Physiological Profiling of Agitation in Dementia: Insights From Wearable Sensor Data.

Background and Objectives: The number of people with dementia is expected to triple to 152 million in 2050, with 90% having accompanying behavioral and psychological symptoms (BPSD). Agitation is among the most critical BPSD and can lead to decreased quality of life for people with dementia and their caregivers. This study aims to explore objective quantification of agitation in people with dementia by analyzing the relationships between physiological and movement data from wearables and observational measures of agitation. Research Design and Methods: The data presented here is from 30 people with dementia, each included for 1 week, collected following our previously published multimodal data collection protocol. This observational protocol has a cross-sectional repeated measures design, encompassing data from both wearable and fixed sensors. Generalized linear mixed models were used to quantify the relationship between data from different wearable sensor modalities and agitation, as well as motor and verbal agitation specifically. Results: Several features from wearable data are significantly associated with agitation, at least the p < .05 level (absolute ß: 0.224-0.753). Additionally, different features are informative depending on the agitation type or the patient the data were collected from. Adding context with key confounding variables (time of day, movement, and temperature) allows for a clearer interpretation of feature differences when a person with dementia is agitated. Discussion and Implications: The features shown to be significantly different, across the study population, suggest possible autonomic nervous system activation when agitated. Differences when splitting the data by agitation type point toward a need for future detection models to tailor to the primary type of agitation expressed. Finally, patient-specific differences in features indicate a need for patient- or group-level model personalization. The findings reported in this study both reinforce and add to the fundamental understanding of and can be used to drive the objective quantification of agitation.

Advancing the Local Pulse Wave Velocity Measurement - Wave Confluence Decomposition using a Double Gaussian Propagation Model.

Background Pulse wave velocity (PWV) is a marker of arterial stiffness and local measurements could facilitate its widescale clinical use. However, confluence of incident and early reflected waves leads to biased spatiotemporal PWV estimates. Objective We introduce the Double Gaussian Propagation Model (DGPM) to measure local PWV in consideration of wave confluence (PWVDGPM) and compare it against conventional spatiotemporal PWV (PWVST), with Bramwell-Hill PWV (PWVBH) and blood pressure (BP) as reference measures. Methods Ten subjects ranging from normotension to hypertension were repeatedly measured at rest and with induced PWV changes. Carotid distension waveforms over a 19 mm wide segment were acquired from ultrasonography, simultaneously with noninvasive continuous BP. Per cardiac cycle, the 8-parameter DGPM (amplitude, centroid, width, and velocity, respectively of forward and backward propagating wave) was fitted to the distension waveforms' systolic foot and dicrotic notch complexes. Corresponding PWVST was computed from linear fittings of respective feature timings and distances. Regression analyses were conducted with PWVDGPM and PWVST as predictors, and various PWV and BP measures as response variables. Results Whereas PWVST correlations were insignificant, PWVDGPM estimated the reference PWVBH with a significant reduction in errors (P<0.001), explained up to 65% PWVBH variability at rest, demonstrated higher intra-method consistency and correlated significantly with all BP measures (P<0.001). Conclusion The proposed DGPM measures local carotid PWV in consideration of wave confluence, showing significant correlations with Bramwell-Hill PWV and BP at two distinct waveform complexes. Thereby PWVDGPM outperforms the conventional PWVST in all investigated respects, potentially enabling PWV assessment in routine clinical practice.

The Feasibility of Semi-Continuous and Multi-Frequency Thoracic Bioimpedance Measurements by a Wearable Device during Fluid Changes in Hemodialysis Patients.

Repeated single-point measurements of thoracic bioimpedance at a single (low) frequency are strongly related to fluid changes during hemodialysis. Extension to semi-continuous measurements may provide longitudinal details in the time pattern of the bioimpedance signal, and multi-frequency measurements may add in-depth information on the distribution between intra- and extracellular fluid. This study aimed to investigate the feasibility of semi-continuous multi-frequency thoracic bioimpedance measurements by a wearable device in hemodialysis patients. Therefore, thoracic bioimpedance was recorded semi-continuously (i.e., every ten minutes) at nine frequencies (8-160 kHz) in 68 patients during two consecutive hemodialysis sessions, complemented by a single-point measurement at home in-between both sessions. On average, the resistance signals increased during both hemodialysis sessions and decreased during the interdialytic interval. The increase during dialysis was larger at 8 kHz (∆ 32.6 Ω during session 1 and ∆ 10 Ω during session 2), compared to 160 kHz (∆ 29.5 Ω during session 1 and ∆ 5.1 Ω during session 2). Whereas the resistance at 8 kHz showed a linear time pattern, the evolution of the resistance at 160 kHz was significantly different (p < 0.0001). Measuring bioimpedance semi-continuously and with a multi-frequency current is a major step forward in the understanding of fluid dynamics in hemodialysis patients. This study paves the road towards remote fluid monitoring.

