Assessing the Preservation of Lumbar Lordotic Curvature in Everyday Sitting Conditions Assessed with an Inertial Measurement System.

Background/Objectives: Lumbar lordotic curvature (LLC), closely associated with low back pain (LBP) when decreased, is infrequently assessed in clinical settings due to the spatiotemporal limitations of radiographic methods. To overcome these constraints, this study used an inertial measurement system to compare the magnitude and maintenance of LLC across various sitting conditions, categorized into three aspects: verbal instructions, chair type, and desk task types. Methods: Twenty-nine healthy participants were instructed to sit for 3 min with two wireless sensors placed on the 12th thoracic vertebra and the 2nd sacral vertebra. The lumbar lordotic angle (LLA) was measured using relative angles for the mediolateral axis and comparisons were made within each sitting category. Results: The maintenance of LLA (LLAdev) was significantly smaller when participants were instructed to sit upright (-3.7 ± 3.9°) compared to that of their habitual sitting posture (-1.2 ± 2.4°) (p = 0.001), while the magnitude of LLA (LLAavg) was significantly larger with an upright sitting posture (p = 0.001). LLAdev was significantly larger when using an office chair (-0.4 ± 1.1°) than when using a stool (-3.2 ± 7.1°) (p = 0.033), and LLAavg was also significantly larger with the office chair (p < 0.001). Among the desk tasks, LLAavg was largest during keyboard tasks (p < 0.001), followed by mouse and writing tasks; LLAdev showed a similar trend without statistical significance (keyboard, -1.2 ± 3.0°; mouse, -1.8 ± 2.2°; writing, -2.9 ± 3.1°) (p = 0.067). Conclusions: Our findings suggest that strategies including the use of an office chair and preference for computer work may help preserve LLC, whereas in the case of cueing, repetition may be necessary.

Validation of Electrocardiogram Based Photoplethysmogram Generated Using U-Net Based Generative Adversarial Networks.

Photoplethysmogram (PPG) performs an important role in alarming atrial fibrillation (AF). While the importance of PPG is emphasized, there is insufficient amount of openly available atrial fibrillation PPG data. We propose a U-net-based generative adversarial network (GAN) which synthesize PPG from paired electrocardiogram (ECG). To measure the performance of the proposed GAN, we compared the generated PPG to reference PPG in terms of morphology similarity and also examined its influence on AF detection classifier performance. First, morphology was compared using two different metrics against the reference signal: percent root mean square difference (PRD) and Pearson correlation coefficient. The mean PRD and Pearson correlation coefficient were 27% and 0.94, respectively. Heart rate variability (HRV) of the reference AF ECG and the generated PPG were compared as well. The p-value of the paired t-test was 0.248, indicating that no significant difference was observed between the two HRV values. Second, to validate the generated AF PPG dataset, four different datasets were prepared combining the generated PPG and real AF PPG. Each dataset was used to optimize a classification model while maintaining the same architecture. A test dataset was prepared to test the performance of each optimized model. Subsequently, these datasets were used to test the hypothesis whether the generated data benefits the training of an AF classifier. Comparing the performance metrics of each optimized model, the training dataset consisting of generated and real AF PPG showed a test accuracy result of 0.962, which was close to that of the dataset consisting only of real AF PPG data at 0.961. Furthermore, both models yielded the same F1 score of 0.969. Lastly, using only the generated AF PPG dataset resulted in test accuracy of 0.945, indicating that the trained model was capable of generating valuable AF PPG. Therefore, it can be concluded that the generated AF PPG can be used to augment insufficient data. To summarize, this study proposes a GAN-based method to generate atrial fibrillation PPG that can be used for training atrial fibrillation PPG classification models.

Development of generalizable automatic sleep staging using heart rate and movement based on large databases.

