Publications

Respiration rate can be assessed by analyzing respiratory changes of the electrocardiogram (ECG). Several methods can be applied to derive the respiratory signal from the ECG (EDR signal). In this study, four EDR estimation methods based on QRS features were analyzed. A database with 44 healthy subjects (16 females) in supine and sitting positions was analyzed. Respiratory flow and ECG recordings on leads I, II, III and a Chest lead was studied. A QR slope-based method, an RS slope-based method, an QRS angle-based method and an QRS area-based method were applied. Their performance was evaluated by the correlation coefficient with the reference respiratory volume signal. Significantly higher correlation coefficients in the range r = 0.77 – 0.86 were obtained with the Chest lead for all methods. The EDR estimation method based on the QRS angle provided the highest similarity with the volume signal for all recording leads and subject positions. We found no statistically significant differences according to gender or subject position.Clinical Relevance— This work analyzes the EDR signal from four electrocardiographic leads to obtain the respiratory signal and contributes to a simplified analysis of respiratory activity.

JTD Keywords: Electrocardiography, Lead, Estimation, Correlation coefficient, Databases, Heart, Correlation

Neural respiratory drive as measured by the electromyography allows the study of the imbalance between the load on respiratory muscles and its capacity. Surface respiratory electromyography (sEMG) is a non-invasive tool used for indirectly assessment of NRD. It also provides a way to evaluate the level and pattern of respiratory muscle activation. The prevalence of electrocardiographic activity (ECG) in respiratory sEMG signals hinders its proper evaluation. Moreover, the occurrence of abnormal heartbeats or cardiac arrhythmias in respiratory sEMG measures can make even more challenging the NRD estimation. Respiratory sEMG can be evaluated using the fixed sample entropy (fSampEn), a technique which is less affected by cardiac artefacts. The aim of this work was to investigate the performance of the fSampEn, the root mean square (RMS) and the average rectified value (ARV) on respiratory sEMG signals with supraventricular arrhythmias (SVA) for NRD estimation. fSampEn, ARV and RMS parameters increased as the inspiratory load increased during the test. fSampEn was less influenced by ECG with SVAs for the NRD estimation showing a greater response to respiratory sEMG, reflected with a higher percentage increase with increasing load (228 % total increase, compared to 142 % and 135 % for ARV and RMS, respectively).

JTD Keywords: Electrocardiography, Muscles, Electrodes, Estimation, Band-pass filters, Electromyography, Heart beat

Lung sound (LS) signals are often contaminated by impulsive artifacts that complicate the estimation of lung sound intensity (LSI) using conventional amplitude estimators. Fixed sample entropy (fSampEn) has proven to be robust to cardiac artifacts in myographic respiratory signals. Similarly, fSampEn is expected to be robust to artifacts in LS signals, thus providing accurate LSI estimates. However, the choice of fSampEn parameters depends on the application and fSampEn has not previously been applied to LS signals. This study aimed to perform an evaluation of the performance of the most relevant fSampEn parameters on LS signals, and to propose optimal fSampEn parameters for LSI estimation. Different combinations of fSampEn parameters were analyzed in LS signals recorded in a heterogeneous population of healthy subjects and chronic obstructive pulmonary disease patients during loaded breathing. The performance of fSampEn was assessed by means of its cross-covariance with flow signals, and optimal fSampEn parameters for LSI estimation were proposed.

