Introduction Blood pressure (BP) has been a potential risk factor for cardiovascular diseases. BP measurement is one of the most useful parameters for early diagnosis, prevention, and treatment of cardiovascular diseases. At present, BP measurement mainly relies on cuff-based techniques that cause inconvenience and discomfort to users. Although some of the present prototype cuffless BP measurement techniques are able to reach overall acceptable accuracies, they require an electrocardiogram (ECG) and a photoplethysmograph (PPG) that make them unsuitable for true wearable applications. Therefore, developing a single PPG-based cuffless BP estimation algorithm with enough accuracy would be clinically and practically useful. Methods The University of Queensland vital sign dataset (online database) was accessed to extract raw PPG signals and its corresponding reference BPs (systolic BP and diastolic BP). The online database consisted of PPG waveforms of 32 cases from whom 8133 (good quality) signal segments (5 s for each) were extracted, preprocessed, and normalised in both width and amplitude. Three most significant pulse features (pulse area, pulse rising time, and width 25%) with their corresponding reference BPs were used to train and test three machine learning algorithms (regression tree, multiple linear regression (MLR), and support vector machine (SVM)). A 10-fold cross-validation was applied to obtain overall BP estimation accuracy, separately for the three machine learning algorithms. Their estimation accuracies were further analysed separately for three clinical BP categories (normotensive, hypertensive, and hypotensive). Finally, they were compared with the ISO standard for noninvasive BP device validation (average difference no greater than 5 mmHg and SD no greater than 8 mmHg). Results In terms of overall estimation accuracy, the regression tree achieved the best overall accuracy for SBP (mean and SD of difference: −0.1 ± 6.5 mmHg) and DBP (mean and SD of difference: −0.6 ± 5.2 mmHg). MLR and SVM achieved the overall mean difference less than 5 mmHg for both SBP and DBP, but their SD of difference was >8 mmHg. Regarding the estimation accuracy in each BP categories, only the regression tree achieved acceptable ISO standard for SBP (−1.1 ± 5.7 mmHg) and DBP (−0.03 ± 5.6 mmHg) in the normotensive category. MLR and SVM did not achieve acceptable accuracies in any BP categories. Conclusion This study developed and compared three machine learning algorithms to estimate BPs using PPG only and revealed that the regression tree algorithm was the best approach with overall acceptable accuracy to ISO standard for BP device validation. Furthermore, this study demonstrated that the regression tree algorithm achieved acceptable measurement accuracy only in the normotensive category, suggesting that future algorithm development for BP estimation should be more specific for different BP categories.

Objective:The waveform of the photoplethysmography (PPG) signal depends on the measurement site and individual physiological conditions. Filtering can distort the morphology of the original PPG signal waveform and change the timing of pulse feature points on PPG signals. We aim to quantitatively investigate the effect of PPG signal morphology (related to measurement site) and type of pulse feature on the filtering-induced time shift (TS). Approach: 60-second PPG signals were measured from six body sites [finger, wrist under (volar), wrist upper (dorsal), earlobe, and forehead] of 36 healthy adults. Using infinite impulse response digital filters which are common in PPG signal processing, PPG signals were prefiltered (band-pass, pass and stop bands: >0.5Hz and <0.2Hz for high-pass filter, <20Hz and >30Hz for low-pass filter) then filtered (low-pass, pass and stop bands: <3Hz and >5Hz). Four pulse feature points were defined and extracted (peak, valley, maximal first derivative, and maximal second derivative). For each subject, overall TS and intra-subject TS variability in feature points were calculated as the mean and standard deviation of TS between prefiltered and filtered PPG signals in 50 cardiac cycles. Statistical testing was performed to investigate the effect of measurement site and type of pulse feature on overall TS and intra-subject TS variability. Results: Measurement site, type of pulse feature, and their interaction had significant impacts on the overall TS and intra-subject TS variability (p<0.001 for all). Valley and maximal second derivative showed higher overall TS than peak and maximal first derivative. Finger has higher overall TS and lower intra-subject TS variability than other measurement sites. Significance: Measurement site and type of pulse feature can significantly influence the timing of feature point on filtered PPG signals. Filtering parameters should be quoted to support the reproducibility of PPG-related studies.

