A system for investigating oesophageal photoplethysmographic signals in anaesthetised patients

The monitoring of arterial blood oxygen saturation in patients with compromised peripheral perfusion is often difficult, because conventional non-invasive techniques such as pulse oximetry (SpO2) can fail. Poor peripheral circulation commonly occurs after major surgery including cardiopulmonary bypass. The difficulties in these clinical situations might be overcome if the sensor were to monitor a better perfused central part of the body such as the oesophagus. A new oesophageal photoplethysmographic (PPG) probe and an isolated processing system have been developed to investigate the pulsatile signals of anaesthetised adult patients undergoing routine surgery. Measurements were made in the middle third of the oesophagus, 25 cm to 30 cm from the upper incisors. The AC PPG signals are sampled by a data acquisition system connected to a laptop computer. The signals recorded correspond to infrared and red AC PPGs from the middle third oesophagus and the finger. Preliminary results from 20 patients show that good quality AC PPG signals can be measured in the human oesophagus. The ratio of the oesophageal to finger AC PPG amplitudes was calculated for the infrared and red wavelengths for each patient. The mean (+/- standard deviation) of this ratio was 2.9 +/- 2.1 (n = 19) for the infrared wavelength and 3.1 +/- 2.4 (n = 16) for the red wavelength. The red and infrared wavelengths used are appropriate for pulse oximetry and this investigation indicates that the mid-oesophagus may be a suitable site for the reliable monitoring of SpO2 in patients with poor peripheral perfusion.     

Filtering techniques for the removal of ventilator artefact in oesophageal pulse oximetry

The oesophagus has been shown to be a reliable site for monitoring blood oxygen saturation (SpO₂). However, the photoplethysmographic (PPG) signals from the lower oesophagus are frequently contaminated by a ventilator artefact making the estimation of SpO₂ impossible. A 776th order finite impulse response (FIR) filter and a 695th order interpolated finite impulse response (IFIR) filter were implemented to suppress the artefact. Both filters attenuated the ventilator artefact satisfactorily without distorting the morphology of the PPG when processing recorded data from ten cardiopulmonary bypass patients. The IFIR filter was the better since it conformed more closely to the desired filter specifications and allowed real-time processing. The average improvements in signal-to-noise ratio (SNR) achieved by the FIR and IFIR filters for the fundamental component of the red PPG signals with respect to the fundamental component of the artefact were 57.96 and 60.60 dB, respectively. The corresponding average improvements achieved by the FIR and IFIR filters for the infrared PPG signals were 54.83 and 60.96 dB, respectively. Both filters were also compared with their equivalent tenth order Butterworth filters. The average SNR improvements for the FIR and IFIR filters were significantly higher than those for the Butterworth filters.     

Optical assessment of recovery of tissue blood supply after removal of externally applied pressure

The authors use photoelectric plethysmography to determine the external occlusion pressure for blood vessels in human tissue in vivo. Three wavelengths are employed; 950 nm (infra-red), 640 nm (red) and 583 nm (yellow). Each probe is applied in turn to one finger of each subject. Pressure is applied, using a neonatal blood pressure cuff, to the finger via the probe. This pressure is increased linearly to 20 kPa (150 mmHg) over 15 s and then decreased linearly to zero over 15 s. The pressure at which perfusion returns is obtained for four repeat measurements at each wavelength. The mean (±standard deviation) occlusion pressures for all 13 subjects investigated are 7.1(±1.9) kPa for infra-red, 6.3(±1.7) kPa for red and 5.8(±1.8) kPa for yellow. The pressure is 0.79(±0.83) kPa lower for red compared with infra-red (P<0.01), 0.54(±0.60) kPa lower for yellow compared with red (P<0.002) and 1.3(±1.0) kPa lower for yellow compared with infra-red (P<0.005). The reduced penetration of shorter optical wavelengths can be used to detect the lower occlusion pressures of the smaller blood vessels nearer the skin surface.     