JTD Keywords: Large scale integration, Lung, Estimation, Entropy, Loading, Robustness, Diseases

Providing force feedback as relevant information in current Robot-Assisted Minimally Invasive Surgery systems constitutes a technological challenge due to the constraints imposed by the surgical environment. In this context, force estimation techniques represent a potential solution, enabling to sense the interaction forces between the surgical instruments and soft-tissues. Specifically, if visual feedback is available for observing soft-tissues’ deformation, this feedback can be used to estimate the forces applied to these tissues. To this end, a force estimation model, based on Convolutional Neural Networks and Long-Short Term Memory networks, is proposed in this work. This model is designed to process both, the spatiotemporal information present in video sequences and the temporal structure of tool data (the surgical tool-tip trajectory and its grasping status). A series of analyses are carried out to reveal the advantages of the proposal and the challenges that remain for real applications. This research work focuses on two surgical task scenarios, referred to as pushing and pulling tissue. For these two scenarios, different input data modalities and their effect on the force estimation quality are investigated. These input data modalities are tool data, video sequences and a combination of both. The results suggest that the force estimation quality is better when both, the tool data and video sequences, are processed by the neural network model. Moreover, this study reveals the need for a loss function, designed to promote the modeling of smooth and sharp details found in force signals. Finally, the results show that the modeling of forces due to pulling tasks is more challenging than for the simplest pushing actions.

JTD Keywords: Convolutional neural networks, Force estimation, LSTM networks, Robotic surgery

This study evaluates the onset and offset of neural inspiratory time estimated from surface diaphragm electromyographic (EMGdi) recordings. EMGdi and airflow signals were recorded in ten healthy subjects according to two respiratory protocols based on respiratory rate (RR) increments, from 15 to 40 breaths per minute (bpm), and fractional inspiratory time (Ti/Ttot) decrements, from 0.54 to 0.18. The analysis of diaphragm electromyographic (EMGdi) signal amplitude is an alternative approach for the quantification of neural respiratory drive (NRD). The EMGdi amplitude was estimated using the fixed sample entropy computed over a 250 ms moving window of the EMGdi signal (EMGdifse). The neural onset was detected through a dynamic threshold over the EMGdifse using the kernel density estimation method, while neural offset was detected by finding when the EMGdifse had decreased to 70 % of the peak value reached during inspiration. The Bland-Altman analysis between airflow and neural onsets showed a global bias of 46 ms in the RR protocol and 22 ms in the Ti/Ttot protocol. The Bland-Altman analysis between airflow and neural offsets reveals a global bias of 11 ms in the RR protocol and -2 ms in the Ti/Ttot protocol. The relationship between pairs of RR values (Pearson’s correlation coefficient of 0.99, Bland- Altman limits of -2.39 to 2.41 bpm, and mean bias of 0.01 bpm) and between pairs of Ti/Ttot values (Pearson’s correlation coefficient of 0.86, Bland-Altman limits of -0.11 to 0.10, and mean bias of -0.01) showed a good agreement. In conclusion, we propose a method for determining neural onset and neural offset based on non-invasive recordings of the electrical activity of the diaphragm that requires no filtering of cardiac muscle interference.

JTD Keywords: Kernel density estimation (KDE),, Surface diaphragm electromyographic,, (EMGdi) signal,, Inspiratory time,, Neural respiratory drive (NRD),, Neural inspiratory time,, Fixed sample entropy (fSampEn)

The time delay estimation problem associated with an ensemble of misaligned, repetitive signals is revisited. Each observed signal is assumed to be composed of an unknown, deterministic signal corrupted by Gaussian, white noise. This paper shows that maximum likelihood (ML) time delay estimation can be viewed as the maximization of an eigenvalue ratio, where the eigenvalues are obtained from the ensemble correlation matrix. A suboptimal, one-step time delay estimate is proposed for initialization of the ML estimator, based on one of the eigenvectors of the inter-signal correlation matrix. With this approach, the ML estimates can be determined without the need for an intermediate estimate of the underlying, unknown signal. Based on respiratory flow signals, simulations show that the variance of the time delay estimation error for the eigenvalue-based method is almost the same as that of the ML estimator. Initializing the maximization with the one-step estimates, rather than using the ML estimator alone, the computation time is reduced by a factor of 5M when using brute force maximization (M denoting the number of signals in the ensemble), and a factor of about 1.5 when using particle swarm maximization. It is concluded that eigenanalysis of the ensemble correlation matrix not only provides valuable insight on how signal energy, jitter, and noise influence the estimation process, but it also leads to a one-step estimator which can make the way for a substantial reduction in computation time.

