Passive detection of accelerometer-recorded fetal movements using a time–frequency signal processing approach

doi:10.1016/j.dsp.2013.10.002

Digital Signal Processing

Volume 25, February 2014, Pages 134-155

https://doi.org/10.1016/j.dsp.2013.10.002 Get rights and content

Abstract

This paper describes a multi-sensor fetal movement (FetMov) detection system based on a time–frequency (TF) signal processing approach. Fetal motor activity is clinically useful as a core aspect of fetal screening for well-being to reduce the current high incidence of fetal deaths in the world. FetMov are present in early gestation but become more complex and sustained as the fetus progresses through gestation. A decrease in FetMov is an important element to consider for the detection of fetal compromise. Current methods of FetMov detection include maternal perception, which is known to be inaccurate, and ultrasound imaging which is intrusive and costly. An alternative passive method for the detection of FetMov uses solid-state accelerometers, which are safe and inexpensive. This paper describes a digital signal processing (DSP) based experimental approach to the detection of FetMov from recorded accelerometer signals. The paper provides an overview of the significant measurement and signal processing challenges, followed by an approach that uses quadratic time–frequency distributions (TFDs) to appropriately deal with the non-stationary nature of the signals. The paper then describes a proof-of-concept with a solution consisting of a detection method that includes (1) a new experimental set-up, (2) an improved data acquisition procedure, and (3) a TF approach for the detection of FetMov including TF matching pursuit (TFMP) decomposition and TF matched filter (TFMF) based on high-resolution quadratic TFDs. Detailed suggestions for further refinement are provided with preliminary results to establish feasibility, and considerations for application to clinical practice are reviewed.

Section snippets

Identifying the at-risk fetus

The main interest in a DSP approach to the problem of automatically detecting a FetMov reduction stems from the fact that current methods of fetal surveillance and monitoring fail to identify compromised fetuses at risk of death in utero [1]. Such a DSP approach has been shown to be effective in the detection of abnormalities in several physiological signals including EEG signal, HRV and ECG. There are marked differences in stillbirth rates between low and middle-income countries (30 to 50 per

Participants, ethics and data

The research team invited women attending for antenatal care at the Royal Brisbane and Womenʼs Hospital, Brisbane, Australia, to participate in the study if they had a live, singleton pregnancy and the gestational age was 32 weeks or more. Fetuses with major nervous system or musculoskeletal disorders likely to affect motor activity were excluded. The study was approved by the Royal Brisbane and Womenʼs Hospital Human Research Ethics Committee and the University of Queensland Medical Research

Time–frequency formulation

The Fourier Transform (FT) is well suited for analyzing stationary signals in the frequency domain. For non-stationary signals such as EEG, HRV or FetMov signals, the FT cannot provide for accurate frequency localization together with accurate time localization and therefore a time–frequency analysis is the preferred approach. One effective way to represent and analyze a non-stationary signal in the TF domain is the class of quadratic TF distributions (QTFDs). This class is defined as smoothed

Performance measures and assumptions

Using the aforementioned methodology on accelerometer data from six subjects, an automated process for the detection of FetMov was defined for both TFMP decomposition and TF matched filter. The performance of these two detection processes is compared to ultrasoundʼs masks, as presented in Table 1; the comparison criterion are the true detection rate (TDR), the positive predictive value (PPV), expressed as: $TDR = 100 \times \frac{DME}{TME}$ $PPV = 100 \times \frac{DME}{(DME + FD)}$ where TME represents the total movement events (as

Conclusions

There is a clinical need for an accurate method to monitor fetal movements (FetMovs) over prolonged periods in late gestation in order to safely and non-invasively identify the onset of fetal compromise and allow intervention before fetal death occurs. The experimental study presented in this paper indicates that the use of accelerometers constitutes a promising advance in the automated detection of FetMov. A proof-of-concept is presented based on two time–frequency (TF) techniques, namely TF