Objective: Based on different physiological mechanisms, the respiratory modulations of photoplethysmography (PPG) signals differ in strength and resultant accuracy of respiratory frequency (RF) estimations. We aimed to investigate the strength of different respiratory modulations and the accuracy of resultant RF estimations in different body sites and two breathing patterns. Approach: PPG and reference respiratory signals were simultaneously measured over 60 s from 36 healthy subjects in six sites (arm, earlobe, finger, forehead, wrist-under (volar side), wrist-upper (dorsal side)). Respiratory signals were extracted from PPG recordings using four demodulation approaches: amplitude modulation (AM), baseline wandering (BW), frequency modulation (FM) and filtering. RFs were calculated from the PPG-derived and reference respiratory signals. To investigate the strength of respiratory modulations, the energy proportion in the range that covers 75% of the total energy in the reference respiratory signal, with RF in the middle, was calculated and compared between different modulations. Analysis of variance and the Scheirer–Ray–Hare test were performed with post hoc analysis. Main results: In normal breathing, FM was the only modulation whose RF was not significantly different from the reference RF (p > 0.05). Compared with other modulations, FM was significantly higher in energy proportion (p < 0.05) and lower in RF estimation error (p < 0.05). As to energy proportion, measurements from the finger and the forehead were not significantly different (p > 0.05), but both were significantly different from the other four sites (p < 0.05). In deep breathing, the RFs derived by BW, filtering and FM were not significantly different from the reference RF (p > 0.05). The RF estimation error of FM was significantly less than that of AM or BW (p < 0.05). The energy proportion of FM was significantly higher than that of other modulations (p < 0.05). Significance: Of all the respiratory modulations, FM has the highest strength and is appropriate for accurate RF estimation from PPG signals recorded at different sites and for different breathing patterns.

Traditional cuff-based blood pressure (BP) monitoring procedure causes inconvenience and discomfort to the users. To overcome these limitations, cuffless BP estimation based on pulse transit time (PTT) and single-channel photoplethysmography (PPG) has been proposed. However, existing studies based on PTT and PPG for BP estimation did not achieve AAMI/ISO standard criteria for BP measurement (mean difference within ±5mmHg and SD of difference within ±8mmHg) under each BP category (Hypotensive, Normotensive and Hypertensive). This study aims to validate an innovative two-step method for PPGbased cuffless BP estimation. A combined database was derived from two online databases (Queensland and MIMIC II) to cover a wide range of corresponding BPs. In total, there were 18010 raw PPG signal segments (5 seconds for each) with corresponding BPs, separated into two halves for training and testing of algorithms (independent datasets). Each PPG signal segment was pre-processed to extract 16 signal features. Later, three significant features have been selected using multicollinearity test. The traditional generic (trained with uncategorized BP) algorithm and two-step algorithm (specifically optimized for each BP category) were developed using machine learning. Generally, the two-step algorithm achieved the AAMI/ISO standard in estimating systolic BP (mean ±SD: 0.07±7.1 mmHg, p<0.001) and diastolic BP (−0.08±6.0 mmHg, p<0.001). Categorically, the two-step method also achieved standard accuracy in all BP categories except Hypotensive systolic BP whereas generic algorithm did not conform to standard accuracy in any BP category except Hypotensive diastolic BP and Normotensive categories. Compared to the traditional generic algorithm, the two-step algorithm specifically designed for three different BP category patients and achieved standard accuracy for cuffless BP estimation. INDEX TERMS Cuffless BP, pulse transit time, Photoplethysmography & categorical BP estimation.