JTD Keywords: Biomedical signals, Time delay estimation, Eigenanalysis, Ensemble analysis

Cardiac motion estimation is an important diagnostic tool for detecting heart diseases and it has been explored with modalities such as MRI and conventional ultrasound (US) sequences. US cardiac motion estimation still presents challenges because of complex motion patterns and the presence of noise. In this work, we propose a novel approach to estimate cardiac motion using ultrafast ultrasound data. Our solution is based on a variational formulation characterized by the L 2-regularized class. Displacement is represented by a lattice of b-splines and we ensure robustness, in the sense of eliminating outliers, by applying a maximum likelihood type estimator. While this is an important part of our solution, the main object of this work is to combine low-rank data representation with topology preservation. Low-rank data representation (achieved by finding the k-dominant singular values of a Casorati matrix arranged from the data sequence) speeds up the global solution and achieves noise reduction. On the other hand, topology preservation (achieved by monitoring the Jacobian determinant) allows one to radically rule out distortions while carefully controlling the size of allowed expansions and contractions. Our variational approach is carried out on a realistic dataset as well as on a simulated one. We demonstrate how our proposed variational solution deals with complex deformations through careful numerical experiments. The low-rank constraint speeds up the convergence of the optimization problem while topology preservation ensures a more accurate displacement. Beyond cardiac motion estimation, our approach is promising for the analysis of other organs that exhibit motion.

JTD Keywords: Cardiac analysis, Low-rank representation, Motion estimation, Topology preservation, Ultrafast ultrasound

Robotic-assisted minimally invasive surgeries have gained a lot of popularity over conventional procedures as they offer many benefits to both surgeons and patients. Nonetheless, they still suffer from some limitations that affect their outcome. One of them is the lack of force feedback which restricts the surgeon's sense of touch and might reduce precision during a procedure. To overcome this limitation, we propose a novel force estimation approach that combines a vision based solution with supervised learning to estimate the applied force and provide the surgeon with a suitable representation of it. The proposed solution starts with extracting the geometry of motion of the heart's surface by minimizing an energy functional to recover its 3D deformable structure. A deep network, based on a LSTM-RNN architecture, is then used to learn the relationship between the extracted visual-geometric information and the applied force, and to find accurate mapping between the two. Our proposed force estimation solution avoids the drawbacks usually associated with force sensing devices, such as biocompatibility and integration issues. We evaluate our approach on phantom and realistic tissues in which we report an average root-mean square error of 0.02 N.

JTD Keywords: Computer-assisted surgery, Deep networks, Force estimation, Visual deformation

Measuring diaphragmatic electromyography (EMGdi) provides an indirect quantification of neural respiratory drive and allows the delimitation of diaphragm neural activation and deactivation during inspiration. EMGdi recordings have been incorporated in novel modes of assisted mechanical ventilation, such as neurally adjusted ventilatory assist (NAVA), to trigger and cycle-off the ventilator. The EMGdi signal improves the assistance delivered by more conventional ventilatory modes, in which the ventilator is synchronized with the patient employing a pneumatic triggering. In this work, we evaluate the time delay between the onset and offset of inspiratory activity estimated from EMGdi and three respiratory mechanical signals: the respiratory flow (FL), the transdiaphragmatic pressure (Pdi) and the diaphragm length (Ldi) signals. To this purpose, these signals were acquired in three mongrel dogs surgically instrumented under general anesthesia. Onsets and offsets were estimated manually and by automatic algorithms on these signals. The highest delays were obtained between EMGdi and FL (100 ms) while the lowest delays were obtained between EMGdi and Pdi (8 ms). Moreover, differences between manual and automatic estimations showed a mean absolute error lower than 45 ms. In conclusion, our study points out that both EMGdi and Pdi signals detect the onset and offset of inspiratory activity earlier than the FL signal, and would therefore be better for the improvement of patient-ventilator synchrony.

