Detection of Unilateral Arm Paresis after Stroke by Wearable Accelerometers and Machine Learning

Recent advances in stroke treatment have provided effective tools to successfully treat ischemic stroke, but still a majority of patients are not treated due to late arrival to hospital. With modern stroke treatment, earlier arrival would greatly improve the overall treatment results. This prospective study was performed to asses the capability of bilateral accelerometers worn in bracelets 24/7 to detect unilateral arm paralysis, a hallmark symptom of stroke, early enough to receive treatment. Classical machine learning algorithms as well as state-of-the-art deep neural networks were evaluated on detection times between 15 min and 120 min. Motion data were collected using triaxial accelerometer bracelets worn on both arms for 24 h. Eighty-four stroke patients with unilateral arm motor impairment and 101 healthy subjects participated in the study. Accelerometer data were divided into data windows of different lengths and analyzed using multiple machine learning algorithms. The results show that all algorithms performed well in separating the two groups early enough to be clinically relevant, based on wrist-worn accelerometers. The two evaluated deep learning models, fully convolutional network and InceptionTime, performed better than the classical machine learning models with an AUC score between 0.947–0.957 on 15 min data windows and up to 0.993–0.994 on 120 min data windows. Window lengths longer than 90 min only marginally improved performance. The difference in performance between the deep learning models and the classical models was statistically significant according to a non-parametric Friedman test followed by a post-hoc Nemenyi test. Introduction of wearable stroke detection devices may dramatically increase the portion of stroke patients eligible for revascularization and shorten the time to treatment. Since the treatment effect is highly time-dependent, early stroke detection may dramatically improve stroke outcomes.

PredictPTB: An Interpretable Preterm Birth Prediction Model Using Attention-based Recurrent Neural Networks

Background: Early identification of pregnant women at risk for preterm birth (PTB), a major cause of infant mortality and morbidity, has a significant potential to improve prenatal care. However, we lack effective predictive models which can accurately forecast PTB and complement these predictions with appropriate interpretations for clinicians. In this work, we introduce a clinical prediction model (PredictPTB) which combines variables (medical codes) readily accessible through electronic health record (EHR) to accurately predict the risk of preterm birth at 1, 3, 6, and 9 months prior to delivery. Methods: The architecture of PredictPTB employs recurrent neural networks (RNNs) to model the longitudinal patient's EHR visits and exploits a single code-level attention mechanism to improve the predictive performance, while providing temporal code-level and visit-level explanations for the prediction results. We compare the performance of different combinations of prediction time-points, data modalities, and data windows. We also present a case-study of our model's interpretability illustrating how clinicians can gain some transparency into the predictions. Results: Leveraging a large cohort of 222,436 deliveries, comprising a total of 27,100 unique clinical concepts, our model was able to predict preterm birth with an ROC-AUC of 0.82, 0.79, 0.78, and PR-AUC of 0.40, 0.31, 0.24, at 1, 3, and 6 months prior to delivery, respectively. Results also confirm that observational data modalities (such as diagnoses) are more predictive for preterm birth than interventional data modalities (e.g., medications and procedures). Conclusions: Our results demonstrate that PredictPTB can be utilized to achieve accurate and scalable predictions for preterm birth, complemented by explanations that directly highlight evidence in the patient's EHR timeline.

Exploration of Person-Independent BCIs for Internal and External Attention-Detection in Augmented Reality

Adding attention-awareness to an Augmented Reality setting by using a Brain-Computer Interface promises many interesting new applications and improved usability. The possibly complicated setup and relatively long training period of EEG-based BCIs however, reduce this positive effect immensely. In this study, we aim at finding solutions for person-independent, training-free BCI integration into AR to classify internally and externally directed attention. We assessed several different classifier settings on a dataset of 14 participants consisting of simultaneously recorded EEG and eye tracking data. For this, we compared the classification accuracies of a linear algorithm, a non-linear algorithm, and a neural net that were trained on a specifically generated feature set, as well as a shallow neural net for raw EEG data. With a real-time system in mind, we also tested different window lengths of the data aiming at the best payoff between short window length and high classification accuracy. Our results showed that the shallow neural net based on 4-second raw EEG data windows was best suited for real-time person-independent classification. The accuracy for the binary classification of internal and external attention periods reached up to 88% accuracy with a model that was trained on a set of selected participants. On average, the person-independent classification rate reached 60%. Overall, the high individual differences could be seen in the results. In the future, further datasets are necessary to compare these results before optimizing a real-time person-independent attention classifier for AR.