Acknowledgements

The authors thank S. Callan, S. Dann, D. Travers and L. Brookes of the Royal Brisbane and Women Hospital, Brisbane, Australia, for collecting the accelerometer and ultrasound data and for marking the ultrasound videos. S. Layeghy is thanked for assisting in data acquisition. This research was made possible by a National Priorities Research Program (NPRP 096262243) grant from the Qatar National Research Fund (a member of The Qatar Foundation). Prior funding for related aspects of this research

Boualem Boashash (IEEE Fellow ‘99’) is a Scholar, Professor and Senior Academic Manager with experience in 5 leading Universities in France and Australia. He has published over 500 technical publications, including over 100 journal publications covering Engineering, Applied Mathematics and Bio-medicine. He pioneered the field of Time–Frequency Signal Processing for which he published the most comprehensive book and most powerful software package. Among many initiatives, he founded ISSPA, a

References (61)

V. Flenady et al.
Stillbirths: The way forward in high-income countries
Lancet
(May 2011)
R.L. Goldenberg et al.
Stillbirths: the vision for 2020
Lancet
(May 2011)
J.A. Low et al.
Limitations in the clinical prediction of intrapartum fetal asphyxia
Am. J. Obstet. Gynecol.
(March 1995)
C. Einspieler et al.
The qualitative assessment of general movements in preterm, term and young infants – Review of the methodology
Early Hum. Dev.
(1997)
J. ten Hof et al.
Quantitative analysis of fetal general movements: Methodological considerations
Early Hum. Dev.
(1999)
K. Marsal
Ultrasonic assessment of fetal activity
Clin. Obstet. Gynaecol.
(December 1983)
W.F. Rayburn
Fetal body movement monitoring
Obstet. Gynecol. Clin. North Am.
(March 1990)
B.S. Richardson et al.
Electrocortical activity, electroocular activity, and breathing movements in fetal sheep with prolonged and graded hypoxemia
Am. J. Obstet. Gynecol.
(August 1992)
D.J. Bekedam et al.
Motor behaviour in the growth retarded fetus
Early Hum. Dev.
(1985)
D.A. Sival et al.
The effect of intrauterine growth retardation on the quality of general movements in the human fetus
Early Hum. Dev.
(February 1992)

F. Kainer et al.

Assessment of the quality of general movements in fetuses and infants of women with type-I diabetes mellitus

Early Hum. Dev.

(1997)

H. Goovaerts et al.

A comparative clinical study of fetal phono- and movement-sensors from Amsterdam, Cambridge and Edinburgh

J. Biomed. Eng.

(1991)

K. Nishihara et al.

A long-term monitoring of fetal movement at home using a newly developed sensor: An introduction of maternal micro-arousals evoked by fetal movement during maternal sleep

Early Hum. Dev.

(2008)

I.E. Timor-Tritsch et al.

Classification of human foetal movement

Am. J. Obstet. Gynecol.

(September 1976)

Z. Hussain et al.

The T-class of time–frequency distributions: time-only kernels with amplitude estimation

J. Franklin Inst.

(2006)

E. Ryo et al.

A new method for long-term home monitoring of fetal movement by pregnant women themselves

Med. Eng. Phys.

(June 2012)

S.S. Abeysekera et al.

Methods of signal classification using the images produced by the Wigner–Ville distribution

Pattern Recogn. Lett.

(1991)

J.V.H. Tveit et al.

Concerns for decreased foetal movements in uncomplicated pregnancies – increased risk of foetal growth restriction and stillbirth among women being overweight advanced age or smoking

J. Matern.-Fetal Neonatal Med.

(October 2010)

Z. Li et al.

Australiaʼs mothers and babies 2009

T. Stampalija et al.

Utero-placental Doppler ultrasound for improving pregnancy outcome

Cochrane Database Syst. Rev.

(2010)

H.B. Westergaard et al.