Investigation of photoplethysmographic signals and blood oxygen saturation values on healthy volunteers during cuff-induced hypoperfusion using a multimode PPG/SpO2 sensor

Photoplethysmography (PPG) is a technique widely used to monitor volumetric blood changes induced by cardiac pulsations. Pulse oximetry uses the technique of PPG to estimate arterial oxygen saturation values (SpO?). In poorly perfused tissues, SpO? readings may be compromised due to the poor quality of the PPG signals. A multimode finger PPG probe that operates simultaneously in reflectance, transmittance and a combined mode called "transreflectance" was developed, in an effort to improve the quality of the PPG signals in states of hypoperfusion. Experiments on 20 volunteers were conducted to evaluate the performance of the multimode PPG sensor and compare the results with a commercial transmittance pulse oximeter. A brachial blood pressure cuff was used to induce artificial hypoperfusion. Results showed that the amplitude of the transreflectance AC PPG signals were significantly different (p < 0.05) than the AC PPG signals obtained from the other two conventional PPG sensors (reflectance and transmittance). At induced brachial pressures between 90 and 135 mmHg, the reflectance finger pulse oximeter failed 25 times (failure rate 42.2 %) to estimate SpO? values, whereas the transmittance pulse oximeter failed 8 times (failure rate 15.5 %). The transreflectance pulse oximeter failed only 3 times (failure rate 6.8 %) and the commercial pulse oximeter failed 17 times (failure rate 29.4 %).     

Photoplethysmograph Signal Reconstruction Based on a Novel Hybrid Motion Artifact Detection–Reduction Approach. Part I: Motion and Noise Artifact Detection

Motion and noise artifacts (MNA) are a serious obstacle in utilizing photoplethysmogram (PPG) signals for real-time monitoring of vital signs. We present a MNA detection method which can provide a clean vs. corrupted decision on each successive PPG segment. For motion artifact detection, we compute four time-domain parameters: (1) standard deviation of peak-to-peak intervals (2) standard deviation of peak-to-peak amplitudes (3) standard deviation of systolic and diastolic interval ratios, and (4) mean standard deviation of pulse shape. We have adopted a support vector machine (SVM) which takes these parameters from clean and corrupted PPG signals and builds a decision boundary to classify them. We apply several distinct features of the PPG data to enhance classification performance. The algorithm we developed was verified on PPG data segments recorded by simulation, laboratory-controlled and walking/stair-climbing experiments, respectively, and we compared several well-established MNA detection methods to our proposed algorithm. All compared detection algorithms were evaluated in terms of motion artifact detection accuracy, heart rate (HR) error, and oxygen saturation (SpO₂) error. For laboratory controlled finger, forehead recorded PPG data and daily-activity movement data, our proposed algorithm gives 94.4, 93.4, and 93.7% accuracies, respectively. Significant reductions in HR and SpO₂ errors (2.3 bpm and 2.7%) were noted when the artifacts that were identified by SVM-MNA were removed from the original signal than without (17.3 bpm and 5.4%). The accuracy and error values of our proposed method were significantly higher and lower, respectively, than all other detection methods. Another advantage of our method is its ability to provide highly accurate onset and offset detection times of MNAs. This capability is important for an automated approach to signal reconstruction of only those data points that need to be reconstructed, which is the subject of the companion paper to this article. Finally, our MNA detection algorithm is real-time realizable as the computational speed on the 7-s PPG data segment was found to be only 7 ms with a Matlab code.     