Purpose: With the advancement of deep neural networks in biosignals processing, the performance of automatic sleep staging algorithms has improved significantly. However, sleep staging using only non-electroencephalogram features has not been as successful, especially following the current American Association of Sleep Medicine (AASM) standards. This study presents a fine-tuning based approach to widely generalizable automatic sleep staging using heart rate and movement features trained and validated on large databases of polysomnography. Methods: A deep neural network is used to predict sleep stages using heart rate and movement features. The model is optimized on a dataset of 8731 nights of polysomnography recordings labeled using the Rechtschaffen & Kales scoring system, and fine-tuned to a smaller dataset of 1641 AASM-labeled recordings. The model prior to and after fine-tuning is validated on two AASM-labeled external datasets totaling 1183 recordings. In order to measure the performance of the model, the output of the optimized model is compared to reference expert-labeled sleep stages using accuracy and Cohen's κ as key metrics. Results: The fine-tuned model showed accuracy of 76.6% with Cohen's κ of 0.606 in one of the external validation datasets, outperforming a previously reported result, and showed accuracy of 81.0% with Cohen's κ of 0.673 in another external validation dataset. Conclusion: These results indicate that the proposed model is generalizable and effective in predicting sleep stages using features which can be extracted from non-contact sleep monitors. This holds valuable implications for future development of home sleep evaluation systems.

Estimation and Validation of Arterial Blood Pressure Using Photoplethysmogram Morphology Features in Conjunction With Pulse Arrival Time in Large Open Databases.

Although various predictors and methods for BP estimation have been proposed, differences in study designs have led to difficulties in determining the optimal method. This study presents analyses of BP estimation methods using 2.4 million cardiac cycles of two commonly used non-invasive biosignals, electrocardiogram (ECG) and photoplethysmogram (PPG), from 1376 surgical patients. Feature selection methods were used to determine the best subset of predictors from a total of 42 including PAT, heart rate (HR), and various PPG morphology features, and BP estimation models constructed using linear regression (LR), random forest (RF), artificial neural network (ANN), and recurrent neural network (RNN) were evaluated. 28 features out of 42 were determined as suitable for BP estimation, in particular two PPG morphology features outperformed PAT, which has been conventionally seen as the best non-invasive indicator of BP. By modelling the low frequency component of BP using ANN and the high frequency component using RNN with the selected predictors, mean errors of 0.05 ± 6.92 mmHg for systolic BP, and -0.05 ± 3.99 mmHg for diastolic BP were achieved. External validation of the model using another biosignal database consisting of 334 intensive care unit patients led to similar results, satisfying three standards for accuracy of BP monitors. The results indicate that the proposed method can contribute to the realization of ubiquitous non-invasive continuous BP monitoring.

Reconstruction of 12-Lead Electrocardiogram from a Three-Lead Patch-Type Device Using a LSTM Network.

Reconstructing a standard 12-lead electrocardiogram (ECG) from signals received from electrodes packed into a patch-type device is a challenging task in the field of medical instrumentation. All attempts to obtain a clinically valid 12-lead ECG using a patch-type device were not satisfactory. In this study, we designed the hardware for a three-lead patch-type ECG device and employed a long short-term memory (LSTM) network that can overcome the limitations of the linear regression algorithm used for ECG reconstruction. The LSTM network can overcome the issue of reduced horizontal components of the vector in the electric signal obtained from the patch-type device attached to the anterior chest. The reconstructed 12-lead ECG that uses the LSTM network was tested against a standard 12-lead ECG in 30 healthy subjects and ECGs of 30 patients with pathologic findings. The average correlation coefficient of the LSTM network was found to be 0.95. The ability of the reconstructed ECG to detect pathologic abnormalities was identical to that of the standard ECG. In conclusion, the reconstruction of a standard 12-lead ECG using a three-lead patch-type device is feasible, and such an ECG is an equivalent alternative to a standard 12-lead ECG.

Analysis of Pulse Arrival Time as an Indicator of Blood Pressure in a Large Surgical Biosignal Database: Recommendations for Developing Ubiquitous Blood Pressure Monitoring Methods.

As non-invasive continuous blood pressure monitoring (NCBPM) has gained wide attraction in the recent decades, many pulse arrival time (PAT) or pulse transit time (PTT) based blood pressure (BP) estimation studies have been conducted. However, most of the studies have used small homogeneous subject pools to generate models of BP based on particular interventions for induced hemodynamic change. In this study, a large open biosignal database from a diverse group of 2309 surgical patients was analyzed to assess the efficacy of PAT, PTT, and confounding factors on the estimation of BP. After pre-processing the dataset, a total of 6,777,308 data pairs of BP and temporal features between electrocardiogram (ECG) and photoplethysmogram (PPG) were extracted and analyzed. Correlation analysis revealed that PAT or PTT extracted from the intersecting-tangent (IT) point of PPG showed the highest mean correlation to BP. The mean correlation between PAT and systolic blood pressure (SBP) was -0.37 and the mean correlation between PAT and diastolic blood pressure (DBP) was -0.30, outperforming the correlation between BP and PTT at -0.12 for SBP and -0.11 for DBP. A linear model of BP with a simple calibration method using PAT as a predictor was developed which satisfied international standards for automatic oscillometric BP monitors in the case of DBP, however, SBP could not be predicted to a satisfactory level due to higher errors. Furthermore, multivariate regression analyses showed that many confounding factors considered in previous studies had inconsistent effects on the degree of correlation between PAT and BP.