A critical appraisal of the use of umbilical artery Doppler ultrasound in high-risk pregnancies: Use of meta-analyses in evidence-based obstetrics

Ultrasound Obstet. Gynecol.

(June 2001)

J.F. Froen et al.

Restricted fetal growth in sudden intrauterine unexplained death

Acta Obstet. Gynecol. Scand.

(September 2004)

A. Imdad et al.

Screening and triage of intrauterine growth restriction (IUGR) in general population and high risk pregnancies: a systematic review with a focus on reduction of IUGR related stillbirths

BMC Public Health

(2011)

J. Gardosi et al.

Maternal and fetal risk factors for stillbirth: population based study

BMJ

(2013)

C. Serena et al.

Stillbirth and fetal growth restriction

J. Matern.-Fetal Neonatal Med.

(2013)

B. Boashash

Time–Frequency Signal Analysis and Processing: A Comprehensive Reference

(2003)

Analog Devices, ADXL 330, 3-axis ± g iMEMS accelerometers, D05677-0-6/07(A)

A.D. Bocking

Assessment of fetal heart rate and fetal movements in detecting oxygen deprivation in-utero

Eur. J. Obstet. Gynecol. Reprod. Biol.

(2003)

M.D. Velazquez et al.

Antenatal evaluation of the fetus using fetal movement monitoring

Clin. Obstet. Gynaecol.

(December 2002)

D. Blyton et al.

Treatment of sleep disordered breathing reverses low fetal activity levels in preeclampsia

Sleep

(2013)

Cited by (36)