Photoplethysmograph Signal Reconstruction based on a Novel Motion Artifact Detection-Reduction Approach. Part II: Motion and Noise Artifact Removal

We introduce a new method to reconstruct motion and noise artifact (MNA) contaminated photoplethysmogram (PPG) data. A method to detect MNA corrupted data is provided in a companion paper. Our reconstruction algorithm is based on an iterative motion artifact removal (IMAR) approach, which utilizes the singular spectral analysis algorithm to remove MNA artifacts so that the most accurate estimates of uncorrupted heart rates (HRs) and arterial oxygen saturation (SpO₂) values recorded by a pulse oximeter can be derived. Using both computer simulations and three different experimental data sets, we show that the proposed IMAR approach can reliably reconstruct MNA corrupted data segments, as the estimated HR and SpO₂ values do not significantly deviate from the uncorrupted reference measurements. Comparison of the accuracy of reconstruction of the MNA corrupted data segments between our IMAR approach and the time-domain independent component analysis (TD-ICA) is made for all data sets as the latter method has been shown to provide good performance. For simulated data, there were no significant differences in the reconstructed HR and SpO₂ values starting from 10 dB down to ?15 dB for both white and colored noise contaminated PPG data using IMAR; for TD-ICA, significant differences were observed starting at 10 dB. Two experimental PPG data sets were created with contrived MNA by having subjects perform random forehead and rapid side-to-side finger movements show that; the performance of the IMAR approach on these data sets was quite accurate as non-significant differences in the reconstructed HR and SpO₂ were found compared to non-contaminated reference values, in most subjects. In comparison, the accuracy of the TD-ICA was poor as there were significant differences in reconstructed HR and SpO₂ values in most subjects. For non-contrived MNA corrupted PPG data, which were collected with subjects performing walking and stair climbing tasks, the IMAR significantly outperformed TD-ICA as the former method provided HR and SpO₂ values that were non-significantly different than MNA free reference values.     

Contactless Multiple Wavelength Photoplethysmographic Imaging: A First Step Toward “SpO<Subscript>2</Subscript> Camera” Technology

We describe a route toward contactless imaging of arterial oxygen saturation (SpO₂) distribution within tissue, based upon detection of a two-dimensional matrix of spatially resolved optical plethysmographic signals at different wavelengths. As a first step toward SpO₂-imaging we built a monochrome CMOS-camera with apochromatic lens and 3λ-LED-ringlight (λ₁ = 660 nm, λ₂ = 810 nm, λ₃ = 940 nm; 100 LEDs λ⁻¹). We acquired movies at three wavelengths while simultaneously recording ECG and respiration for seven volunteers. We repeated this experiment for one volunteer at increased frame rate, additionally recording the pulse wave of a pulse oximeter. Movies were processed by dividing each image frame into discrete Regions of Interest (ROIs), averaging 10 × 10 raw pixels each. For each ROI, pulsatile variation over time was assigned to a matrix of ROI-pixel time traces with individual Fourier spectra. Photoplethysmograms correlated well with respiration reference traces at three wavelengths. Increased frame rates revealed weaker pulsations (main frequency components 0.95 and 1.9 Hz) superimposed upon respiration-correlated photoplethysmograms, which were heartbeat-related at three wavelengths. We acquired spatially resolved heartbeat-related photoplethysmograms at multiple wavelengths using a remote camera. This feasibility study shows potential for non-contact 2-D imaging reflection-mode pulse oximetry. Clinical devices, however, require further development.     

Respiratory variations in the reflection mode photoplethysmographic signal. Relationships to peripheral venous pressure

Photoplethysmography (PPG) is a non-invasive optical way of measuring variations in blood volume and perfusion in the tissue, used in pulse oximetry for instance. Respiratory-induced intensity variations (RIIVs) in the PPG signal exist, but the physiological background is not fully understood. Respiration causes variations in the blood volume in the peripheral vascular bed. It was hypothesised that the filling of peripheral veins is one of the important factors involved. In 16 healthy subjects, the respiratory synchronous variations from a PPG reflection mode signal and the peripheral venous pressure (PVP) were recorded. Variations of tidal volume, respiratory rate and contribution from abdominal and thoracic muscles gave significant and similar amplitude changes in both RIIV and the respiratory variation of PVP (p<0.01). The highest amplitudes of both signals were found at the largest tidal volume, lowest respiratory rate and during mainly thoracic breathing, respectively. The coherence between PVP and RIIV signals was high, the median (quartile range) being 0.78 (0.42). Phase analysis showed that RIIV was usually leading PVP, but variations between subjects were large. Although respiratory-induced variations in PVP and PPG showed a close correlation in amplitude variation, a causal relationship between the signals could not be demonstrated.     