JTD Keywords: Estimation, Ventilation, Anesthesia, Dogs, Manuals, Power harmonic filters

Changes in the left ventricle function produce alternans in the hemodynamic and electric behavior of the cardiovascular system. A total of 49 cardiomyopathy patients have been studied based on the blood pressure signal (BP), and were classified according to the left ventricular ejection fraction (LVEF) in low risk (LR: LVEF>35%, 17 patients) and high risk (HR: LVEF≤35, 32 patients) groups. We propose to characterize these patients using a linear and a nonlinear methods, based on the spectral estimation and the recurrence plot, respectively. From BP signal, we extracted each systolic time interval (STI), upward systolic slope (BPsl), and the difference between systolic and diastolic BP, defined as pulse pressure (PP). After, the best subset of parameters were obtained through the sequential feature selection (SFS) method. According to the results, the best classification was obtained using a combination of linear and nonlinear features from STI and PP parameters. For STI, the best combination was obtained considering the frequency peak and the diagonal structures of RP, with an area under the curve (AUC) of 79%. The same results were obtained when comparing PP values. Consequently, the use of combined linear and nonlinear parameters could improve the risk stratification of cardiomyopathy patients.

JTD Keywords: Feature extraction, Blood pressure, Heart rate, Estimation, Data mining, Covariance matrices, Hospitals

Fuzzy theory was motivated by the need to create human-like solutions that allow representing vagueness and uncertainty that exist in the real-world. These capabilities have been recently further enhanced by deep learning since it allows converting complex relation between data into knowledge. In this paper, we present a novel Deep-Neuro-Fuzzy strategy for unsupervised estimation of the interaction forces in Robotic Assisted Minimally Invasive scenarios. In our approach, the capability of Neuro-Fuzzy systems for handling visual uncertainty, as well as the inherent imprecision of real physical problems, is reinforced by the advantages provided by Deep Learning methods. Experiments conducted in a realistic setting have demonstrated the superior performance of the proposed approach over existing alternatives. More precisely, our method increased the accuracy of the force estimation and compared favorably to existing state of the art approaches, offering a percentage of improvement that ranges from about 35% to 85%.

JTD Keywords: Estimation, Force, Machine learning, Robots, Three-dimensional displays, Uncertainty, Visualization

Obstructive Sleep Apnea (OSA) is a sleep disorder highly prevalent in the general population. Cardiorespiratory Phase Synchronization (CRPS) is a form of non-linear interaction between respiratory and cardiovascular systems that was found to be reduced in severe OSA patients. The Hilbert Transform (HT) method was the recommended choice for estimating the respiratory phase in CRPS studies. But we have noticed that HT provides a phase that is aligned to the transition between the exhalation and the inhalation parts of different breathing cycles, instead of being aligned to the breathing onsets. In this work we proposed a Realigned HT phase estimation method (RHT) and we compared it to the conventional HT and to the Linear Phase (LP) approximation for estimating CRPS in a database of 28 patients with different OSA severity levels. RHT provided similar synchronization percentages (%Sync) as HT, and it enhanced the significant differences in %Sync between mild and severe OSA patients. %Sync showed the highest negative correlation with the Apnea-Hypopnea Index (AHI) when using RHT (rAHI=-0.692, p<;0.001), which only had an 10% extra computational cost. On the other hand, LP method significantly overestimated %Sync especially in the more severe patients, because it was unable to track the phase non-linearities that can be observed during sleep disordered breathing. Therefore, the newly proposed RHT can be the preferred alternative over the conventional HT or the LP approximation for estimating CRPS in OSA patients.

