Objective: The objective of this study is to obtainaccurate estimations of breathing rate (BR), using only the electrocardiogram (ECG) or the photoplethysmogram (PPG) signals, to avoid wearing cumbersome and uncomfortable sensors for direct measurements. Methods: Several respiration waveforms are derived from ECG or PPG signals based on amplitude, frequency and baseline wander modulations. It is however difficult to determine their optimal combination for BR estimation due to the noise and patient specificity. We first propose to quantify the quality of modulation waveforms using respiratory quality indices (RQIs). We then present two methods: the first automatically selects the modulation signal with highest RQI for BR estimation, the second tracks the respiration signal using the Kalman Smoother (KS) to fuse modulation signals with highest RQI. Results: These two methods are evaluated on two independent datasets, one benchmark DB with immobilized patients recordings and the second with those performing daily activities. Our results outperform existing methods in the literature in both cases. Conclusion: Experimental results show that the RQIs coupled with a fusion algorithm increases the accuracy for BR estimations in dealing with derived modulation signals. Significance: This work describes a robust Kalman Smoother method applicable in multiple clinical contexts to improve breathing rate estimation from data fusion.

Robust, real-time event detection from physiological signals acquired during long-term ambulatory monitoring still represents a major challenge for highly-artifacted signals. In this paper, we propose an original and generic multi-feature probabilistic detector (MFPD) and apply it to real-time QRS complex detection under noisy conditions. The MFPD method calculates a binary Bayesian probability for each derived feature and makes a centralized fusion, using the Kullback-Leibler divergence. The method is evaluated on two ECG databases: 1) the MIT-BIH arrhythmia database from Physionet containing clean ECG signals, 2) a benchmark noisy database created by adding noise recordings of the MIT-BIH noise stress test database, also from Physionet, to the MIT-BIH arrhythmia database. Results are compared with a well-known wavelet-based detector, and two recently published detectors: one based on spatiotemporal characteristic of the QRS complex and the second, as the MFDP, based on feature calculations from the University of New South Wales detector (UNSW). For both benchmark Physionet databases, the proposed MFPD method achieves the lowest standard deviation in sensitivity and positive predictivity (+P) despite its online algorithm architecture. While the statistics are comparable for low-to mildly artifactual ECG signals, the MFPD outperforms reference methods for artifacted ECG with low SNR levels reaching 87.48 ± 14.21% in sensitivity and 89.39 ± 14.67% in +P as compared to 88.30 ± 17.66% and 86.06 ± 19.67% respectively from UNSW, the best performing reference method. With demonstrations on the extensively studied QRS detection problem, we consider that the proposed generic structure of the multi-feature probabilistic detector should offer promising perspectives for long-term monitoring applications for highly-artifacted signals.

The precise measurement of respiratory variables, such as tidal volume, minute ventilation, and respiratory rate, is necessary to monitor respiratory status, overcome several diseases, improve patient health conditions and reduce health care costs. This measurement has conventionally been performed by breathing into a mouthpiece connected to a flow rate measuring device. However, a mouthpiece can be uncomfortable for the subject and is difficult to use for long-term monitoring. Other noninvasive systems and devices have been developed that do not require a mouthpiece to quantitatively measure respiratory variables. These techniques are based on measuring size changes of the rib cage (RC) and abdomen (ABD), as lung volume is known to be a function of these variables. Among these systems, we distinguish respiratory inductive plethysmography (RIP), respiratory magnetometer plethysmography (RMP), and optoelectronic plethysmography devices. However, these devices should be previously calibrated for the correct evaluation of respiratory variables. The most popular calibration methods are isovolume manoeuvre calibration (ISOCAL), qualitative diagnostic calibration (QDC), multiple linear regression (MLR) and artificial neural networks (ANNs). The aim of this review is first to present how thoracoabdominal breathing distances can be used to estimate respiratory variables and second to present the different techniques and calibration methods used for this purpose.

Intramuscular electromyography (EMG) signals are usually decomposed with semiautomatic procedures that involve the interaction with an expert operator. In this paper, a Bayesian statistical model and a maximum a posteriori (MAP) estimator are used to solve the problem of multiunit EMG decomposition in a fully automatic way. The MAP estimation exploits both the likelihood of the reconstructed EMG signal and some physiological constraints, such as the discharge pattern regularity and the refractory period of muscle fibers, as prior information integrated in a Bayesian framework. A Tabu search is proposed to efficiently tackle the nondeterministic polynomial-time-hard problem of optimization w.r.t the motor unit discharge patterns. The method is fully automatic and was tested on simulated and experimental EMG signals. Compared with the semiautomatic decomposition performed by an expert operator, the proposed method resulted in an accuracy of 90.0% ± 3.8% when decomposing single-channel intramuscular EMG signals recorded from the abductor digiti minimi muscle at contraction forces of 5% and 10% of the maximal force. The method can also be applied to the automatic identification and classification of spikes from other neural recordings.

The decomposition of multiunit signals consists of the restoration of spike trains and action potentials in neural or muscular recordings. Because of the complexity of automatic decomposition, semiautomatic procedures are sometimes chosen. The main difficulty in automatic decomposition is the resolution of temporally overlapped potentials. In a previous study , we proposed a Bayesian model coupled with a maximum a posteriori (MAP) estimator for fully automatic decomposition of multiunit recordings and we showed applications to intramuscular EMG signals. In this study, we propose a more complex signal model that includes the variability in amplitude of each unit potential. Moreover, we propose the Markov Chain Monte Carlo (MCMC) simulation and a Bayesian minimum mean square error (MMSE) estimator by averaging on samples that converge in distribution to the joint posterior law. We prove the convergence property of this approach mathematically and we test the method representatively on intramuscular multiunit recordings. The results showed that its average accuracy in spike identification is greater than 90% for intramuscular signals with up to 8 concurrently active units. In addition to intramuscular signals, the method can be applied for spike sorting of other types of multiunit recordings.

As a task-based approach for medical image quality assessment, model observers (MOs) have been proposed as surrogates for human observers. While most MOs treat only signal-known-exactly tasks, there are few studies on signal-known-statistically (SKS) MOs, which are clinically more relevant. In this paper, we present a new SKS MO named channelized joint detection and estimation observer (CJO), capable of detecting and estimating signals with unknown amplitude, orientation, and size. We evaluate its estimation and detection performance using both synthesized (correlated Gaussian) backgrounds and real clinical (magnetic resonance) backgrounds. The results suggest that the CJO has good performance in the SKS detection–estimation task.