Photoplethysmography

When the microvascular blood perfusion in human skin is measured by photoplethysmography (PPG), infra-red light (800-960 nm) is normally used as the light source. The PPG signal, which consists of a pulsatile (AC) and a slowly fluctuating (DC) component, was studied at different optical wavelengths utilising optical fibres for guiding the light to and from the skin surface. Finger and forearm skin was examined and high and low skin blood perfusion was brought about by local water-induced temperature provocation. The analysis of the measurement results provided evidence that the use of shorter wavelengths in PPG (AC) for monitoring skin perfusion changes could be applicable. The use of different optical wavelengths also raises the possibility of recording perfusing changes at different depths in the superficial tissue. The sweat water content in stratum corneum of human skin will probably determine the total amount of reflected and backscattered radiation reaching the photodetector. This is important when the skin perfusion is changed by alterations in the environmental temperature conditions activating the sweat glands in tissue. Temperature-dependent optical characteristics of blood-free skin tissue may explain the limited ability of the DC component of PPG to monitor skin perfusion changes.     

Spectral signature and heterodyne efficiency for different wavelengths in laser Doppler flowmetry

Laser Doppler perfusion monitoring and imaging technologies generate time traces and two-dimensional flow maps of the microcirculation. With the goal of reaching different tissue depths, these technologies are equipped with lassers operating at different wavelengths λ. The fact that the average scattering angle, at a single scattering event, between a photon and a red blood cell increases with λ is compensated for by a 1/λ effect in the scattering vector, rendering the average frequency shift virtually independent of the choice of wavelength. Monte Carlo simulations showed that the corresponding spectral signature of the Doppler signals for λ=632.8nm and 780nm were close to identical. The theoretical predictions were verified by calculating the centre-of-gravity (COG) frequency of the laser Doppler power spectral density for the two wavelengths from forearm and finger skin, representing a low and high perfusion area, respectively (forearm COG=123 against 121Hz, finger COG=220 against 212 Hz). When the wavelength changes from 632.8nm to 780nm, the heterodyne efficiency of the detector and, thereby, the inherent system amplifcation increase. For tissues with identical microvascular flow conditions, the output signal therfore tends to increase in magnitude when shifting to longer wavelengths.     

Computer analysis system of blood oxygen saturation in an animal hypoxia model

An experimental animal hypoxia model has been developed. It consists of two sensors (an in vitro and in vivo model), an experimental device and a computer signal processing system. This method can easily be applied to determine and analyse blood oxygen saturation at various hypoxia levels. It can also be used to evaluate the accuracy of pulse oximetry over a wide range of oxyhemoglobin desaturation levels. The DC and AC components of recorded red and infra-red signals, the dual-wavelength ratio R₁₂ and the reading of a pulse oximeter (SpO₂) can be automatically calculated and displayed on a computer screen. Preliminary results of the animal hypoxia test indicate that the measurements made by the instrument correlate well with the oxygen saturation readings of the automatic blood gas analyser AVL945. The computer analysis system is suitable for repeated estimations in the animal model.     

A comparison of pulse oximetry and cerebral oxygenation in children with severe sleep apnea–hypopnea syndrome: a pilot study

