe neonatal acidemia was observed both in the ultrasound transducer (odds ratio, 2.78; 95% confidence interval, 2.23-3.48) and concurrent ultrasound transducer and maternal heart rate recording (odds ratio, 1.37; 95% confidence interval, 1.05-1.78) groups compared with those monitored with fetal scalp electrodes. No difference in risk of neonatal encephalopathy was found between newborns monitored with concurrent ultrasound transducer and maternal heart rate recording and those monitored with fetal scalp electrodes.

CONCLUSION

The use of external ultrasound transducer monitoring of fetal heart rate without simultaneous maternal heart rate recording is associated with higher rates of neonatal encephalopathy and severe neonatal acidemia. We suggest that either external fetal heart rate monitoring with concurrent maternal heart rate recording or internal fetal scalp electrode be used routinely as a fetal surveillance tool in term deliveries.

Multichannel high noise level ECG denoising based on adversarial deep learning

This paper proposes a denoising method based on an adversarial deep learning approach for the post-processing of multi-channel fetal electrocardiogram (ECG) signals. As it's well known, noise leads to misinterpretations of fetal ECG signals and thus limits the use of fetal electrocardiography for healthcare applications. Therefore, denoising algorithms are essential for the exploitation of non-invasive fetal ECG. The proposed method is based on the combination of three end-to-end trained sub-networks to convert noisy fetal ECG signals into clean signals. The first two sub-networks are linked by skip connections and form a deep convolutional network that downsamples the noisy signals into a latent representation and subsequently upsamples this latent representation to recover clean signals. The third sub-network aims to boost the decoder sub-network to generate realistic clean signals. Experiments carried out on synthetic and real data showed that the proposed method improved by the signal-to-noise (SNR) of fetal ECG signals with input SNR ranging from [Formula: see text] to 0 dB by an average of 20 dB, and improve fetal signal quality by significantly increasing the number of true detected QRS complexes and halving QRS complex detection errors.

Sporadic accelerations during labor strongly indicate normal pH, whereas periodic accelerations do not: a case-control study

PURPOSE

To determine the association between the occurrence of sporadic and periodic fetal heart rate accelerations during labor and acidemia at birth.

MATERIALS AND METHODS

This is a case-control study of fetal heart rate patterns from 364 neonates with acidemia at birth (cord blood pH <7.05 at vaginal birth, or pH <7.10 at birth after first stage cesarean delivery) and 731 controls with pH ≥7.15. The last 30-60 min of the cardiotocographic traces before birth from the neonates born with acidemia and from the corresponding stage in labor for the controls were scrutinized. Odds ratios (OR) with 95% confidence interval for acidemia at birth were determined.

RESULTS

During the first stage, ≥2 sporadic accelerations were present in 16% of cases and 78% of controls; OR for acidemia (compared to 0-1 accelerations) 0.05 (0.02-0.10). In the second stage, the corresponding rates were 13% and 60%, OR 0.09 (0.06-0.14). Isolated periodic accelerations were infrequent. A weak negative association between ≥2 periodic accelerations and acidemia (compared with 0-1 accelerations) was found in the second stage, OR 0.51 (0.30-0.86), but was not significant in the first stage, OR 0.24 (0.04-1.4). Even among fetuses with normal fetal heart rate variability (5-25 beats per minute) the occurrence of less than two sporadic accelerations was associated with an increased risk of acidemia, OR 10.3 (7.2-14.8).

CONCLUSIONS

Sporadic accelerations indicate a very low probability of acidosis but are absent in 40% of fetuses with normal pH during a 30-60 min second-stage recording.

Computerized analysis of cardiotocograms in clinical practice and the SisPorto® system thirty-two years after: technological, physiopathological and clinical studies

OBJECTIVES

The objective of this study is to present the why, what and how about computerized analysis of cardiotocograms (cCTG) and the SisPorto system for cCTG.

CONTENT

A narrative review about cCTG and the SisPorto system for cCTG is presented. The meta-analysis of randomized controlled trials (RCT) performed so far have evidenced that cCGT compared to traditional CTG analysis may save time spent in hospital for women, in the antepartum period, and is objective with at least equivalent results in maternal and perinatal outcomes, both in the ante and intrapartum periods. The SisPorto system for cCTG closely follows the FIGO guidelines for fetal monitoring. It may be used both in the ante and intrapartum periods, alone or connected to a central monitoring station, with simultaneous monitoring of fetal and maternal signals, not only in singletons but also in twins. It has been assessed in technical, physiopathological and clinical studies, namely in one large multicentric international RCT during labor and two meta-analysis.