JTD Keywords: Correlation, Databases, Electrocardiography, Phase estimation, Sleep apnea, Synchronization, Transforms

The lack of force feedback is considered one of the major limitations in Robot Assisted Minimally Invasive Surgeries. Since add-on sensors are not a practical solution for clinical environments, in this paper we present a force estimation approach that starts with the reconstruction of a 3D deformation structure of the tissue surface by minimizing an energy functional. A Recurrent Neural Network-Long Short Term Memory (RNN-LSTM) based architecture is then presented to accurately estimate the applied forces. According to the results, our solution offers long-term stability and shows a significant percentage of accuracy improvement, ranging from about 54% to 78%, over existing approaches.

JTD Keywords: Computer architecture, Estimation, Force, Microprocessors, Robot sensing systems, Surgery

This paper addresses the issue of lack of force feedback in robotic-assisted minimally invasive surgeries. Force is an important measure for surgeons in order to prevent intra-operative complications and tissue damage. Thus, an innovative neuro-vision based force estimation approach is proposed. Tissue surface displacement is first measured via minimization of an energy functional. A neuro approach is then used to establish a geometric-visual relation and estimate the applied force. The proposed approach eliminates the need of add-on sensors, carrying out biocompatibility studies and is applicable to tissues of any shape. Moreover, we provided an improvement from 15.14% to 56.16% over other approaches which demonstrate the potential of our proposal.

JTD Keywords: Estimation, Force, Minimally invasive surgery, Robot sensing systems, Three-dimensional displays

Automated signal processing in analytical instrumentation is today required for the analysis of highly complex biomedical samples. Baseline estimation techniques are often used to correct long term instrument contamination or degradation. They are essential for accurate peak area integration. Some methods approach the baseline estimation iteratively, trying to ignore peaks which do not belong to the baseline. The proposed method in this work consists of a modification of the Asymmetric Least Squares (ALS) baseline removal technique developed by Eilers and Boelens. The ALS technique suffers from bias in the presence of intense peaks (in relation to the noise level). This is typical of diverse instrumental techniques such as Gas Chromatography-Mass Spectrometry (GC-MS) or Gas Chromatography-Ion Mobility Spectrometry (GC-IMS). In this work, we propose a modification (named psalsa) to the asymmetry weights of the original ALS method in order to better reject large peaks above the baseline. Our method will be compared to several versions of the ALS algorithm using synthetic and real GC signals. Results show that our proposal improves previous versions being more robust to parameter variations and providing more accurate peak areas.

JTD Keywords: Gas chromatography, Instruments, Radioactivity measurement, Signal processing, Analytical instrument, Analytical Instrumentation, Asymmetric least squares, Baseline estimation, Baseline removal, Gas chromatography-mass spectrometries (GC-MS), Instrumental techniques, Noise levels, Iterative methods

Due to the increasing elderly population and the extensive number of comorbidities that affect them, studies are required to determine future increments in admission to emergency departments. Some of these studies could focus on the relation between chronic diseases and breathing pattern in elderly patients. Variations in the fractal properties of respiratory signals can be associated with several diseases. To determine the relationship between these variations and breathing patterns, and to quantify the fractal properties of respiratory flow signals, we estimated the Hurst exponent (H). Detrended fluctuation analysis (DFA) and discrete wavelet transform-based estimation (DWTE) methods were applied. The estimation methods were analyzed using simulated data series generated by fractional Gaussian noise. 43 elderly patients (19 patients with a non-periodic breathing pattern - nPB, and 24 patients with a periodic breathing pattern - PB) were studied. The results were evaluated according to the length of data and the number of averaged data series used to obtain a good estimation. The DWTE method estimated the respiratory flow signals better than the DFA method, and obtained Hurst values clustered by group. We found significant differences in the H exponent (p = 0.002) between PB and nPB patients, which showed different behavior in the fractal properties.