Near infrared spectroscopy (NIRS) has been used to assess the impact of obstructive sleep apnea–hypopnea syndrome (OSAHS) on cerebral oxygenation. However, the relationship between the variations in the cerebral tissue oxygen saturation (ΔTOI) and pulse oximetry (ΔSpO₂) has not been assessed in children with OSAHS. Consecutive clinically stable children with severe OSAHS [apnea–hypopnea index (AHI) >15 events h⁻¹] diagnosed during a night‐time polygraphy with simultaneous recording of cerebral oxygenation with NIRS (NIRO‐200NX, Hamamatsu Photonics KK) were included between September 2015 and June 2016. Maximal ΔSpO₂ (SpO₂ drop from the value preceding desaturation to nadir) and concomitant variations in transcutaneous carbon dioxide (ΔPtcCO₂), maximal ΔTOI and maximal variations in cerebral oxygenated (O₂Hb) and deoxygenated (HHb) haemoglobin were reported. The relationships between ΔSpO₂, ΔPtcCO₂ and ΔTOI, ΔO₂Hb and ΔHHb were investigated. The data from five children (three boys, aged 9.6 ± 6.7 years, AHI 16–91 events h⁻¹) were analysed. Strong correlations were found between ΔSpO₂ and ΔTOI (r = 0.887, P < 0.001), but also with ΔO₂Hb and ΔHHb with a particular pattern in the youngest child with a dark skin pigmentation. Mean ΔSpO₂ was 20 ± 17% and mean ΔTOI was 8 ± 7%. Maximal ΔSpO₂ of approximately 70% were coupled with ΔTOI of no more than 35%. ΔPtcCO₂ correlated only weakly with the cerebral oxygenation indexes. This pilot study shows a strong relationship between pulse oximetry and cerebral oxygenation in children with OSAHS, with lower changes in TOI compared to SpO₂. Future studies should address the clinical impact of respiratory events on cerebral oxygenation and its consequences.     

Non-invasive continuous estimation of blood flow changes in human patellar bone

A photoplethysmographic (PPG) technique to assess blood flow in bone tissue has been developed and tested. The signal detected by the PPG consists of a constant-level (DC) component—which is related to the relative vascularization of the tissue—and a pulsatile (AC) component—which is synchronous with the pumping action of the heart. The PPG probe was applied on the skin over the patella. The probe uses near-infrared (804 nm) and green (560 nm) light sources and the AC component of the PPG signals of the two wavelengths was used to monitor pulsatile blood flow in the patellar bone and the overlying skin, respectively. Twenty healthy subjects were studied and arterial occlusion resulted in elimination of PPG signals at both wavelengths, whereas occlusion of skin blood flow by local surface pressure eliminated only the PPG signal at 560 nm. In a parallel study on a physical model with a rigid tube we showed that the AC component of the PPG signal originates from pulsations of blood flow in a rigid structure and not necessarily from volume pulsations. We conclude that pulsatile blood flow in the patellar bone can be assessed with the present PPG technique.     

In vitro performance test system for pulse oximeters

An in vitro system was developed capable of testing the accuracy and reproducibility of pulse oximeter readings. The pulse oximeter probe receives signals through a pulsating blood cuvette. The development of the design of the cuvette is described. Using the final design (or ‘model finger’), a comparison is made between readings from a Datex Satlite pulse oximeter (SpO₂) and saturation values obtained by use of a multiwavelength bench oximeter (SaO₂). Linear regression analysis of the data gives SpO₂=0·88 SaO₂+11·2 (r=0·979, p<0·001).     

无线传输在脉搏血氧饱和度无创监测中的应用

随着光电子技术的快速发展,基于光电检测原理的血氧饱和度在线测量方法得到了广泛研究。本文对传统双波长透射式血氧饱和度测量方法进行了改进,并在系统中应用了无线数字信号传输和数据小波处理方法。经实验验证,基于以上原理提出的脉搏血氧饱和度检测方法和无线数字信号传输方法,有效地消除了非血氧成分与其他干扰因素对测量结果的影响,并且突破了测量的空间限制。     

Accuracy of pulse oximetry at various haematocrits and during haemolysis in anin vitro model