Multi-modal detection of fetal movements using a wearable monitor
2024, Information Fusion
The importance of Fetal Movement (FM) patterns as a biomarker for fetal health has been extensively argued in obstetrics. However, the inability of current FM monitoring methods, such as ultrasonography, to be used outside clinical environments has made it challenging to understand the nature and evolution of FM. A small body of work has introduced wearable sensor-based FM monitors to address this gap. Despite promises in controlled environments, reliable instrumentation to monitor FM out-of-clinic remains unresolved, particularly due to the challenges of separating FMs from interfering artifacts arising from maternal activities. To date, efforts have been focused almost exclusively on homogenous (single) sensing and information fusion modalities, such as decoupled acoustic or accelerometer sensors. However, FM and related signal artifacts have varying power and frequency bandwidths that homogeneous sensor arrays may not capture or separate efficiently. In this investigation, we introduce a novel wearable FM monitor with an embedded heterogeneous sensor suite combining accelerometers, acoustic sensors, and piezoelectric diaphragms designed to capture a broad range of FM and interfering artifact signal features enabling more efficient isolation of both. We further outline a novel data fusion architecture combining data-dependent thresholding and machine learning to automatically detect FM and separate it from signal artifacts in real-world (home) environments. The performance of the device and the data fusion architecture are validated using 33 h of at-home use through concurrent recording of maternal perception of FM. The FM monitor detected an impressive 82 % of maternally sensed FMs with an overall accuracy of 90 % in recognizing FM and non-FM events. Consistency of detection was strongest from 32 gestational weeks onwards, which overlaps with the critical FM monitoring window for stillbirth prevention. We believe the multi-modal sensor fusion approach presented in this research will be a major milestone in the development of low-cost wearable FM monitors enabling pervasive monitoring of FM in unsupervised environments.
Automatic fetal movement recognition from multi-channel accelerometry data
2021, Computer Methods and Programs in Biomedicine
Citation Excerpt :
For the MPD-based feature, however, a different approach was adopted for determining the decomposition stopping criterion. In [12], the stopping criterion for the iterative MPD was arbitrarily fixed. In this study, however, the MPD stopping criterion – still based on energy of the residual signal – was carefully selected as described below.
Significant health care resources are allocated to monitoring high risk pregnancies to minimize growth compromise, reduce morbidity and prevent stillbirth. Fetal movement has been recognized as an important indicator of fetal health. Studies have shown that 25% of pregnancies with decreased fetal movement in the third trimester led to poor outcomes at birth. The studies have also shown that maternal perception of fetal movement is highly subjective and varies from person to person. A non-invasive system for fetal movement detection that can be used outside hospital would represent an advance in at-home monitoring of at-risk pregnancies. This is a challenging task that requires the use of advanced signal processing techniques to differentiate genuine fetal movements from contaminating artefacts.
This manuscript proposes a novel algorithm for automatic fetal movement recognition using data collected from wearable tri-axial accelerometers strategically placed on the maternal abdomen. The novelty of the work resides in the efficient removal of artefacts and in distinctive feature extraction. The proposed algorithm used independent component analysis (ICA) for dimensionality reduction and artefact removal. A supplemental technique based on discrete wavelet transform (DWT) was also used to remove artefacts.
To identify fetal movements, 31 features were extracted from the acceleration data. Based on these features, several classifiers were used to distinguish fetal from non-fetal movements. Robustness of the classifiers was tested for various concentrations of artefacts in the classification data. The best performance was achieved by Bagging classifier algorithm, with random forest as its basis classifier, yielding an accuracy ranging from 87.6% to 95.8% depending on the artefact concentration level.
A high performance detection of fetal movements can be achieved using accelerometery-based systems suitable for long-term monitoring.
A Wearable System for In-Home and Long-Term Assessment of Fetal Movement
2020, IRBM
Citation Excerpt :
Besides, in order to feel her baby's activities, the mother has to keep quiet and stay still, causing a large impact on her daily routine. In recent years, with the advancement of microelectronic technology as well as the development in signal processing, automatic detection of fetal movements by using accelerometers and advanced signal processing technologies has gained a lot of attention [9–19]. When contacted with the maternal abdominal wall, a movement of fetal body parts with sufficient force generates vibrations that could be detected by one or a set of accelerometers placed on the surface of the abdomen.
Objectives: This paper presents a novel wearable system for in-home and long-term fetal movement monitoring on a reliable and easily accessible basis.
Material and methods: The system mainly consists of four accelerometers for fetal movement signal acquisition, a microcontroller for signal processing and an Android-based device interacting with the microcontroller via Bluetooth Low Energy (BLE), providing the mother with information related to the fetal movement in an intelligible way.
Results: The proposed system can deliver reliable results with a specificity of 0.99 and a sensitivity of 0.77 for fetal movement time series signal classification.