Comparison of whole body versus thoracic bioimpedance in relation to ultrafiltration volume and systolic blood pressure during hemodialysis.

In contrast to whole body bioimpedance, which estimates fluid status at a single point in time, thoracic bioimpedance applied by a wearable device could enable continuous measurements. However, clinical experience with thoracic bioimpedance in patients on dialysis is limited. To test the reproducibility of whole body and thoracic bioimpedance measurements and to compare their relationship with hemodynamic changes during hemodialysis, these parameters were measured pre- and end-dialysis in 54 patients during two sessions. The resistance from both bioimpedance techniques was moderately reproducible between two dialysis sessions (intraclass correlations of pre- to end-dialysis whole body and thoracic resistance between session 1 and 2 were 0.711 [0.58-0.8] and 0.723 [0.6-0.81], respectively). There was a very high to high correlation between changes in ultrafiltration volume and changes in whole body thoracic resistance. Changes in systolic blood pressure negatively correlated to both bioimpedance techniques. Although the relationship between changes in ultrafiltration volume and changes in resistance was stronger for whole body bioimpedance, the relationship with changes in blood pressure was at least comparable for thoracic measurements. These results suggest that thoracic bioimpedance, measured by a wearable device, may serve as an interesting alternative to whole body measurements for continuous hemodynamic monitoring during hemodialysis.NEW & NOTEWORTHY We examined the role of whole body and thoracic bioimpedance in hemodynamic changes during hemodialysis. Whole body and thoracic bioimpedance signals were strongly related to ultrafiltration volume and moderately, negatively, to changes in blood pressure. This work supports the further development of a wearable device measuring thoracic bioimpedance longitudinally in patients on hemodialysis. As such, it may serve as an innovative tool for continuous hemodynamic monitoring during hemodialysis in hospital or in a home-based setting.

Comparison between Speckle Plethysmography and Photoplethysmography during Cold Pressor Test Referenced to Finger Arterial Pressure.

Speckle Plethysmography (SPG) and Photoplethysmography (PPG) are different biophotonics technologies that allow for measurement of haemodynamics. As the difference between SPG and PPG under low perfusion conditions is not fully understood, a Cold Pressor Test (CPT-60 s full hand immersion in ice water), was used to modulate blood pressure and peripheral circulation. A custom-built setup simultaneously derived SPG and PPG from the same video streams at two wavelengths (639 nm and 850 nm). SPG and PPG were measured at the right index finger location before and during the CPT using finger Arterial Pressure (fiAP) as a reference. The effect of the CPT on the Alternating Component amplitude (AC) and Signal-to-Noise Ratio (SNR) of dual-wavelength SPG and PPG signals was analysed across participants. Furthermore, waveform differences between SPG, PPG, and fiAP based on frequency harmonic ratios were analysed for each subject (n = 10). Both PPG and SPG at 850 nm show a significant reduction during the CPT in both AC and SNR. However, SPG showed significantly higher and more stable SNR than PPG in both study phases. Harmonic ratios were found substantially higher in SPG than PPG. Therefore, in low perfusion conditions, SPG seems to offer a more robust pulse wave monitoring with higher harmonic ratios than PPG.

Comparing Remote Speckle Plethysmography and Finger-Clip Photoplethysmography with Non-Invasive Finger Arterial Pressure Pulse Waves, Regarding Morphology and Arrival Time.