In situations in which it may be impossible and/or unethical to evaluate pulse oximetry in humans, an in vitro model with circulating blood may be a necessity. The main objective was to develop such an in vitro model and, in this model, validate the pulse oximetry technique at various haematocrit levels. The pulsating character of arterial blood flow in a tubing system was simulated by using a specially constructed pressure-regulated roller pump. The tubing system was designed to minimise damage to red blood cells. The pulse oximeter readings (SpO₂) were compared with oxygen saturation analyses by a haemoximeter (SaO₂). The pulse oximetry readings were recorded at various haematocrit levels and during haemolysis in the SaO₂ range 60–100 per cent. At a haematocrit level of 41–44 per cent, there was no correlation between SaO₂ and SpO₂ readings. After diluting the blood with normal saline to a haematocrit of 10–11 per cent, a good correlation between SaO₂ and SpO₂ was found. Following haemolysis, the agreement between SaO₂ and SpO₂ was further improved. Using the developed in vitro model, the results indicate that the accuracy of a pulse oximeter may be dependent on the haematocrit level.     

Right-left correlation of the sympathetically induced fluctuations of photoplethysmographic signal in diabetic and non-diabetic subjects

Photoplethysmography (PPG) records the cardiac-induced changes in tissue blood volume by light-transmission measurements. The baseline and amplitude of the PPG signal show very low-frequency (VLF) spontaneous fluctuations, which are mediated by the sympathetic nervous system, and high correlation between right and left extremities of healthy subjects. As sympathetic neuropathy is one of the diabetic complications, the right-left correlation of the PPG fluctuations was examined in diabetic patients. The PPG signal was simultaneously measured in the two index fingers and the two second toes of 35 diabetic patients and 33 non-diabetic subjects. For each PPG pulse, the baseline and amplitude were determined, and the right-left correlation coefficients of the VLF fluctuations in the baseline and amplitude were derived. The VLF fluctuations in the baseline showed high right-left correlation, both for fingers (0.93±0.05) and toes (0.93±0.06), for the non-diabetic subjects, and significantly lower correlation (0.78±0.22 and 0.84±0.17, respectively) for the diabetic patients. Similar results were obtained for the amplitude VLF fluctuations. The right-left correlation coefficients for diabetic patients decreased with the disease duration for the toe baseline and toe amplitude fluctuations and correlated with heart rate response to deep breathing for the finger baseline and toe amplitude fluctuations. The right-left correlation coefficients of the PPG fluctuations provide a simple and convenient means for assessing the adequacy of the sympathetic nervous system function.     

基于深度学习的指-桡端脉搏波信号转换方法

针对目前市面上大多数脉搏波检测仪器检测的是指端脉搏波信号,提出一种基于卷积神经网络的指-桡端脉搏波信号转换方法,在仅获取指端脉搏波信号的情况下得到对应的桡动脉脉搏波信号。该方法主要由一维卷积神经网络通过端到端的训练实现,模型包含编码器、解码器和跳跃连接3个部分,通过编码器网络提取指端脉搏波信号的特征,再通过解码器网络将特征图进行扩展,并且利用跳跃连接的方式实现特征图的融合。采集60份指端和桡端的脉搏波信号进行实验,并与传递函数模型和弹性腔模型进行对比。实验结果表明,该模型转换所得的桡端脉搏波信号在MAE和PRD的指标上分别达到1.4%±0.3%和3.6%±1.2%,优于其他模型。研究表明,该模型能够较精确地实现指端脉搏波信号到桡端脉搏波信号的转化。     

Photoplethysmography

Photoplethysmography (PPG) and laser Doppler flowmetry (LDF) were compared and evaluated. The comparison was accomplished considering differences in physical principles and geometrical and optical conditions. Changes in human skin perfusion were induced by cold and hot water provocation on limited areas of the finger and the forearm. The results showed that LDF and PPG, using red light at 630-660 nm and a commercial PPG probe, were in general equally sensitive in detecting a blood perfusion increase following a skin temperature elevation. However, we also found that PPG occasionally showed an inverse response to a skin temperature elevation, especially in finger skin. Furthermore, the study indicated that laser light is unsuitable as a light source in PPG using optical fibres of small diameter and with no fibre separation. It was also found that the physical dimensions of the probe (including the light source and photodetector) play an important role in determining the measuring volume and the quality of the signal.