SUMMARY AND OUTLOOK

There is evidence that cCTG may be useful in clinical practice with advantages compared to traditional CTG analysis, although without clear impact on the decrease of preventable maternal and perinatal mortality and morbidity. More studies are warranted, namely on technical improvements and assessment in larger studies in a wider range of clinical settings.

Technical Update No. 429: Maternal Heart Rate Artefact During Intrapartum Fetal Health Surveillance

OBJECTIVE

To improve perinatal outcomes and minimize provider error by increasing awareness of strategies to detect intrapartum maternal heart rate artefact and to respond when such artefact is suspected.

TARGET POPULATION

All pregnant patients during labour.

OPTIONS

Maternal heart rate artefact may be detected based on clinical features or through technology. Suspected maternal heart rate artefact may be assessed by applying a fetal scalp electrode (preferred) or through external fetal monitoring, augmented by point-of-care sonography (alternative).

OUTCOMES

Unrecognized intrapartum maternal heart rate artefact increases the risk that abnormal/atypical fetal heart rate patterns will go undetected and, hence, the risk of adverse perinatal outcomes.

BENEFITS, HARMS, AND COSTS

Unrecognized maternal heart rate artefact can lead to adverse perinatal outcomes (hypoxic-ischemic encephalopathy, fetal death, and neonatal death) and adverse maternal outcomes (unnecessary cesarean delivery or operative vaginal delivery). Timely recognition of such artefact may avoid these adverse outcomes. The costs of early recognition of maternal heart rate artefact are relatively small: increased use of fetal scalp electrodes and point-of-care sonography, as well as additional assessments by the health care provider. The cost savings are significant, as a result of lower risk of adverse perinatal outcomes. Potential harms are false-positive diagnoses of maternal heart rate artefact, expediting delivery unnecessarily when the fetal status cannot be reliably determined but is normal, and the rare complications associated with increased use of fetal scalp electrodes.

EVIDENCE

Two PubMed searches were completed. The first was for articles published between January 1, 1970, and November 25, 2021, using the medical subject headings (MeSH) "fetal monitoring" and "artifacts" (38 articles). The second was for articles published during the same period using the MeSH "fetal monitoring" and "maternal heart rate" (841 articles).

VALIDATION METHODS

The authors rated the quality of evidence and strength of recommendations using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach. See online Appendix A (Tables A1 for definitions and A2 for interpretations of strong and conditional [weak] recommendations).

INTENDED AUDIENCE

All health care providers involved in obstetrical care.

SUMMARY STATEMENTS

RECOMMENDATIONS.

Mise à jour technique no 429 : Artéfact de la fréquence cardiaque maternelle pendant la surveillance du bien-être fœtal per partum

OBJECTIF

Améliorer les issues périnatales et réduire au minimum le risque d'erreurs chez les fournisseurs en améliorant leurs connaissances sur les stratégies de détection des artéfacts de la fréquence cardiaque maternelle per partum et sur les modes d'intervention lorsque de tels artéfacts sont soupçonnés.

POPULATION CIBLE

Toutes les parturientes.