OBJECTIVE: The goal was to compare Speckle plethysmography (SPG) and Photoplethysmography (PPG) with non-invasive finger Arterial Pressure (fiAP) regarding Pulse Wave Morphology (PWM) and Pulse Arrival Time (PAT). METHODS: Healthy volunteers (n = 8) were connected to a Non-Invasive Blood Pressure (NIBP) monitor providing fiAP pulse wave and PPG from a clinical transmission-mode SpO2 finger clip. Biopac recorded 3-lead ECG. A camera placed at a 25 cm distance recorded a video stream (100 fps) of a finger illuminated by a laser diode at 639 nm. A chest belt (Polar) monitored respiration. All signals were recorded simultaneously during episodes of spontaneous breathing and paced breathing. ANALYSIS: Post-processing was performed in Matlab to obtain SPG and analyze the SPG, PPG and fiAP mean absolute deviations (MADs) on PWM, plus PAT modulation. RESULTS: Across 2599 beats, the average fiAP MAD with PPG was 0.17 (0-1) and with SPG 0.09 (0-1). PAT derived from ECG-fiAP correlated as follows: 0.65 for ECG-SPG and 0.67 for ECG-PPG. CONCLUSION: Compared to the clinical NIBP monitor fiAP reference, PWM from an experimental camera-derived non-contact reflective-mode SPG setup resembled fiAP significantly better than PPG from a simultaneously recorded clinical transmission-mode finger clip. For PAT values, no significant difference was found between ECG-SPG and ECG-PPG compared to ECG-fiAP.

Exploring the perception of stress in childhood and early adolescence.

Stressful life experiences may jeopardize the healthy development of children. To improve interventions, more knowledge is needed on the perception of stress by children. In adults, stress is regarded as a state of low valence and high arousal. It is unclear whether children perceive stress similarly. In the current study, 35 children of the general population completed three tasks aiming to provide insight into their knowledge of the concept stress. In the first task, participants were asked about their verbal knowledge of the concept stress. In the second task, they rated the valence and arousal of eight emotion-evoking vignettes. In the final task, participants completed an experience sampling survey for at least 1 day, consisting of a stress thermometer and pictorial scales of valence and arousal. Participants' perception of stress was found to be mainly valence focused. Age and sex were found to play a role in the degree of arousal focus. Older participants differentiated more in arousal levels than younger participants, as did girls in comparison with boys. Because the perception of stress depends on developmental and other individual factors, using stress as a single measurement dimension in a survey is not recommended.

Electric field gradient focusing with electro-osmotic flow to reduce analyte dispersion: Concept and numerical investigation.

In proteomics, the need to precisely examine the protein compounds of small samples, requires sensitive analytical methods which can separate and enrich compounds with high precision. Current techniques require a minimal analysis time to obtain satisfactory compound separation where longer analysis time means better separation of compounds. But, molecular diffusion will create broadening of the separated compound bands over time, increasing the peak width, and thus reducing the resolution and the enrichment. Electric field gradient focusing (EFGF) is a separation technique, in which proteins are simultaneously separated and enriched by balancing a gradient electrostatic force with a constant hydrodynamic drag force. Because of this balance, analytes are continuously pushed back to their focusing point, limiting the time-dependent peak broadening due to molecular diffusion. Current EFGF techniques are however still suffering from peak broadening because of flow-profile inhomogeneities. In this paper, we propose to use AC electro-osmotic flow (AC EOF) to create a homogeneous flow in EFGF. The interference between the electric field gradient and the AC EOF was thoroughly analysed and the concept was validated using numerical simulations. The results show that a plug flow is obtained on top of a small, distorted boundary layer. While applying different DC electric fields in the electrolyte, a constant flow velocity can be obtained by including a DC offset to the electrodes generating the AC EOF. The plug flow is then maintained over the whole separation channel length, while an electric field gradient is applied. This way, the flow-induced contribution to peak broadening can be minimized in EFGF devices. By modelling the separation of green fluorescent protein (GFP) and R-Phycoerythrin (R-PE), it was shown that the peak width of separated compounds can be reduced and that the separation resolution can be improved, compared to current EFGF methods.

The cumulative effect of chronic stress and depressive symptoms affects heart rate in a working population.