Fault Location Method for Two-Terminal Untransposed Overhead Lines in Distribution Networks Measured by μMPMU

With the rapid development of distribution networks, two-terminal overhead lines have been used on a large scale for higher power supply reliability, thus the fault location has attracted much attention. Accurate fault location is helpful to shorten the outage time and improve the economy of operation greatly. However, since insufficient standardization of equipment selection and poor management, line parameters are usually inaccurate or even unknown, mature fault location methods based on impedance can’t be applied anymore. Also, the asymmetry caused by non-transposition in distribution networks affects the accuracy of fault location. This paper proposes a fault location method for two-terminal untransposed overhead lines without requiring line parameters. Firstly, this paper considers parameter asymmetry, and the mutual impedances between the three phases are set as different values. Secondly, the location equations rely on three-phase networks, then the self-impedance and mutual impedances are regarded as unknowns and solved directly. Finally, this method takes the average value of fundamental frequency components from different data windows, which reduces error and improves accuracy. The simulation results show that the fault location method has high accuracy, and can effectively overcome the influence of unknown line parameters and non-transposition.

Weather Station IoT Educational Model Using Cloud Services

IoT technology is gaining more and more popularity in practice, as it collects, processes, evaluates and stores important measured data. The IoT is used every day in the work, in the home or smart houses or in public areas. It realizes the connectivity between real world and digital world which means, that it converts physical quantities of the real world in the form of analog signals into digital numbers stored in clauds. It is essential that students gain practical experience in the design and implementation of the IoT systems during their studies. The article first describes IoT issues and communication protocols used in IoT generally are closer described. Then the design and implementation of an educational model of IoT system - Weather station with the ThingSpeak cloud support is described. The created IoT model interconnects microcontroller programming, sensors and measuring, cloud API interfaces, MATLAB scripts which are useful to analyses the stored data, Windows and Android application developing.

Fully automated, real-time, calibration-free, continuous noninvasive estimation of intracranial pressure in children

OBJECTIVEIn the search for a reliable, cooperation-independent, noninvasive alternative to invasive intracranial pressure (ICP) monitoring in children, various approaches have been proposed, but at the present time none are capable of providing fully automated, real-time, calibration-free, continuous and accurate ICP estimates. The authors investigated the feasibility and validity of simultaneously monitored arterial blood pressure (ABP) and middle cerebral artery (MCA) cerebral blood flow velocity (CBFV) waveforms to derive noninvasive ICP (nICP) estimates.METHODSInvasive ICP and ABP recordings were collected from 12 pediatric and young adult patients (aged 2–25 years) undergoing such monitoring as part of routine clinical care. Additionally, simultaneous transcranial Doppler (TCD) ultrasonography–based MCA CBFV waveform measurements were performed at the bedside in dedicated data collection sessions. The ABP and MCA CBFV waveforms were analyzed in the context of a mathematical model, linking them to the cerebral vasculature’s biophysical properties and ICP. The authors developed and automated a waveform preprocessing, signal-quality evaluation, and waveform-synchronization “pipeline” in order to test and objectively validate the algorithm’s performance. To generate one nICP estimate, 60 beats of ABP and MCA CBFV waveform data were analyzed. Moving the 60-beat data window forward by one beat at a time (overlapping data windows) resulted in 39,480 ICP-to-nICP comparisons across a total of 44 data-collection sessions (studies). Moving the 60-beat data window forward by 60 beats at a time (nonoverlapping data windows) resulted in 722 paired ICP-to-nICP comparisons.RESULTSGreater than 80% of all nICP estimates fell within ± 7 mm Hg of the reference measurement. Overall performance in the nonoverlapping data window approach gave a mean error (bias) of 1.0 mm Hg, standard deviation of the error (precision) of 5.1 mm Hg, and root-mean-square error of 5.2 mm Hg. The associated mean and median absolute errors were 4.2 mm Hg and 3.3 mm Hg, respectively. These results were contingent on ensuring adequate ABP and CBFV signal quality and required accurate hydrostatic pressure correction of the measured ABP waveform in relation to the elevation of the external auditory meatus. Notably, the procedure had no failed attempts at data collection, and all patients had adequate TCD data from at least one hemisphere. Last, an analysis of using study-by-study averaged nICP estimates to detect a measured ICP > 15 mm Hg resulted in an area under the receiver operating characteristic curve of 0.83, with a sensitivity of 71% and specificity of 86% for a detection threshold of nICP = 15 mm Hg.CONCLUSIONSThis nICP estimation algorithm, based on ABP and bedside TCD CBFV waveform measurements, performs in a manner comparable to invasive ICP monitoring. These findings open the possibility for rational, point-of-care treatment decisions in pediatric patients with suspected raised ICP undergoing intensive care.

Simultaneous reconstruction of seismic reflections and diffractions using a global hyperbolic Radon dictionary

We have developed a new transform with basis functions that closely resemble seismic reflections and diffractions. The new transform is an extension of the classic hyperbolic Radon transform and accounts for the apex shifts of the seismic reflection hyperbolas and the asymptote shifts of the seismic diffraction hyperbolas. The adjoint and forward operators of the proposed transform are computed using Stolt operators in the frequency domain to increase the computational efficiency of the transform. This new transform is used, in conjunction with a sparse inversion algorithm, to reconstruct common-shot gathers. Our tests indicate that this new transform is an efficient tool for interpolating coarsely sampled seismic data in cases in which one cannot use small data windows to validate the linear event assumption that is often made by Fourier-based reconstruction methods.