Estimating Maximal Oxygen Uptake From Daily Activity Data Measured by a Watch-Type Fitness Tracker: Cross-Sectional Study.

BACKGROUND: Cardiorespiratory fitness (CRF), an important index of physical fitness, is the ability to inhale and provide oxygen to the exercising muscle. However, despite its importance, the current gold standard for measuring CRF is impractical, requiring maximal exercise from the participants. OBJECTIVE: This study aimed to develop a convenient and practical estimation model for CRF using data collected from daily life with a wristwatch-type device. METHODS: A total of 191 subjects, aged 20 to 65 years, participated in this study. Maximal oxygen uptake (VO2 max), a standard measure of CRF, was measured with a maximal exercise test. Heart rate (HR) and physical activity data were collected using a commercial wristwatch-type fitness tracker (Fitbit; Fitbit Charge; Fitbit) for 3 consecutive days. Maximal activity energy expenditure (aEEmax) and slope between HR and physical activity were calculated using a linear regression. A VO2 max estimation model was built using multiple linear regression with data on age, sex, height, percent body fat, aEEmax, and the slope. The result was validated with 2 different cross-validation methods. RESULTS: aEEmax showed a moderate correlation with VO2 max (r=0.50). The correlation coefficient for the multiple linear regression model was 0.81, and the SE of estimate (SEE) was 3.518 mL/kg/min. The regression model was cross-validated through the predicted residual error sum of square (PRESS). The PRESS correlation coefficient was 0.79, and the PRESS SEE was 3.667 mL/kg/min. The model was further validated by dividing it into different subgroups and calculating the constant error (CE) where a low CE showed that the model does not significantly overestimate or underestimate VO2 max. CONCLUSIONS: This study proposes a CRF estimation method using data collected by a wristwatch-type fitness tracker without any specific protocol for a wide range of the population.

Lumbar Stability in Healthy Individuals and Low Back Pain Patients Quantified by Wall Plank-and-Roll Test.

BACKGROUND: Low back pain (LBP) has been linked to the degree of lumbar stability, but evaluating lumbar stability has remained a challenge. Previous research has shown that inertial sensors could be used to quantify motor patterns during the wall plank-and-roll (WPR) test, and that LBP may cause deviations in movement from the general motor patterns observed in healthy individuals. OBJECTIVE: To generalize the lumbar motor patterns during the WPR test in healthy individuals, and to analyze the effect of aging and LBP on the motor patterns during the WPR test. DESIGN: A descriptive, exploratory research with a convenience sample. This study is registered at the Clinical Research Information Service (Korea) under public trial registration numbers KCT0002481 and KCT0002533. SETTING: A biomechanics laboratory of a tertiary university hospital. PARTICIPANTS: 57 healthy individuals (23 men 36.7 ± 15.4 years old and 34 women 42.4 ± 17.7 years old) and 17 patients (5 men 48.4 ± 10.9 years old and 12 women 33.7 ± 9.9 years old) with axial LBP. METHODS: Participants performed the WPR test with 2 inertial sensors placed on the thoracic spine and sacrum. Relative angles between the sensors were calculated to quantify and examine lumbar motion in 3 anatomical planes: axial twist, kyphosis-lordosis, and lateral bending. MAIN OUTCOME MEASURES: General motor patterns during the WPR test in healthy participants were examined, stratified based on age, and changes based on age were analyzed. Motor patterns of LBP patients were compared with those from the healthy group. RESULTS: Movement in the kyphosis-lordosis and lateral bending axes showed little variation in healthy participants, whereas in the axial twist axis there were 2 dominant patterns. A χ 2 test revealed that the distributions of 2 motor patterns in the axial twist axis between the younger group and the older group were significantly different (P < .05). Furthermore, the older group had decreased lordosis at the static position (P = .02) and at the maximal rotating position (P = .03). Compared with the healthy group, LBP patients showed increasing lateral bending at the maximal rotating position (P = .007) and increased lateral bending excursion angle (P = .04) during the WPR test. CONCLUSIONS: A general lumbar motor pattern was observed during the WPR test in the healthy participants, but age contributed to variations in this general pattern. Comparison of motor patterns between healthy individuals and LBP patients revealed a different type of variation in the LBP patients. The results presented should be scrutinized with further research, characterizing specific variations in different subgroups of LBP patients. LEVEL OF EVIDENCE: III.