JTD Keywords: Discrete wavelet transforms, Diseases, Estimation, Fractals, Modulation, Senior citizens, Time series analysis

The aim of the present study was to measure bilateral asynchrony in patients suffering from Chronic Obstructive Pulmonary Disease (COPD) performing an incremental inspiratory load protocol. Bilateral asynchrony was estimated by the comparison of respiratory movements derived from diaphragm mechanomyographic (MMGdi) signals, acquired by means of capacitive accelerometers placed on left and right sides of the rib cage. Three methods were considered for asynchrony evaluation: Lissajous figure, Hilbert transform and Motto's algorithm. Bilateral asynchrony showed an increase at 20, 40 and 60% (values of normalized inspiratory pressure by their maximum value reached in the last inspiratory load) while the very severe group showed and increase at 20, 40, 80, and 100 % during the protocol. These increments in the phase's shift can be due to an increase of the inspiratory load along the protocol, and also as a consequence of distress and fatigue. In summary, this work evidenced the capability to estimate bilateral asynchrony in COPD patients. These preliminary results also showed that the use of capacitive accelerometers can be a suitable sensor for recording of respiratory movement and evaluation of asynchrony in COPD patients.

JTD Keywords: Accelerometers, Diseases, Estimation, Fatigue, IP networks, Protocols, Transforms

Robotic-assisted minimally invasive surgery has demonstrated its benefits in comparison with traditional procedures. However, one of the major drawbacks of current robotic system approaches is the lack of force feedback. Apart from space restrictions, the main problems of using force sensors are their high cost and the biocompatibility. In this work a proposal based on Vision Based Force Measurement is presented, in which the deformation mapping of the tissue is obtained using the `2−Regularized Optimization class, and the force is estimated via a recurrent neural network that has as inputs the kinematic variables and the deformation mapping. Moreover, the capability of RNN for predicting time series is used in order to deal with tool occlusions. The highlights of this proposal, according to the results, are: knowledge of material properties are not necessary, there is no need of adding extra sensors and a good trade-off between accuracy and efficiency has been achieved.

JTD Keywords: Force estimation, Regularized optimization, Deformable tracking, Recurrent neural network

Non-invasive evaluation of respiratory activity is an area of increasing research interest, resulting in the appearance of new monitoring techniques, ones of these being based on the analysis of the diaphragm mechanomyographic (MMGdi) signal. The MMGdi signal can be decomposed into two parts: (1) a high frequency activity corresponding to lateral vibration of respiratory muscles, and (2) a low frequency activity related to excursion of the thoracic cage. The purpose of this study was to apply the empirical mode decomposition (EMD) method to obtain the low frequency of MMGdi signal and selecting the intrinsic mode functions related to the respiratory movement. With this intention, MMGdi signals were acquired from a healthy subject, during an incremental load respiratory test, by means of two capacitive accelerometers located at left and right sides of rib cage. Subsequently, both signals were combined to obtain a new signal which contains the contribution of both sides of thoracic cage. Respiratory rate (RR) measured from the mechanical activity (RR_Mmg) was compared with that measured from inspiratory pressure signal (RR_P). Results showed a Pearson's correlation coefficient (r = 0.87) and a good agreement (mean bias = -0.21 with lower and upper limits of -2.33 and 1.89 breaths per minute, respectively) between RR_mmg and RR_P measurements. In conclusion, this study suggests that RR can be estimated using EMD for extracting respiratory movement from low mechanical activity, during an inspiratory test protocol.

JTD Keywords: Accelerometers, Band-pass filters, Biomedical measurement, Empirical mode decomposition, Estimation, IP networks, Muscles