Background: Chronic stress and depressive symptoms have both been linked to increased heart rate (HR) and reduced HR variability. However, up to date, it is not clear whether chronic stress, the mechanisms intrinsic to depression or a combination of both cause these alterations. Subclinical cases may help to answer these questions. In a healthy working population, we aimed to investigate whether the effect of chronic stress on HR circadian rhythm depends on the presence of depressive symptoms and whether chronic stress and depressive symptoms have differential effects on HR reactivity to an acute stressor. Methods: 1,002 individuals of the SWEET study completed baseline questionnaires, including psychological information, and 5 days of electrocardiogram (ECG) measurements. Complete datasets were available for 516 individuals. In addition, a subset (n = 194) of these participants completed a stress task on a mobile device. Participants were grouped according to their scores for the Depression Anxiety Stress Scale (DASS) and Perceived Stress Scale (PSS). We explored the resulting groups for differences in HR circadian rhythm and stress reactivity using linear mixed effect models. Additionally, we explored the effect of stress and depressive symptoms on night-time HR variability [root mean square of successive differences (RMSSD)]. Results: High and extreme stress alone did not alter HR circadian rhythm, apart from a limited increase in basal HR. Yet, if depressive symptoms were present, extreme chronic stress levels did lead to a blunted circadian rhythm and a lower basal HR. Furthermore, blunted stress reactivity was associated with depressive symptoms, but not chronic stress. Night-time RMSSD data was not influenced by chronic stress, depressive symptoms or their interaction. Conclusion: The combination of stress and depressive symptoms, but not chronic stress by itself leads to a blunted HR circadian rhythm. Furthermore, blunted HR reactivity is associated with depressive symptoms and not chronic stress.

Camera-Derived Photoplethysmography (rPPG) and Speckle Plethysmography (rSPG): Comparing Reflective and Transmissive Mode at Various Integration Times Using LEDs and Lasers.

Background: Although both speckle plethysmography (SPG) and photoplethysmography (PPG) examine pulsatile changes in the vasculature using opto-electronics, PPG has a long history, whereas SPG is relatively new and less explored. The aim of this study was to compare the effects of integration time and light-source coherence on signal quality and waveform morphology for reflective and transmissive rSPG and rPPG. Methods: (A) Using time-domain multiplexing, we illuminated 10 human index fingers with pulsed lasers versus LEDs (both at 639 and 850 nm), in transmissive versus reflective mode. A synchronized camera (Basler acA2000-340 km, 25 cm distance, 200 fps) captured and demultiplexed four video channels (50 fps/channel) in four stages defined by illumination mode. From all video channels, we derived rPPG and rSPG, and applied a signal quality index (SQI, scale: Good > 0.95; Medium 0.95−0.85; Low 0.85−0.8; Negligible < 0.8); (B) For transmission videos only, we additionally calculated the intensity threshold area (ITA), as the area of the imaging exceeding a certain intensity value and used linear regression analysis to understand unexpected similarities between rPPG and rSPG. Results: All mean SQI-values. Reflective mode: Laser-rSPG > 0.965, LED-rSPG < 0.78, rPPG < 0.845. Transmissive mode: 0.853−0.989 for rSPG and rPPG at all illumination settings. Coherent mode: Reflective rSPG > 0.951, reflective rPPG < 0.740, transmissive rSPG and rPPG 0.990−0.898. Incoherent mode: Reflective all <0.798 and transmissive all 0.92−0.987. Linear regressions revealed similar R2 values of rPPG with rSPG (R2 = 0.99) and ITA (R2 = 0.98); Discussion: Laser-rSPG and LED-rPPG produced different waveforms in reflection, but not in transmission. We created the concept of ITA to investigate this behavior. Conclusions: Reflective Laser-SPG truly originated from coherence. Transmissive Laser-rSPG showed a loss of speckles, accompanied by waveform changes towards rPPG. Diffuse spatial intensity modulation polluted spatial-mode SPG.

Measuring Health-Related Quality of Life With Multimodal Data: Viewpoint.

The ability to objectively measure aspects of performance and behavior is a fundamental pillar of digital health, enabling digital wellness products, decentralized trial concepts, evidence generation, digital therapeutics, and more. Emerging multimodal technologies capable of measuring several modalities simultaneously and efforts to integrate inputs across several sources are further expanding the limits of what digital measures can assess. Experts from the field of digital health were convened as part of a multi-stakeholder workshop to examine the progress of multimodal digital measures in two key areas: detection of disease and the measurement of meaningful aspects of health relevant to the quality of life. Here we present a meeting report, summarizing key discussion points, relevant literature, and finally a vision for the immediate future, including how multimodal measures can provide value to stakeholders across drug development and care delivery, as well as three key areas where headway will need to be made if we are to continue to build on the encouraging progress so far: collaboration and data sharing, removal of barriers to data integration, and alignment around robust modular evaluation of new measurement capabilities.

