Physiological measurement最新文献

Objective. Risk stratification of hypertension plays a crucial role in the treatment decisions and medication guidance during clinical practices. Although fruitful achievements have been reported on risk stratification of hypertension, the potential use of ambulatory blood pressure monitoring data is not well investigated. Different from single measuring blood pressure data, long-term blood pressure monitoring data can provide more comprehensive dynamical blood pressure information. Therefore, this paper proposes an intelligent hypertension risk stratification method based on ambulatory blood pressure monitoring data and improved machine learning algorithms.Approach. A total of 262 patients with hypertension are enrolled at People's Hospital of Yangjiang, in which 93 subjects are with simple hypertension and 169 subjects have hypertension with complication. Time-domain features, frequency-domain features, nonlinear dynamics features and correlation features underlying time-varying ambulatory blood pressure monitoring data are extracted to obtain discriminative feature representations. Synthetic minority over-sampling algorithm is applied to solve the problem of data balancing. The particle swarm optimization combined with kernel extreme learning machine is employed for feature fusion and optimization.Main results. The proposed method can yield a diagnostic accuracy of 93.7%, 97.8%, and 98.4% under two-, five- and ten-fold cross-validation, which demonstrates hypertension risk stratification in an intuitive, quantizable manner using multi-dimensional feature representation and learning.Significance. The proposed method is expected to provide early warning for latent serious cardiovascular diseases before obvious symptoms are present.

Objective: Respiratory rate (RR) is an important vital sign but is often neglected. Multiple technologies exist for RR monitoring but are either expensive or impractical. Tri-axial accelerometry represents a minimally intrusive solution for continuous RR monitoring, however, the method has not been validated in a wide RR range. Therefore, the aim of this study was to investigate the agreement between RR estimation from a tri-axial accelerometer and a reference method in a wide RR range. Approach. Twenty-five healthy participants were recruited. For accelerometer RR estimation, the accelerometer was placed on the abdomen for optimal breathing movement detection. The acquired accelerometry data were processed using a lowpass filter, principal component analysis (PCA), and autocorrelation. The subjects were instructed to breathe at slow, normal, and fast paces in segments of 60 seconds. A flow meter was used as reference. Main results. Furthermore, the PCA-autocorrelation method was compared with a similar single axis method. The PCA-autocorrelation method resulted in a bias of 0.0 breaths per minute (bpm) and limits of agreement (LOA) = [-1.9; 1.9 bpm] compared to the reference. Overall, 99% of the RRs estimated by the PCAautocorrelation method were within ±2 bpm of the reference. A Pearson correlation indicated a very strong correlation with r = 0.99 (p<0.001). The single axis method resulted in a bias of 3.7 bpm, LOA = [-14.9; 22.3 bpm], and r = 0.44 (p<0.001). Significance. The results indicate a strong agreement between the PCA-autocorrelation method and the reference. Furthermore, the PCA-autocorrelation method outperformed the single axis method.

Atrial fibrillation (AFib) is a prevalent cardiac arrhythmia associated with significant morbidity and mortality. Early detection and continuous monitoring of AFib are crucial for prevention of complications such as stroke. In this paper, we explore the potential of using a ballistocardiogram (BCG) based bed-sensor to detect AFib through a comprehensive clinical study consisting of night hospital recordings for 116 patients split into 72 training subjects and 44 test subjects. The study uses established methods such as autocorrelation in order to detect AFib from the BCG signals. Spot and continuous Holter ECG were used as the reference methods to detect AFib against which the BCG rhythm classifications were compared Our findings suggest that this innovative approach holds promise for accurate and non-invasive continuous monitoring of AFib, contributing to improved patient care and outcomes. With full overnight recordings we were able to detect AFib with 94% accuracy with the train set by using a rule-based method and achieving AUROC score of 97% for the test set using a machine learning model trained with the training set.

Objective: Forced expiratory volume in one second (FEV1) is an important metric for patients to track at home for their self-management of asthma and chronic obstructive pulmonary disease (COPD). Unfortunately, the state-of-the art for measuring FEV1 at home either depends on the patient's physical effort and motivation, or relies on bulky wearable devices that are impractical for long-term monitoring. This paper explores the feasibility of using a machine learning model to infer FEV1 from 270 seconds of a single-lead electrocardiogram (ECG) signal measured on the fingers with a mobile device.