OPTIONS

L'artéfact de la fréquence cardiaque maternelle peut être détecté à l'aide de caractéristiques cliniques ou de la technologie. On peut évaluer l'artéfact de la fréquence cardiaque maternelle soupçonné en posant une électrode de cuir chevelu fœtal (option à privilégier) ou en recourant à la surveillance fœtale externe optimisée par l'échographie au chevet (solution de rechange). RéSULTATS: Les artéfacts de la fréquence cardiaque maternelle per partum non détectés augmentent le risque que des rythmes anormaux ou atypiques de la fréquence cardiaque fœtale passent inaperçus, ce qui augmente le risque d'issues périnatales défavorables. BéNéFICES, RISQUES ET COûTS: L'artéfact de la fréquence cardiaque maternelle non détecté peut entraîner de graves issues périnatales défavorables (encéphalopathie hypoxo-ischémique, mort fœtale et mort néonatale) et des issues maternelles défavorables (césarienne injustifiée ou accouchement assisté). Ces issues peuvent être évitées par la détection rapide d'un tel artéfact. Le coût de la détection précoce des artéfacts de fréquence cardiaque maternelle est relativement faible (utilisation accrue des électrodes de cuir chevelu fœtal et de l'échographie au chevet avec évaluations supplémentaires par le fournisseur de soins). La réduction des événements périnataux défavorables engendre des économies considérables. Les risques sont : faux positifs d'artéfact de la fréquence cardiaque maternelle; accélération inutile de l'accouchement lorsque l'état du fœtus est normal, mais qu'on ne peut le déterminer de façon fiable; et les rares complications associées à l'utilisation accrue des électrodes de cuir chevelu fœtal. DONNéES PROBANTES: Deux recherches ont été effectuées dans PubMed. La première a été réalisée pour répertorier les articles publiés entre le 1^er janvier 1970 et le 25 novembre 2021 à partir des termes MeSH fetal monitoring et artifacts (38 articles); la deuxième, pour répertorier les articles publiés au cours de la même période à partir des termes MeSH fetal monitoring et maternal heart rate (841 articles). MéTHODES DE VALIDATION: Les auteurs ont évalué la qualité des données probantes et la force des recommandations en utilisant le cadre méthodologique GRADE (Grading of Recommendations, Assessment, Development, and Evaluation). Voir l'annexe A en ligne (tableau A1 pour les définitions et tableau A2 pour l'interprétation des recommandations fortes et conditionnelles [faibles]). PROFESSIONNELS CONCERNéS: Tous les fournisseurs de soins obstétricaux. DÉCLARATIONS SOMMAIRES: RECOMMANDATIONS.

Use of Deep Learning to Detect the Maternal Heart Rate and False Signals on Fetal Heart Rate Recordings

We have developed deep learning models for automatic identification of the maternal heart rate (MHR) and, more generally, false signals (FSs) on fetal heart rate (FHR) recordings. The models can be used to preprocess FHR data prior to automated analysis or as a clinical alert system to assist the practitioner. Three models were developed and used to detect (i) FSs on the MHR channel (the FSMHR model), (ii) the MHR and FSs on the Doppler FHR sensor (the FSDop model), and (iii) FSs on the scalp ECG channel (the FSScalp model). The FSDop model was the most useful because FSs are far more frequent on the Doppler FHR channel. All three models were based on a multilayer, symmetric, GRU, and were trained on data recorded during the first and second stages of delivery. The FSMHR and FSDop models were also trained on antepartum recordings. The training dataset contained 1030 expert-annotated periods (mean duration: 36 min) from 635 recordings. In an initial evaluation of routine clinical practice, 30 fully annotated recordings for each sensor type (mean duration: 5 h for MHR and Doppler sensors, and 3 h for the scalp ECG sensor) were analyzed. The sensitivity, positive predictive value (PPV) and accuracy were respectively 62.20%, 87.1% and 99.90% for the FSMHR model, 93.1%, 95.6% and 99.68% for the FSDop model, and 44.6%, 87.2% and 99.93% for the FSScalp model. We built a second test dataset with a more solid ground truth by selecting 45 periods (lasting 20 min, on average) on which the Doppler FHR and scalp ECG signals were recorded simultaneously. Using scalp ECG data, the experts estimated the true FHR value more reliably and thus annotated the Doppler FHR channel more precisely. The models achieved a sensitivity of 53.3%, a PPV of 62.4%, and an accuracy of 97.29%. In comparison, two experts (blinded to the scalp ECG data) respectively achieved a sensitivity of 15.7%, a PPV of 74.3%, and an accuracy of 96.91% and a sensitivity of 60.7%, a PPV of 83.5% and an accuracy of 98.24%. Hence, the models performed at expert level (better than one expert and worse than the other), although a well-trained expert with good knowledge of FSs could probably do better in some cases. The models and datasets have been included in the Fetal Heart Rate Morphological Analysis open-source MATLAB toolbox and can be used freely for research purposes.

Appraisal of trimester-specific fetal heart rate and its role in gestational age prediction

INTRODUCTION

The aim of this study was to evaluate and report normal sonographic FHR values among low-risk singleton women across the three trimesters of pregnancy and determine FHR role in gestational age prediction.

METHOD

A prospective cross-sectional study of 2727 low-risk singleton pregnant women was undertaken. FHR measurements were obtained by a consultant radiologist and three experienced sonographers using transabdominal approach from January 2019 to December 2020. Two FHR measurements were taken for each participant. The fetal lie and presentation were also documented in the first trimester. Data were analysed using SPSS version 24 (IBM, Armonk, NY, USA).