Data Augmentation and Transfer Learning for Data Quality Assessment in Respiratory Monitoring.

Changes in respiratory rate have been found to be one of the early signs of health deterioration in patients. In remote environments where diagnostic tools and medical attention are scarce, such as deep space exploration, the monitoring of the respiratory signal becomes crucial to timely detect life-threatening conditions. Nowadays, this signal can be measured using wearable technology; however, the use of such technology is often hampered by the low quality of the recordings, which leads more often to wrong diagnosis and conclusions. Therefore, to apply these data in diagnosis analysis, it is important to determine which parts of the signal are of sufficient quality. In this context, this study aims to evaluate the performance of a signal quality assessment framework, where two machine learning algorithms (support vector machine-SVM, and convolutional neural network-CNN) were used. The models were pre-trained using data of patients suffering from chronic obstructive pulmonary disease. The generalization capability of the models was evaluated by testing them on data from a different patient population, presenting normal and pathological breathing. The new patients underwent bariatric surgery and performed a controlled breathing protocol, displaying six different breathing patterns. Data augmentation (DA) and transfer learning (TL) were used to increase the size of the training set and to optimize the models for the new dataset. The effect of the different breathing patterns on the performance of the classifiers was also studied. The SVM did not improve when using DA, however, when using TL, the performance improved significantly (p < 0.05) compared to DA. The opposite effect was observed for CNN, where the biggest improvement was obtained using DA, while TL did not show a significant change. The models presented a low performance for shallow, slow and fast breathing patterns. These results suggest that it is possible to classify respiratory signals obtained with wearable technologies using pre-trained machine learning models. This will allow focusing on the relevant data and avoid misleading conclusions because of the noise, when designing bio-monitoring systems.

Characterizing and Modeling Smoking Behavior Using Automatic Smoking Event Detection and Mobile Surveys in Naturalistic Environments: Observational Study.

BACKGROUND: There are 1.1 billion smokers worldwide, and each year, more than 8 million die prematurely because of cigarette smoking. More than half of current smokers make a serious quit every year. Nonetheless, 90% of unaided quitters relapse within the first 4 weeks of quitting due to the lack of limited access to cost-effective and efficient smoking cessation tools in their daily lives. OBJECTIVE: This study aims to enable quantified monitoring of ambulatory smoking behavior 24/7 in real life by using continuous and automatic measurement techniques and identifying and characterizing smoking patterns using longitudinal contextual signals. This work also intends to provide guidance and insights into the design and deployment of technology-enabled smoking cessation applications in naturalistic environments. METHODS: A 4-week observational study consisting of 46 smokers was conducted in both working and personal life environments. An electric lighter and a smartphone with an experimental app were used to track smoking events and acquire concurrent contextual signals. In addition, the app was used to prompt smoking-contingent ecological momentary assessment (EMA) surveys. The smoking rate was assessed based on the timestamps of smoking and linked statistically to demographics, time, and EMA surveys. A Poisson mixed-effects model to predict smoking rate in 1-hour windows was developed to assess the contribution of each predictor. RESULTS: In total, 8639 cigarettes and 1839 EMA surveys were tracked over 902 participant days. Most smokers were found to have an inaccurate and often biased estimate of their daily smoking rate compared with the measured smoking rate. Specifically, 74% (34/46) of the smokers made more than one (mean 4.7, SD 4.2 cigarettes per day) wrong estimate, and 70% (32/46) of the smokers overestimated it. On the basis of the timestamp of the tracked smoking events, smoking rates were visualized at different hours and were found to gradually increase and peak at 6 PM in the day. In addition, a 1- to 2-hour shift in smoking patterns was observed between weekdays and weekends. When moderate and heavy smokers were compared with light smokers, their ages (P<.05), Fagerström Test of Nicotine Dependence (P=.01), craving level (P<.001), enjoyment of cigarettes (P<.001), difficulty resisting smoking (P<.001), emotional valence (P<.001), and arousal (P<.001) were all found to be significantly different. In the Poisson mixed-effects model, the number of cigarettes smoked in a 1-hour time window was highly dependent on the smoking status of an individual (P<.001) and was explained by hour (P=.02) and age (P=.005). CONCLUSIONS: This study reported the high potential and challenges of using an electronic lighter for smoking annotation and smoking-triggered EMAs in an ambulant environment. These results also validate the techniques for smoking behavior monitoring and pave the way for the design and deployment of technology-enabled smoking cessation applications. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): RR2-10.1136/bmjopen-2018-028284.