Methods: We evaluated the model's inferred FEV1 values against the ground truth of hospital-grade spirometry tests, which were performed by twenty-five patients with obstructive respiratory disease.

Results: The model-inferred FEV1 compared to the spirometry-measured FEV1 with a correlation coefficient of r = 0.73, a mean absolute percentage error of 23% and a bias of -0.08.

Conclusions: These results suggest that the ECG signal contains useful information about FEV1, although a larger, richer dataset might be necessary to train a machine learning model that can extract this information with better accuracy.

Significance: The benefit of a mobile ECG-based solution for measuring FEV1 is that it would require minimal effort, thus encouraging patient adherence and promoting successful self-management of asthma and COPD.

Objective.Occult hemorrhage (OH) can emerge subtly post-trauma, especially when internal bleeding is not yet severe enough to result in noticeable hemodynamic changes or shock. Despite normal appearances of traditional vital signs like heart rate (HR) and blood pressure (BP), clinically significant OH may be present, posing a critical diagnostic challenge. Early detection of OH, before vital signs begin to deteriorate, is vital as delays in identifying such conditions are linked to poorer patient outcomes. We analyze the performance of poly-anatomic multivariate technologies-including electrical impedance tomography (EIT), near-infrared spectroscopy (NIRS), electrical impedance spectroscopy (EIS), plethysmography (Pleth), and ECG-in a porcine model of OH. The goal was to detect OH without the need to know the subject's pre-established normal baseline.Approach.Forty female swine were bled at slow rates to create an extended period of subclinical hemorrhage, during which the animals' HR and BP remained stable before hemodynamic deterioration. Continuous vital signs, Pleth, and continuous non-invasive data were recorded and analyzed with the objective of developing an improved means of detecting OH. This detection was set up as a supervised voting classification problem where the measurement of each technology (minimally transformed) was used to train a classifier. A soft majority voting classification technique was then used to detect the existence of OH.Main Results.When comparing the prediction performance of the most significant univariate technology (EIT) to that of a poly-anatomic multivariate approach, the latter achieved higher area-under-the-curve (AUC) values from receiver operating characteristic analyses in almost every observation interval duration. In particular, after 21 min of continuous observation, the best AUC of the multivariate approach was 0.98, while that of the univariate approach was 0.92. The best multivariate technologies, in descending order, appeared to be EIT on the thorax, NIRS on the abdomen, and EIS on the thorax.Significance.In this clinically relevant porcine model of clinically OH, multivariate non-invasive measurements may be superior to univariate ones in detecting OH. Advanced technologies such as EIT, NIRS, and EIS exhibit considerably greater potential to accurately predict OH than standard physiological measurements. From a practical standpoint, our approach would not require the medical device to have prior access to non-hemorrhage baseline data for each patient. Early detection of OH using these technologies could improve patient outcomes by allowing for timely intervention before vital signs begin to deteriorate.

Objective: Electrical impedance tomography (EIT) generates cross-sectional images through non-invasive impedance measurements from surface electrodes. While impedance above 200 kHz can reveal intracellular properties, most existing EIT images are published at frequencies below 200 kHz. When frequencies exceed 200 kHz, the accuracy of impedance measurements declines due to the influence of distributed circuit parameters such as parasitic capacitance, on-resistance of switch and the series inductance, with a more significant impact on larger impedance. To overcome this limitation, this paper proposes an approach for precision impedance measurement through self-identification of distributed parameter.

Approach: Firstly, the distributed circuit parameters are identified via correction measurements of precision resistances in the frequency range from 5 kHz to 1 MHz; then, the circuit is accurately modeled; finally, transfer impedance measurements during imaging process are corrected using the established circuit model.

Main results: The distributed circuit parameter self-identification method was verified through a goodness-of-fit test, confirming the consistency between the model's predicted values and the actual values of the component. The test results indicate that at 1 MHz, the relative residuals follow a right-skewed distribution with an average value of 0.08%, which demonstrates high model accuracy. At 1 MHz, the measurement relative error after correction for the 499 Ω precision resistor is reduced by 12.01%, and for the 56 pF precision capacitor, the relative error after correction is 0.46%.

Significance: The proposed method can extend the frequency range of EIT and other impedance technologies from below 200 kHz to up to 1 MHz, while ensuring good measurement accuracy.

