BIOSIGNALS 2019 Abstracts


Full Papers
Paper Nr: 6
Title:

Elimination of Boundary Effects at the Numerical Implementation of Continuous Wavelet Transform to Nonstationary Biomedical Signals

Authors:

S. V. Bozhokin, I. B. Suslova and D. A. Tarakanov

Abstract: The main objective of the paper is to develop an algorithm improving the situation with boundary effects at the numerical implementation of continuous wavelet transform, what makes it possible to hold down important information in the study of many nonstationary biological signals. As a basis for the research, we propose the mathematical model of nonstationary signal (NS) with a complex amplitude variation in time. Such signals show peaks near the beginning and end of the observation period. The model leads to the analytical expression for continuous wavelet transform (CWT). Thus, we get to know how boundary effects relate to the finiteness in time of the mother wavelet function. To avoid boundary effects arising at the procedure of numerical wavelet transform, we develop the algorithm of signal time shift (STS). The results of analytical solution and numerical CWT calculation with and without the use of STS show the benefit of using STS in the processing of nonstationary signals. We have applied the technique to the study of nonstationary heart tachogram and discovered heart activity bursts in different spectral ranges, which appear and disappear at different time moments.

Paper Nr: 11
Title:

Classification of Cardiac Arrhythmias from Single Lead ECG with a Convolutional Recurrent Neural Network

Authors:

Jérôme Van Zaen, Olivier Chételat, Mathieu Lemay, Enric M. Calvo and Ricard Delgado-Gonzalo

Abstract: While most heart arrhythmias are not immediately harmful, they can lead to severe complications. In particular, atrial fibrillation, the most common arrhythmia, is characterized by fast and irregular heart beats and increases the risk of suffering a stroke. To detect such abnormal heart conditions, we propose a system composed of two main parts: a smart vest with two cooperative sensors to collect ECG data and a neural network architecture to classify heart rhythms. The smart vest uses two dry bi-electrodes to record a single lead ECG signal. The biopotential signal is then streamed via a gateway to the cloud where a neural network detects and classifies the heart arrhythmias. We selected an architecture that combines convolutional and recurrent layers. The convolutional layers extract relevant features from sliding windows of ECG and the recurrent layer aggregates them for a final softmax layer that performs the classification. Our neural network achieves an accuracy of 87.50% on the dataset of the challenge of Computing in Cardiology 2017.

Paper Nr: 14
Title:

Heart Rate Variability and Electrodermal Activity in Mental Stress Aloud: Predicting the Outcome

Authors:

Rodrigo Lima, Daniel Osório and Hugo Gamboa

Abstract: The assessment of changes in the autonomous nervous system (ANS), have important prognostic and diagnostic value, and can be used to assess stress levels. There are many approaches to directly measure the sympathetic and parasympathetic nervous system, although, most of them are invasive and unable to provide continuous monitoring. Heart rate variability (HRV) and Electrodermal activity (EDA) are noninvasive methods to assess the autonomous nervous system, by computing the spectral analysis of both HRV and EDA biosignals. In order to provide continuous monitoring, a wearable device is used, obtaining HRV features with photoplethysmography signals from the wrist and EDA from the fingers. The extraction of the HRV and EDA features, were obtained by submitting the subjects to a mental arithmetic stress test. The distinct response to stress was then classified using machine-learning techniques. The constructed models have the ability to predict how the subjects will respond, with an accuracy of approximately 80% in terms of HRV features in baseline and an accuracy of approximately 77% in terms of HRV and EDA simultaneous baseline features, when submitted to a situation of stress.

Paper Nr: 15
Title:

Comparing Parkinson's Disease Dysarthria and Aging Speech using Articulation Kinematics

Authors:

A. Gómez-Rodellar, D. Palacios-Alonso, J. Mekyska, A. Álvarez-Marquina and P. Gómez-Vilda

Abstract: Speech is being considered a pervasive and costless means to detect and monitor neurodegenerative disease progression. Many different approaches have been reported to differentiate normative subject speech from neurodegenerative patient speech. Most of them are focussed on statistical pattern recognition approaches to improve detection results on a baseline, considering only patient speech and normative controls. The definition of a normative control is not well established in itself, usually being subjects free of any pathology aligned in the same age range as patients. But one question which is not taken into account is the effects of aging in healthy controls, as usually neurodegenerative diseases may include mostly patients affected by certain effects, as dysphonia or dysarthria, as a consequence of aging. The present research introduces a methodology based on information theory to compare the effects produced by aging dysarthria with those due to Parkinson’s Disease, using the statistical distribution of speech articulation kinematics as a marker. On the one hand, it may be concluded that articulation kinematics is substantially different for PD and HC with respect to normative subjects. On the other hand, this does not seem to be the case between PD and HC subjects, as these subsets may share some dysarthric features which may be contributed more by aging than by neuromotor degeneration. This differentiation problem needs to be evaluated as well in the case of phonation features, otherwise there will not be full guarantee in using phonation features to assess neuromotor degeneration. In this sense new methodologies have to be designed to distinguish neurodegenerative from aging speech granting better guarantees.

