An ML-Based Approach to Reconstruct Heart Rate from PPG in Presence of Motion Artifacts

Vicente-Samper, José María; Tamantini, Christian; Ávila-Navarro, Ernesto; Casa-Lillo, Miguel A. de la de la; Zollo, Loredana; Navarro, José María Sabater; Cordella, Francesca

doi:10.3390/bios13070718

Cited by 7 publications

(5 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Second, short-length segments can make the model more efficient because they require less data to train the model, which can lead to faster training times and lower computational costs. Finally, short-length segments can make the model more interpretable and easier to understand than long sequences, which can help to identify the patterns that the model is learning [ 15 ].…”

Section: Discussionmentioning

confidence: 99%

“…The use of SpO2 only for OSA screening has several benefits [ 14 ] Pulse oximetry is non-invasive, cost-effective, portable and convenient. This means that it does not require any electrodes or sensors to be placed on the body, and it does not cost as much as other methods, and can be used in home settings [ 15 ]. Theoretical research on the use of SpO2 only to automate OSA scoring has been limited.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Adoption of Transformer Neural Network to Improve the Diagnostic Performance of Oximetry for Obstructive Sleep Apnea

Almarshad,

Al-Ahmadi,

Islam

et al. 2023

Sensors

View full text Add to dashboard Cite

Scoring polysomnography for obstructive sleep apnea diagnosis is a laborious, long, and costly process. Machine learning approaches, such as deep neural networks, can reduce scoring time and costs. However, most methods require prior filtering and preprocessing of the raw signal. Our work presents a novel method for diagnosing obstructive sleep apnea using a transformer neural network with learnable positional encoding, which outperforms existing state-of-the-art solutions. This approach has the potential to improve the diagnostic performance of oximetry for obstructive sleep apnea and reduce the time and costs associated with traditional polysomnography. Contrary to existing approaches, our approach performs annotations at one-second granularity. Allowing physicians to interpret the model’s outcome. In addition, we tested different positional encoding designs as the first layer of the model, and the best results were achieved using a learnable positional encoding based on an autoencoder with structural novelty. In addition, we tried different temporal resolutions with various granularity levels from 1 to 360 s. All experiments were carried out on an independent test set from the public OSASUD dataset and showed that our approach outperforms current state-of-the-art solutions with a satisfactory AUC of 0.89, accuracy of 0.80, and F1-score of 0.79.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Adoption of Transformer Neural Network to Improve the Diagnostic Performance of Oximetry for Obstructive Sleep Apnea

Almarshad,

Al-Ahmadi,

Islam

et al. 2023

Sensors

View full text Add to dashboard Cite

show abstract

“…However, not all EMG portable devices share a similar protocol [65][66][67]. Once the performance of portable EMG has been improved, it could be used for concordance studies against other types of novel tools that are emerging due to the evolution of technology, big data, and artificial intelligence [68][69][70]. Such studies would allow them to be performed longitudinally and more fluently than with PSG in a sleep lab.…”

Section: Discussionmentioning

confidence: 99%

Diagnostic Accuracy of a Portable Electromyography and Electrocardiography Device to Measure Sleep Bruxism in a Sleep Apnea Population: A Comparative Study