Bidirectional Recurrent Auto-Encoder for Photoplethysmogram Denoising.

Photoplethysmography (PPG) has become ubiquitous with the development of smart watches and the mobile healthcare market. However, PPG is vulnerable to various types of noises that are ever present in uncontrolled environments, and the key to obtaining meaningful signals depends on successful denoising of PPG. In this context, algorithms have been developed to denoise PPG, but many were validated in controlled settings or are reliant on multiple steps that must all work correctly. This paper proposes a novel PPG denoising algorithm based on bidirectional recurrent denoising auto-encoder (BRDAE) that requires minimal pre-processing steps and have the benefit of waveform feature accentuation beyond simple denoising. The BRDAE was trained and validated on a dataset with artificially augmented noise, and was tested on a large open database of PPG signals collected from patients enrolled in intensive care units as well as from PPG data collected intermittently during the daily routine of nine subjects over 24 h. Denoising with the trained BRDAE improved signal-to-noise ratio of the noise-augmented data by 7.9 dB during validation. In the test datasets, the denoised PPG showed statistically significant improvement in heart rate detection as compared with the original PPG in terms of correlation to reference and root-mean-squared error. These results indicate that the proposed method is an effective solution for denoising the PPG signal, and promises values beyond traditional denoising by providing PPG feature accentuation for pulse waveform analysis.

Novel blood pressure and pulse pressure estimation based on pulse transit time and stroke volume approximation.

BACKGROUND: Non-invasive continuous blood pressure monitors are of great interest to the medical community due to their value in hypertension management. Recently, studies have shown the potential of pulse pressure as a therapeutic target for hypertension, but not enough attention has been given to non-invasive continuous monitoring of pulse pressure. Although accurate pulse pressure estimation can be of direct value to hypertension management and indirectly to the estimation of systolic blood pressure, as it is the sum of pulse pressure and diastolic blood pressure, only a few inadequate methods of pulse pressure estimation have been proposed. METHODS: We present a novel, non-invasive blood pressure and pulse pressure estimation method based on pulse transit time and pre-ejection period. Pre-ejection period and pulse transit time were measured non-invasively using electrocardiogram, seismocardiogram, and photoplethysmogram measured from the torso. The proposed method used the 2-element Windkessel model to model pulse pressure with the ratio of stroke volume, approximated by pre-ejection period, and arterial compliance, estimated by pulse transit time. Diastolic blood pressure was estimated using pulse transit time, and systolic blood pressure was estimated as the sum of the two estimates. The estimation method was verified in 11 subjects in two separate conditions with induced cardiovascular response and the results were compared against a reference measurement and values obtained from a previously proposed method. RESULTS: The proposed method yielded high agreement with the reference (pulse pressure correlation with reference R ≥ 0.927, diastolic blood pressure correlation with reference R ≥ 0.854, systolic blood pressure correlation with reference R ≥ 0.914) and high estimation accuracy in pulse pressure (mean root-mean-squared error ≤ 3.46 mmHg) and blood pressure (mean root-mean-squared error ≤ 6.31 mmHg for diastolic blood pressure and ≤ 8.41 mmHg for systolic blood pressure) over a wide range of hemodynamic changes. CONCLUSION: The proposed pulse pressure estimation method provides accurate estimates in situations with and without significant changes in stroke volume. The proposed method improves upon the currently available systolic blood pressure estimation methods by providing accurate pulse pressure estimates.

Development of a pressure sensor system for unobtrusive monitoring of abdominal muscle activities.