RESULT

The maternal mean ± SD age was 25.8 ± 6.5 years and mean FHR for first, second and third trimesters were 151 ± 16, 145 ± 6 and 125±6 bpm respectively. The mean ± SD gestational age were 10 ± 2, 19 ± 3 and 34 ± 2 weeks for the first, second and third trimester respectively. Using ANOVA, there were statistically significant differences in FHR across the three trimesters (p ≤ 0.05). A positive correlation existed between maternal age and FHR (r = 0.57, p ≤ 0.05).

CONCLUSION

This study has established normal values for FHR in first, second and third trimester respectively. Referring physicians, radiologists, sonographers, obstetricians and gynaecologists may consider FHR of (135-167) bpm (139-151) bpm and (119-131) bpm as normal FHR ranges for the first, second and third trimester respectively. This study has also revealed the possibility of gestational age prediction using FHR with the equation [Gestational Age = 87.8 - (0.47) FHR].

IMPLICATIONS FOR PRACTICE

This paper provides the most up-to-date sonographic FHR recommendations for fetal management. More importantly, findings from this study also suggests that ultrasound practitioners can use FHR measurements as a reliable alternative for fetal dating.

Characteristics of fetal and maternal heart rate tracings during labor: A prospective observational study

BACKGROUND

Fetal well-being is assured during labor and delivery with the employment of electronic fetal heart monitoring (EFHM). In uncommon instances, maternal heart rate (MHR) instead of fetal heart rate (FHR) can be the source of signals on monitors (signal ambiguity) leading to erroneous interpretation and management. Information about MHR characteristics are comparatively inadequate. We aim to analyze and compare MHR and FHR characteristics during the first and second stages of labor.

METHODS

A prospective cohort study was conducted in a single tertiary care center during a one year period. Fifty one healthy full term women with singleton pregnancies during labor were enrolled. Uterine contractions, MHR and FHR were recorded simultaneously during both stages of labor by monitors designed for twin gestation.

RESULTS

When compared to FHR, MHR had significantly lower baseline rate during 1st and 2nd stages (p < 0.0001). It demonstrated also more marked beat-to-beat variability during both stages (p < 0.0001). MHR showed significantly more accelerations (p = 0.03 and p = 0.008) and less decelerations (p < 0.0001 and p = 0.021) during 1st and 2nd stages respectively.

CONCLUSIONS

All characteristic parameters and patterns produced by FHR could be mimicked by MHR as well, though, at different frequencies. Understanding EFHM patterns suspected to be MHR artefacts and the employment of modern monitors that simultaneously obtain and display FHR and MHR can unmask ambiguity and avert related misinterpretation problems. Similar studies should be conducted in high-risk groups where the potential for fetal hypoxia/acidosis is increased.

Unrecognized maternal heart rate artefact in cases of perinatal mortality reported to the United States Food and Drug Administration from 2009 to 2019: a critical patient safety issue

BACKGROUND

Maternal heart rate artefact is a signal processing error whereby the fetal heart rate is masked by the maternal pulse, potentially leading to danger by failure to recognize an abnormal fetal heart rate or a pre-existing fetal death. Maternal heart rate artefact may be exacerbated by autocorrelation algorithms in modern fetal monitors due to smooth transitions between maternal and fetal heart rates rather than breaks in the tracing. In response, manufacturers of cardiotocography monitors recommend verifying fetal life prior to monitoring and have developed safeguards including signal ambiguity detection technologies to simultaneously and continuously monitor the maternal and fetal heart rates. However, these safeguards are not emphasized in current cardiotocography clinical practice guidelines, potentially leading to a patient safety gap.

METHODS

The United States Food and Drug Administration Manufacturer and User Facility Device Experience database was reviewed for records with event type "Death" for the time period March 31, 2009 to March 31, 2019, in combination with search terms selected to capture all cases reported involving cardiotocography devices. Records were reviewed to determine whether maternal heart rate artefact was probable and/or whether the report contained a recommendation from the device manufacturer regarding maternal heart rate artefact.