Conclusion: The proposed wearable system will provide a good alternative to optimize the use of medical professionals and hospital resources, and has potential applications in the field of e-Health home care. Besides, the fetal movement acceleration signals acquired with volunteers (pregnant women) help establish an initial database for future medical analysis of sensor-recorded fetal behaviors.
Surface electromyography based muscle fatigue detection using high-resolution time-frequency methods and machine learning algorithms
2018, Computer Methods and Programs in Biomedicine
Surface electromyography (sEMG) based muscle fatigue research is widely preferred in sports science and occupational/rehabilitation studies due to its noninvasiveness. However, these signals are complex, multicomponent and highly nonstationary with large inter-subject variations, particularly during dynamic contractions. Hence, time-frequency based machine learning methodologies can improve the design of automated system for these signals.
In this work, the analysis based on high-resolution time-frequency methods, namely, Stockwell transform (S-transform), B-distribution (BD) and extended modified B-distribution (EMBD) are proposed to differentiate the dynamic muscle nonfatigue and fatigue conditions. The nonfatigue and fatigue segments of sEMG signals recorded from the biceps brachii of 52 healthy volunteers are preprocessed and subjected to S-transform, BD and EMBD. Twelve features are extracted from each method and prominent features are selected using genetic algorithm (GA) and binary particle swarm optimization (BPSO). Five machine learning algorithms, namely, naïve Bayes, support vector machine (SVM) of polynomial and radial basis kernel, random forest and rotation forests are used for the classification.
The results show that all the proposed time-frequency distributions (TFDs) are able to show the nonstationary variations of sEMG signals. Most of the features exhibit statistically significant difference in the muscle fatigue and nonfatigue conditions. The maximum number of features (66%) is reduced by GA and BPSO for EMBD and BD-TFD respectively. The combination of EMBD- polynomial kernel based SVM is found to be most accurate (91% accuracy) in classifying the conditions with the features selected using GA.
The proposed methods are found to be capable of handling the nonstationary and multicomponent variations of sEMG signals recorded in dynamic fatiguing contractions. Particularly, the combination of EMBD- polynomial kernel based SVM could be used to detect the dynamic muscle fatigue conditions.
Time–frequency analysis in infant cry classification using quadratic time frequency distributions
2018, Biocybernetics and Biomedical Engineering
Citation Excerpt :
Hence in the literature, generally the MFCCs and LPCs feature extraction is enclosed with the dimensionality reduction process [17–20]. Generally, in applications involving the multi component signals like infant cry signals, the time–frequency (t–f) based techniques are demonstrated as knowledgeable rather than the methods based on time or frequency domains [23–25]. Nevertheless, the application of t–f based approaches in the development of infant cry based classification tools are not greatly highlighted and is identifies as a gap in this line of research.
This paper presents a new investigation of time–frequency (t–f) based signal processing approach using quadratic time–frequency distributions (QTFDs) namely spectrogram (SPEC), Wigner–Ville distribution (WVD), Smoothed–Wigner Ville distribution (SWVD), Choi–William distribution (CWD) and modified B-distribution (MBD) for classification of infant cry signals. t–f approaches have proved as an efficient approach for applications involving the non stationary signals. In feature extraction, a cluster of t–f based features were extracted by extending the time-domain and frequency-domain features to the joint t–f domain from the generated t–f representation. Conventional features such as mel-frequency cepstral coefficients (MFCCs) and linear prediction coefficients (LPCs) were also extracted in order to compare the effectiveness of the t–f methods. The efficacy of the extracted feature vectors was validated using probabilistic neural network (PNN) and general regression neural network (GRNN). The proposed methodology was implemented to classify different sets of binary classification problems of infant cry signals from different native. The best empirical result of above 90% was reported and revealed the good potential of t–f methods in the context of infant cry classification.
An efficient inverse short-time Fourier transform algorithm for improved signal reconstruction by time-frequency synthesis: Optimality and computational issues
2017, Digital Signal Processing: A Review Journal
This paper presents an improved signal reconstruction method based on a new inverse short-time Fourier transform (ISTFT) estimator. The main challenge addressed in this study is to design a more computationally efficient algorithm called exact formal approach (EFA) which overcomes the drawbacks of the popular overlap and add (OLA) and least squares (LS) methods in several cases of practical interest. The proposed EFA algorithm is based on a vector formulation and the exploitation of properties of the matrix formed by signal samples and frames corresponding to signals segments with overlap. A detailed comparative study shows the advantages of the EFA compared to the OLA and LS methods. Several experiments illustrate the performance and properties of the different estimators. The criteria of comparison are based on synthesis quality and denoising efficiency. The results indicate that, (1) from a computational point of view, the proposed algorithm EFA outperforms other popular ISTFT algorithms including OLA and LS and (2) that the EFA and LS have similar results in terms of synthesis quality and both outperform the algorithm currently most used for ISTFT, the OLA. The proposed EFA estimator can then improve ISTFT based applications involving signal enhancement and denoising.