Surface electromyogram (sEMG) is often used by to objectively measure muscular activity during rehabilitation exercises. sEMG is accurate, but it is unsuitable for uses outside the clinic, and patients can benefit from an unobtrusive device which can be readily used to ubiquitously measure abdominal muscle activation. In this study, we present a pressure sensor system which can be latched onto a belt to measure abdominal muscle activation. sEMG and pressure sensor output were measured in 15 healthy young males during isometric trunk flexion exercise (public trials registration number, KCT0002351), and the results were highly correlated (median R > 0.939). As initial contact force can change the pressure sensor sensitivity, the experiment was performed at two different levels of belt tightness, but the correlations did not significantly improve after tightening the belt, suggesting that the system can be used to ubiquitously and unobtrusively monitor abdominal muscle activity with minimal discomfort.

Unobtrusive Estimation of Cardiorespiratory Fitness with Daily Activity in Healthy Young Men.

Despite the importance of cardiorespiratory fitness, no practical method exists to estimate maximal oxygen consumption (VO2max) without a specific exercise protocol. We developed an estimation model of VO2max, using maximal activity energy expenditure (aEEmax) as a new feature to represent the level of physical activity. Electrocardiogram (ECG) and acceleration data were recorded for 4 days in 24 healthy young men, and reference VO2max levels were measured using the maximal exercise test. aEE was calculated using the measured acceleration data and body weight, while heart rate (HR) was extracted from the ECG signal. aEEmax was obtained using linear regression, with aEE and HR as input parameters. The VO2max was estimated from the aEEmax using multiple linear regression modeling in the training group (n = 16) and was verified in the test group (n = 8). High correlations between the estimated VO2max and the measured VO2max were identified in both groups, with a 15-hour recording being sufficient to produce a highly accurate VO2max estimate. Additional recording time did not significantly improve the accuracy of the estimation. Our VO2max estimation method provides a robust alternative to traditional approaches while only requiring minimal data acquisition time in daily life.

Quantification of Lumbar Stability During Wall Plank-and-Roll Activity Using Inertial Sensors.

OBJECTIVE: To develop a simple method of quantifying dynamic lumbar stability by evaluating postural changes of the lumbar spine during a wall plank-and-roll (WPR) activity while maintaining maximal trunk rigidity. DESIGN: A descriptive, exploratory research with a convenience sample. SETTING: A biomechanics laboratory of a tertiary university hospital. PARTICIPANTS: Sixteen healthy young subjects (8 men and 8 women; 30.7 ± 6.8 years old) and 3 patients (2 men 46 and 50 years old; 1 woman 54 years old) with low back pain (LBP). METHODS: The subjects performed the WPR activity with 2 inertial sensors attached on the thoracic spine and sacrum. Relative angles between the sensors were calculated to characterize lumbar posture in 3 anatomical planes: axial twist (AT), kyphosis-lordosis (KL), or lateral bending (LB). Isokinetic truncal flexion and extension power were measured. MAIN OUTCOME MEASURES: AT, KL, and LB were compared between the initial plank and maximal roll positions. Angular excursions were compared between males and females and between rolling sides, and tested for correlation with isokinetic truncal muscle power. Patterns and consistencies of the lumbar postural changes were determined. Lumbar postural changes of each patient were examined in the aspects of pattern and excursion, considering those from the healthy subjects as reference. RESULTS: AT, KL, and LB were significantly changed from the initial plank to the maximal roll position (P < .01); that is, the thoracic spine rotated further, lumbar lordosis increased, and the thoracic spine was bent away from the wall by 6.9° ± 12.0°, 9.5° ± 6.5°, and 7.9° ± 4.9°, respectively. The patterns and amounts of lumbar postural changes were not significantly different between the rolling sides or between male and female participants, except that the excursion in AT was larger on the dominant rolling side. The excursions were not related to isokinetic truncal muscle power. The 3 LBP patients showed varied deviations in pattern and excursion from the average of the healthy subjects. CONCLUSIONS: Certain amounts and patterns of lumbar postural changes were observed in healthy young subjects, with no significant variations based on gender, rolling side, or truncal muscle power. Application of the evaluation on LBP patients revealed prominent deviations from the healthy postural changes, suggesting potential clinical applicability. Therefore, with appropriate development and case stratification, we believe that the quantification of lumbar postural changes during WPR activity can be used to assess dynamic lumbar stability in clinical practice.