RESULTS

Forty-seven cases of perinatal mortality were identified with probable maternal heart rate artefact including 14 with antepartum fetal death prior to initiation of cardiotocography, 14 with intrapartum fetal death or neonatal death after initiation of cardiotocography, and 19 where the temporal relationship between initiation of cardiotocography and death cannot be definitively established from the report. In 29 cases, there was a recommendation from the manufacturer regarding diagnosis and/or management of maternal heart rate artefact.

CONCLUSIONS

This case series indicates a recurring problem with undetected maternal heart rate artefact leading to perinatal mortality and, in cases of pre-existing fetal death, healthcare provider confusion. In response, manufacturers frequently recommend safeguards which are found in their device's instructions for use but not in major intrapartum cardiotocography guidelines. Cardiotocography guidelines should be updated to include the latest safeguards against the risks of maternal heart rate artefact. An additional file summarizing key points for clinicians is included.

Use of automated fetal heart rate analysis to identify risk factors for umbilical cord acidosis at birth

OBJECTIVE

To identify clinical parameters and intrapartum fetal heart rate parameters associated with a risk of umbilical cord acidosis at birth, using an automated analysis method based on empirical mode decomposition.

METHODS

Our single-center study included 381 cases (arterial cord blood pH at birth pHa ≤7.15) and 1860 controls (pHa ≥7.25) extracted from a database comprising 8,383 full datasets for over-18 mothers after vaginal or caesarean non-twin, non-breech deliveries at term (>37 weeks of amenorrhea). The analysis of a 120-min period of the FHR recording (before maternal pushing or the decision to perform a caesarean section during labor) led to the extraction of morphological, frequency-related, and long- and short-term heart rate variability variables. After univariate analyses, sparse partial least square selection and logistic regression were applied.

RESULTS

Several clinical factors were predictive of fetal acidosis in a multivariate analysis: nulliparity (odds ratio (OR) 95% confidence interval (CI)]: 1.769 [1.362-2.300]), a male fetus (1.408 [1.097-1.811]), and the term of the pregnancy (1.333 [1.189-1.497]). The risk of acidosis increased with the time interval between the end of the FHR recording and the delivery (OR [95%CI] for a 1-min increment: 1.022 [1.012-1.031]). The risk factors related to the FHR signal were mainly the difference between the mean baseline and the mean FHR (OR [95%CI]: 1.292 [1.174-1.424]), the baseline range (1.027 [1.014-1.040]), fetal bradycardia (1.038 [1.003-1.075]) and the late deceleration area (1.002 [1.000-1.005]). The area under the curve for the multivariate model was 0.79 [0.76; 0.81].

CONCLUSION

In addition to clinical predictors, the automated FHR analysis highlighted other significant predictors, such as the baseline range, the instability of the FHR signal and the late deceleration area. This study further extends the routine application of automated FHR analysis during labor and, ultimately, contributes to the development of predictive scores for fetal acidosis.

Fetal heart rate baseline computation with a weighted median filter

BACKGROUND

Automated fetal heart rate (FHR) analysis removes inter- and intra-expert variability, and is a promising solution for reducing the occurrence of fetal acidosis and the implementation of unnecessary medical procedures. The first steps in automated FHR analysis are determination of the baseline, and detection of accelerations and decelerations (A/D). We describe a new method in which a weighted median filter baseline (WMFB) is computed and A/Ds are then detected.

METHOD

The filter weightings are based on the prior probability that the sampled FHR is in the baseline state or in an A/D state. This probability is computed by estimating the signal's stability at low frequencies and by progressively trimming the signal. Using a competition dataset of 90 previously annotated FHR recordings, we evaluated the WMFB method and 11 recently published literature methods against the ground truth of an expert consensus. The level of agreement between the WMFB method and the expert consensus was estimated by calculating several indices (primarily the morphological analysis discordance index, MADI). The agreement indices were then compared with the values for eleven other methods. We also compared the level of method-expert agreement with the level of interrater agreement.

RESULTS

For the WMFB method, the MADI indicated a disagreement of 4.02% vs. the consensus; this value is significantly lower (p<10^-13) than that calculated for the best of the 11 literature methods (7.27%, for Lu and Wei's empirical mode decomposition method). The level of inter-expert agreement (according to the MADI) and the level of WMFB-expert agreement did not differ significantly (p=0.22).

CONCLUSION

The WMFB method reproduced the expert consensus analysis better than 11 other methods. No differences in performance between the WMFB method and individual experts were observed. The method Matlab source code is available under General Public Licence at http://utsb.univ-catholille.fr/fhr-wmfb.