Toward Quantification of Agitation in People With Dementia Using Multimodal Sensing.

Background and Objectives: Agitation, a critical behavioral and psychological symptom in dementia, has a profound impact on a patients' quality of life as well as their caregivers'. Autonomous and objective characterization of agitation with multimodal systems has the potential to capture key patient responses or agitation triggers. Research Design and Methods: In this article, we describe our multimodal system design that encompasses contextual parameters, physiological parameters, and psychological parameters. This design is the first to include all three of these facets in an n > 1 study. Using a combination of fixed and wearable sensors and a custom-made app for psychological annotation, we aim to identify physiological markers and contextual triggers of agitation. Results: A discussion of both the clinical as well as the technical implementation of the to-date data collection protocol is presented, as well as initial insights into pilot study data collection. Discussion and Implications: The ongoing data collection moves us toward improved agitation quantification and subsequent prediction, eventually enabling just-in-time intervention.

LSTM-only Model for Low-complexity HR Estimation from Wrist PPG.

Continuous and non-invasive cardiovascular monitoring has gained attention due to the miniaturization of wearable devices. Particularly, wrist-worn photoplethysmography (PPG) sensors present an alternative to electrocardiogram recording for heart rate (HR) monitoring as it is cheaper and non-intrusive for daily activities. Yet, the accuracy of PPG measurements is heavily affected by motion artifacts which are inherent to ambulatory environments. In this paper, we propose a low-complexity LSTM-only neural network for HR estimation from a single PPG channel during intense physical activity. This work explored the trade-off between model complexity and accuracy by exploring different model dataflows, number of layers, and number of training epochs to capture the intrinsic time-dependency between PPG samples. The best model achieves a mean absolute error of 4.47 ± 3.68 bpm when evaluated on 12 IEEE SPC subjects.Clinical relevance- This work aims to improve the quality of HR inference from PPG signals using neural network, enabling continuous vital signal monitoring with little interference in daily activities from embedded monitoring devices.

A 134 DB Dynamic Range Noise Shaping Slope Light-to-Digital Converter for Wearable Chest PPG Applications.

This paper presents a low power, high dynamic range (DR), light-to-digital converter (LDC) for wearable chest photoplethysmogram (PPG) applications. The proposed LDC utilizes a novel 2nd-order noise-shaping slope architecture, directly converting the photocurrent to a digital code. This LDC applies a high-resolution dual-slope quantizer for data conversion. An auxiliary noise shaping loop is used to shape the residual quantization noise. Moreover, a DC compensation loop is implemented to cancel the PPG signal's DC component, thus further boosting the DR. The prototype is fabricated with 0.18 µm standard CMOS and characterized experimentally. The LDC consumes 28 µW per readout channel while achieving a maximum 134 dB DR. The LDC is also validated with on-body chest PPG measurement.

Performance and limitation estimation of a three-tap gated imaging sensor in wide field time-gated fluorescence lifetime imaging systems.

In this paper, a computational performance analysis is presented of a wide-field time-gated fluorescence lifetime imaging microscope (FLIM) using practically realizable properties of the laser, sample, and a three-tap time-gated CMOS image sensor. The impact of these component-level properties on the accuracy and the precision of the measurement results are estimated and discussed based on Monte Carlo simulations. The correlation between the detector speed and the accuracy of the extracted fluorescence lifetime is studied, and the minimum required incident photoelectron number of each pixel is estimated for different detector speeds and different fluorescence lifetime measurements. In addition, the detection limits due to the dark current and the parasitic light sensitivity of the detector are also investigated. This work gives an overview of the required fluorescence emission condition as well as the required detector properties for a three-tap time-gated image sensor to achieve good FLIM data in biological applications.

Technical aspects of cardiorespiratory estimation using subspace projections and cross entropy.