View all citing articles on Scopus

Mohamed Salah Khlif received his BS (1992) and MS (1994) in Mechanical Engineering from The University of Nebraska, Lincoln, USA. He worked (1996–2004) in the data storage industry as a Laser Process Development Engineer and holds nine US patents and two registered trade secret applications. In 2012, he received a PhD in Medical Science, with emphasis on biomedical signal processing, from The University of Queensland, Brisbane, Australia. His research interests include data mining and time–frequency and statistical approaches to signal processing.

Taoufik Ben-Jabeur received the M.S. and Ph.D. degrees in Computer Science and Signal Processing from University of Paris-Descartes, France, in 2005 and 2009, respectively. He was a graduate Teaching Assistant from October 2008 to August 2010 at the same institute. In 2010 through 2011, he joined the University of Reims, France, as a Postdoc in the area of wireless multicarrier communications and sensor networks. Since 2011, he has been with Qatar University as a Postdoc. His research interests include wireless multicarrier communications and time–frequency analysis.

Christine E. East received a MMedSc and PhD from the University of Queensland for her clinical research involving fetal pulse oximetry. Her longstanding interest in methods of monitoring fetal welfare extended from the pulse oximetry work to the development of a clinically robust system for monitoring fetal movements, through her collaboration between the University of Melbourne and the University of Queensland. Now a Professor of Midwifery at Monash University and Honorary Professor at the University of Melbourne, her ongoing research includes the evaluation of fetal scalp blood sampling to determine fetal wellbeing during childbirth.

Paul B. Colditz is Professor of Perinatal Medicine at the University of Queensland. He received a Master degree in Biomedical Engineering from the University of NSW and then a DPhil from Oxford University, studying brain circulation in the newborn. He has continued to undertake research involving the application of technology to clinical problems of the fetus and newborn, with the current focus being prevention of death and brain injury. He has key roles in pediatric research education, medical philanthropy, international research, industry scientific roles and clinical practice.

View full text

Passive detection of accelerometer-recorded fetal movements using a time–frequency signal processing approach

Abstract

Section snippets

Identifying the at-risk fetus

Participants, ethics and data

Time–frequency formulation

Performance measures and assumptions

Conclusions

Acknowledgements

Lancet

Lancet

Am. J. Obstet. Gynecol.

Early Hum. Dev.

Early Hum. Dev.

Clin. Obstet. Gynaecol.

Obstet. Gynecol. Clin. North Am.

Am. J. Obstet. Gynecol.

Early Hum. Dev.

Early Hum. Dev.

Early Hum. Dev.

J. Biomed. Eng.

Early Hum. Dev.

Am. J. Obstet. Gynecol.

J. Franklin Inst.

Med. Eng. Phys.

Pattern Recogn. Lett.

Concerns for decreased foetal movements in uncomplicated pregnancies – increased risk of foetal growth restriction and stillbirth among women being overweight advanced age or smoking

J. Matern.-Fetal Neonatal Med.

Australiaʼs mothers and babies 2009

Utero-placental Doppler ultrasound for improving pregnancy outcome

Cochrane Database Syst. Rev.

A critical appraisal of the use of umbilical artery Doppler ultrasound in high-risk pregnancies: Use of meta-analyses in evidence-based obstetrics

Ultrasound Obstet. Gynecol.

Restricted fetal growth in sudden intrauterine unexplained death

Acta Obstet. Gynecol. Scand.

Screening and triage of intrauterine growth restriction (IUGR) in general population and high risk pregnancies: a systematic review with a focus on reduction of IUGR related stillbirths

BMC Public Health

Maternal and fetal risk factors for stillbirth: population based study

BMJ

Stillbirth and fetal growth restriction

J. Matern.-Fetal Neonatal Med.

Time–Frequency Signal Analysis and Processing: A Comprehensive Reference

Analog Devices, ADXL 330, 3-axis ± g iMEMS accelerometers, D05677-0-6/07(A)

Assessment of fetal heart rate and fetal movements in detecting oxygen deprivation in-utero

Eur. J. Obstet. Gynecol. Reprod. Biol.

Antenatal evaluation of the fetus using fetal movement monitoring

Clin. Obstet. Gynaecol.

Treatment of sleep disordered breathing reverses low fetal activity levels in preeclampsia

Sleep