Cid-Verdejo,

Domínguez Gordillo,

Sánchez-Romero

et al. 2023

Clocks & Sleep

View full text Add to dashboard Cite

Background: The gold standard for diagnosing sleep bruxism (SB) and obstructive sleep apnea (OSA) is polysomnography (PSG). However, a final hypermotor muscle activity often occurs after apnea episodes, which can confuse the diagnosis of SB when using portable electromyography (EMG) devices. This study aimed to compare the number of SB episodes obtained from PSG with manual analysis by a sleep expert, and from a manual and automatic analysis of an EMG and electrocardiography (EKG) device, in a population with suspected OSA. Methods: Twenty-two subjects underwent a polysomnographic study with simultaneous recording with the EMG-EKG device. SB episodes and SB index measured with both tools and analyzed manually and automatically were compared. Masticatory muscle activity was scored according to published criteria. Patients were segmented by severity of OSA (mild, moderate, severe) following the American Academy of Sleep Medicine (AASM) criteria. ANOVA and the Bland–Altman plot were used to quantify the agreement between both methods. The concordance was calculated through the intraclass correlation coefficient (ICC). Results: On average, the total events of SB per night in the PSG study were (8.41 ± 0.85), lower than the one obtained with EMG-EKG manual (14.64 ± 0.76) and automatic (22.68 ± 16.02) analysis. The mean number of SB episodes decreases from the non-OSA group to the OSA group with both PSG (5.93 ± 8.64) and EMG-EKG analyses (automatic = 22.47 ± 18.07, manual = 13.93 ± 11.08). However, this decrease was minor in proportion compared to the automatic EMG-EKG analysis mode (from 23.14 to 22.47). The ICC based on the number of SB episodes in the segmented sample by severity degree of OSA along the three tools shows a moderate correlation in the non-OSA (0.61) and mild OSA (0.53) groups. However, it is poorly correlated in the moderate (0.24) and severe (0.23) OSA groups: the EMG-EKG automatic analysis measures 14.27 units more than PSG. The results of the manual EMG-EKG analysis improved this correlation but are not good enough. Conclusions: The results obtained in the PSG manual analysis and those obtained by the EMG-EKG device with automatic and manual analysis for the diagnosis of SB are acceptable but only in patients without OSA or with mild OSA. In patients with moderate or severe OSA, SB diagnosis with portable electromyography devices can be confused due to apneas, and further study is needed to investigate this.

show abstract

“…Meanwhile, both sound-sensing and remote-based methods are in the research and development (R&D) stage, which cannot be applicable commercially in a short time [12]. Despite ongoing research on the reconstruction of the heart rate from photoplethysmography (PPG) while accounting for motion artifacts through a machine-learning-based approach [19], the diag-nosis of OSA necessitates the development of a distinct algorithm for the transformation of the heart rate into the OSA diagnostic indicator. Furthermore, the accuracy of such an algorithm is not guaranteed.…”

Section: Psg Accuratementioning

confidence: 99%

Minimally Invasive Hypoglossal Nerve Stimulator Enabled by ECG Sensor and WPT to Manage Obstructive Sleep Apnea

Xia,

Li,

et al. 2023

Sensors

View full text Add to dashboard Cite

A hypoglossal nerve stimulator (HGNS) is an invasive device that is used to treat obstructive sleep apnea (OSA) through electrical stimulation. The conventional implantable HGNS device consists of a stimuli generator, a breathing sensor, and electrodes connected to the hypoglossal nerve via leads. However, this implant is bulky and causes significant trauma. In this paper, we propose a minimally invasive HGNS based on an electrocardiogram (ECG) sensor and wireless power transfer (WPT), consisting of a wearable breathing monitor and an implantable stimulator. The breathing external monitor utilizes an ECG sensor to identify abnormal breathing patterns associated with OSA with 88.68% accuracy, achieved through the utilization of a convolutional neural network (CNN) algorithm. With a skin thickness of 5 mm and a receiving coil diameter of 9 mm, the power conversion efficiency was measured as 31.8%. The implantable device, on the other hand, is composed of a front-end CMOS power management module (PMM), a binary-phase-shift-keying (BPSK)-based data demodulator, and a bipolar biphasic current stimuli generator. The PMM, with a silicon area of 0.06 mm2 (excluding PADs), demonstrated a power conversion efficiency of 77.5% when operating at a receiving frequency of 2 MHz. Furthermore, it offers three-voltage options (1.2 V, 1.8 V, and 3.1 V). Within the data receiver component, a low-power BPSK demodulator was ingeniously incorporated, consuming only 42 μW when supplied with a voltage of 0.7 V. The performance was achieved through the implementation of the self-biased phase-locked-loop (PLL) technique. The stimuli generator delivers biphasic constant currents, providing a 5 bit programmable range spanning from 0 to 2.4 mA. The functionality of the proposed ECG- and WPT-based HGNS was validated, representing a highly promising solution for the effective management of OSA, all while minimizing the trauma and space requirements.

show abstract

An ML-Based Approach to Reconstruct Heart Rate from PPG in Presence of Motion Artifacts

Cited by 7 publications

References 35 publications

Adoption of Transformer Neural Network to Improve the Diagnostic Performance of Oximetry for Obstructive Sleep Apnea

Adoption of Transformer Neural Network to Improve the Diagnostic Performance of Oximetry for Obstructive Sleep Apnea

Diagnostic Accuracy of a Portable Electromyography and Electrocardiography Device to Measure Sleep Bruxism in a Sleep Apnea Population: A Comparative Study

Minimally Invasive Hypoglossal Nerve Stimulator Enabled by ECG Sensor and WPT to Manage Obstructive Sleep Apnea

Contact Info

Product

Resources

About