Background.Respiratory sinus arrhythmia (RSA) is a form of cardiorespiratory coupling. Its quantification has been suggested as a biomarker to diagnose different diseases. Two state-of-the-art methods, based on subspace projections and entropy, are used to estimate the RSA strength and are evaluated in this paper. Their computation requires the selection of a model order, and their performance is strongly related to the temporal and spectral characteristics of the cardiorespiratory signals.Objective.To evaluate the robustness of the RSA estimates to the selection of model order, delays, changes of phase and irregular heartbeats as well as to give recommendations for their interpretation on each case.Approach.Simulations were used to evaluate the model order selection when calculating the RSA estimates introduced before, as well as three different scenarios that can occur in signals acquired in non-controlled environments and/or from patient populations: the presence of irregular heartbeats; the occurrence of delays between heart rate variability (HRV) and respiratory signals; and the changes over time of the phase between HRV and respiratory signals.Main results.It was found that using a single model order for all the calculations suffices to characterize RSA correctly. In addition, the RSA estimation in signals containing more than 5 irregular heartbeats in a period of 5 min might be misleading. Regarding the delays between HRV and respiratory signals, both estimates are robust. For the last scenario, the two approaches tolerate phase changes up to 54°, as long as this lasts less than one fifth of the recording duration.Significance.Guidelines are given to compute the RSA estimates in non-controlled environments and patient populations.

Twenty-Four-Hour Heart Rate Is a Trait but Not State Marker for Depression in a Pilot Randomized Controlled Trial With a Single Infusion of Ketamine.

Background: Abnormalities of heart rate (HR) and its variability are characteristic of major depressive disorder (MDD). However, circadian rhythm is rarely taken into account when statistically exploring state or trait markers for depression. Methods: A 4-day electrocardiogram was recorded for 16 treatment-resistant patients with MDD and 16 age- and sex-matched controls before, and for the patient group only, after a single treatment with the rapid-acting antidepressant ketamine or placebo (clinical trial registration available on https://www.clinicaltrialsregister.eu/ with EUDRACT number 2016-001715-21). Circadian rhythm differences of HR and the root mean square of successive differences (RMSSD) were compared between groups and were explored for classification purposes. Baseline HR/RMSSD were tested as predictors for treatment response, and physiological measures were assessed as state markers. Results: Patients showed higher HR and lower RMSSD alongside marked reductions in HR amplitude and RMSSD variation throughout the day. Excellent classification accuracy was achieved using HR during the night, particularly between 2 and 3 a.m. (90.6%). A positive association between baseline HR and treatment response (r = 0.55, p = 0.046) pointed toward better treatment outcome in patients with higher HR. Heart rate also decreased significantly following treatment but was not associated with improved mood after a single infusion of ketamine. Limitations: Our study had a limited sample size, and patients were treated with concomitant antidepressant medication. Conclusion: Patients with depression show a markedly reduced amplitude for HR and dysregulated RMSSD fluctuation. Higher HR and lower RMSSD in depression remain intact throughout a 24-h day, with the highest classification accuracy during the night. Baseline HR levels show potential for treatment response prediction but did not show potential as state markers in this study. Clinical trial registration: EUDRACT number 2016-001715-21.

Artefact Detection in Impedance Pneumography Signals: A Machine Learning Approach.

Impedance pneumography has been suggested as an ambulatory technique for the monitoring of respiratory diseases. However, its ambulatory nature makes the recordings more prone to noise sources. It is important that such noisy segments are identified and removed, since they could have a huge impact on the performance of data-driven decision support tools. In this study, we investigated the added value of machine learning algorithms to separate clean from noisy bio-impedance signals. We compared three approaches: a heuristic algorithm, a feature-based classification model (SVM) and a convolutional neural network (CNN). The dataset consists of 47 chronic obstructive pulmonary disease patients who performed an inspiratory threshold loading protocol. During this protocol, their respiration was recorded with a bio-impedance device and a spirometer, which served as a gold standard. Four annotators scored the signals for the presence of artefacts, based on the reference signal. We have shown that the accuracy of both machine learning approaches (SVM: 87.77 ± 2.64% and CNN: 87.20 ± 2.78%) is significantly higher, compared to the heuristic approach (84.69 ± 2.32%). Moreover, no significant differences could be observed between the two machine learning approaches. The feature-based and neural network model obtained a respective AUC of 92.77±2.95% and 92.51±1.74%. These findings show that a data-driven approach could be beneficial for the task of artefact detection in respiratory thoracic bio-impedance signals.