One of the most challenging problems in intensive care is still the process of discontinuing mechanical ventilation, called weaning process. Both an unnecessary delay in the discontinuation process and a weaning trial that is undertaken too early are undesirable. In this study, we analyzed respiratory pattern variability using the respiratory volume signal of patients submitted to two different levels of pressure support ventilation (PSV), prior to withdrawal of the mechanical ventilation. In order to characterize the respiratory pattern, we analyzed the following time series: inspiratory time, expiratory time, breath duration, tidal volume, fractional inspiratory time, mean inspiratory flow and rapid shallow breathing. Several autoregressive modeling techniques were considered: autoregressive models (AR), autoregressive moving average models (ARMA), and autoregressive models with exogenous input (ARX). The following classification methods were used: logistic regression (LR), linear discriminant analysis (LDA) and support vector machines (SVM). 20 patients on weaning trials from mechanical ventilation were analyzed. The patients, submitted to two different levels of PSV, were classified as low PSV and high PSV. The variability of the respiratory patterns of these patients were analyzed. The most relevant parameters were extracted using the classifiers methods. The best results were obtained with the interquartile range and the final prediction errors of AR, ARMA and ARX models. An accuracy of 95% (93% sensitivity and 90% specificity) was obtained when the interquartile range of the expiratory time and the breath duration time series were used a LDA model. All classifiers showed a good compromise between sensitivity and specificity.

JTD Keywords: autoregressive moving average processes, feature extraction, medical signal processing, patient care, pneumodynamics, signal classification, support vector machines, time series, ARX, autoregressive modeling techniques, autoregressive models with exogenous input, autoregressive moving average model, breath duration time series, classification method, classifier method, discontinuing mechanical ventilation, expiratory time, feature extraction, final prediction errors, fractional inspiratory time, intensive care, interquartile range, linear discriminant analysis, logistic regression analysis, mean inspiratory flow, patient respiratory volume signal, pressure support level, pressure support ventilation, rapid shallow breathing, respiratory pattern variability characterization, support vector machines, tidal volume, weaning trial, Analytical models, Autoregressive processes, Biological system modeling, Estimation, Support vector machines, Time series analysis, Ventilation

The gold standard for diagnosing sleep apnoea-hypopnoea syndrome (SAHS) is polysomnography (PSG), an expensive, labour-intensive and time-consuming procedure. Accordingly, it would be very useful to have a screening method to allow early assessment of the severity of a subject, prior to his/her referral for PSG. Several differences have been reported between simple snorers and SAHS patients in the acoustic characteristics of snoring and its variability. In this paper, snores are fully characterised in the time domain, by their sound intensity and pitch, and in the frequency domain, by their formant frequencies and several shape and energy ratio measurements. We show that accurate multiclass classification of snoring subjects, with three levels of SAHS, can be achieved on the basis of acoustic analysis of snoring alone, without any requiring information on the duration or the number of apnoeas. Several classification methods are examined. The best of the approaches assessed is a Bayes model using a kernel density estimation method, although good results can also be obtained by a suitable combination of two binary logistic regression models. Multiclass snore-based classification allows early stratification of subjects according to their severity. This could be the basis of a single channel, snore-based screening procedure for SAHS.

JTD Keywords: Bayes classifier, Kernel density estimation, Sleep apnoea, Snoring

One objective of mechanical ventilation is the recovery of spontaneous breathing as soon as possible. Remove the mechanical ventilation is sometimes more difficult that maintain it. This paper proposes the study of respiratory flow signal of patients on weaning trials process by autoregressive moving average model (ARMA), through the location of poles and zeros of the model. A total of 151 patients under extubation process (T-tube test) were analyzed: 91 patients with successful weaning (GS), 39 patients that failed to maintain spontaneous breathing and were reconnected (GF), and 21 patients extubated after the test but before 48 hours were reintubated (GR). The optimal model was obtained with order 8, and statistical significant differences were obtained considering the values of angles of the first four poles and the first zero. The best classification was obtained between GF and GR, with an accuracy of 75.3% on the mean value of the angle of the first pole.

JTD Keywords: Analytical models, Biological system modeling, Computational modeling, Estimation, Hospitals, Poles and zeros, Ventilation, Autoregressive moving average processes, Patient care, Patient monitoring, Pneumodynamics, Poles and zeros, Ventilation, ARMA model, T-tube test, Autoregressive moving average model, Extubation process, Mechanical ventilation, Optimal model, Patient classification, Respiratory flow signal, Roots, Spontaneous breathing, Weaning trials