Objective. Electrodermal activity (EDA) is a marker of psychophysiological arousal and is usually a measure of the skin conductance which is associated with sweat gland activity. Recent studies have shown that it is possible to estimate the EDA using contactless video based methods.Approach. Sensor EDA signals (SenEDA) and videos of the the palm were recorded simultaneously from over 30 participants under various stimuli (audio, video, cognitive and physiological). The luminance information from the video data was used to track sweat gland activity on the skin surface and extract the contactless signal luminance based EDA (LumEDA).Main results. Comparison of the SenEDA and LumEDA signals showed a high positive correlation between the two as expected.Significance. Under suitable illumination, simple spatial filters can be used to track sweat gland activity which can then be used to estimate signals analogous to the EDA. Such video based methods also facilitate spatio-temporal analysis of EDA correlates over larger areas of the body.

Objective: Photoplethysmography is widely used for physiological monitoring, whether in clinical devices such as pulse oximeters, or consumer devices such as smartwatches. A key step in the analysis of photoplethysmogram (PPG) signals is detecting heartbeats. The MSPTD algorithm has been found to be one of the most accurate PPG beat detection algorithms, but is less computationally efficient than other algorithms. Therefore, the aim of this study was to develop a more efficient, open-source implementation of the MSPTD algorithm for PPG beat detection, named MSPTDfast (v.2).

Approach: Five potential improvements to MSPTD were identified and evaluated on four datasets. MSPTDfast (v.2) was designed by incorpo- rating each improvement which on its own reduced execution time whilst maintaining a high F1-score. After internal validation, MSPTDfast (v.2) was benchmarked against state-of-the-art beat detection algorithms on four additional datasets.

Main results: MSPTDfast (v.2) incorporated two key improvements: pre-processing PPG signals to reduce the sampling frequency to 20 Hz; and only calculating scalogram scales corresponding to heart rates >30 bpm. During internal validation MSPTDfast (v.2) was found to have an execution time of between approximately one-third and one-twentieth of MSPTD, and a comparable F1-score. During benchmarking MSPTDfast (v.2) was found to have the highest F1-score alongside MSPTD, and amongst one of the lowest execution times with only MSPTDfast (v.1), qppgfast and MMPD (v.2) achieving shorter execution times.

Significance: MSPTDfast (v.2) is an accurate and efficient PPG beat detection algorithm, available in an open-source Matlab toolbox.

Objective: Electrical impedance tomography (EIT) has shown the ability to provide clinically useful functional information on ventilation in humans and other land mammals. We are motivated to use EIT with sea mammals and human divers, since EIT could provide unique information on lung ventilation that can help address diver performance and safety, and veterinary and behavioural questions. However, in-water use of EIT is challenging, primarily because sea water is more conductive than the body.

Approach: We first address this issue by modelling the in-water component and evaluating image reconstruction algorithms.

Main results: EIT is able to produce reasonable images if an outer insulating layer allows a water layer thickness < 2% of the body radius. We next describe the design of custom EIT belts with an outer neoprene insulator to minimize current leakage. We show example underwater EIT recordings in human and dolphin subjects.

Significance: we demonstrate in-water EIT is feasible with appropriate techniques. .

Objective.The detection of arterial pulsating signals at the skin periphery with Photoplethysmography (PPG) are easily distorted by motion artifacts. This work explores the alternatives to the aid of PPG reconstruction with movement sensors (accelerometer and/or gyroscope) which to date have demonstrated the best pulsating signal reconstruction.Approach.A generative adversarial network with fully connected layers is proposed for the reconstruction of distorted PPG signals. Artificial corruption was performed to the clean selected signals from the BIDMC Heart Rate dataset, processed from the larger MIMIC II waveform database to create the training, validation and testing sets.Main results.The heart rate (HR) of this dataset was further extracted to evaluate the performance of the model obtaining a mean absolute error of 1.31 bpm comparing the HR of the target and reconstructed PPG signals with HR between 70 and 115 bpm.Significance.The model architecture is effective at reconstructing noisy PPG signals regardless the length and amplitude of the corruption introduced. The performance over a range of HR (70-115 bpm), indicates a promising approach for real-time PPG signal reconstruction without the aid of acceleration or angular velocity inputs.