Paper Nr: 27
Title:

Comparing Different Settings of Parameters Needed for Pre-processing of ECG Signals used for Blood Pressure Classification

Authors:

Monika Simjanoska, Gregor Papa, Barbara K. Seljak and Tome Eftimov

Abstract: Because a real-time monitoring using electrocardiogram (ECG) signals is a challenging task, the pre-processing techniques used for ECG signal analysis are crucial for obtaining information that is further used for some more complex analysis, such as predictive analyses. We compared different settings of parameters needed for pre-processing of ECG signals in order to estimate the valuable information that can be further used for blood pressure classification. Two parameters were involved in the comparison: i) the signal length used for ECG segmentation; and ii) the cut-off frequency used for baseline removal. The first parameter is the parameter used for obtaining ECG segments that are further used, and the second one is the frequency used for baseline removal. Thirty different combinations, each a combination of a signal length and a cut-off frequency, were evaluated using a dataset that contains data from five commercially available ECG sensors. For signal lengths: 10 s, 20 s, and 30 s, were used for data segmentation, while the cut-off frequency for baseline removal starts from 0.05 Hz, till 0.50 Hz, with a step length of 0.05 Hz. The evaluation of these combinations was done in combination with complexity analysis used for features extraction that are further used for blood pressure classification. Experimental results, obtained using a data-driven approach by comparing the combinations using the results obtained from the classification for 17 performance measures, showed that a signal length of 30 s carries the most information in a combination with cut-off frequency between 0.10 Hz and 0.20 Hz. Results contribute to the arguments published in the literature discussing the optimal ECG sample lengths needed for building predictive models, as well as the lower frequencies where the ECG components overlap with the baseline wander noise.

Paper Nr: 28
Title:

Physiotherapy Exercises Evaluation using a Combined Approach based on sEMG and Wearable Inertial Sensors

Authors:

Ana Pereira, Duarte Folgado, Ricardo Cotrim and Inês Sousa

Abstract: The efficacy of home-based physiotherapy depends on the correct and systematic execution of prescribed exercises. Biofeedback systems enable to accurately track exercise execution and prevent patients from unconsciously introduce incorrect postures or improper muscular loads on the prescribed exercises. This is often achieved using inertial and surface electromyography (sEMG) sensors, as they can be used to monitor human motion variables and muscular activation. In this work, we propose to use machine learning techniques to automatically assess if a given exercise was properly executed. We present two major contributions: (1) a novel sEMG segmentation algorithm based on a syntactic approach and (2) a feature extraction and classification pipeline. The proposed methodology was applied to a controlled laboratory trial, for a set of 3 different exercises often prescribe by physiotherapists. The findings of this study support it is possible to automatically segment and classify exercise repetitions according to a given set of common deviations.

Short Papers
Paper Nr: 1
Title:

An Investigation of Multi-Language Age Classification from Voice

Authors:

Osman Büyük and Levent M. Arslan

Abstract: In this paper, we investigate the use of deep neural networks (DNN) for a multi-language age classification task using speaker’s voice. For this purpose, speech databases in two different languages are combined together to construct a multi-language database. Mel-frequency cepstral coefficients (MFCC) are extracted for each utterance. A Gaussian mixture model (GMM), a support vector machine (SVM) and a feed-forward deep neural network (DNN) systems are trained using the features. In the SVM and DNN methods, the GMM means are concatenated to obtain a GMM supervector. The supervectors are fed into the SVM and DNN for age classification. In the experiments, we observe that the multi-language training does not degrade the performance in the SVM and DNN methods when compared to the matched training where train and test languages are the same. On the other hand, the performance is degraded for the traditional GMM method. Additionally, the SVM and DNN significantly outperform the GMM in the multi-language train-test scenario. The absolute performance improvement with the SVM and DNN is approximately 12% and 7% for female and male speakers, respectively.

Paper Nr: 5
Title:

Measuring Upper-Extremity Use with One IMU

Authors:

Hang Wang, Mohamed M. Refai and Bert-Jan F. van Beijnum

Abstract: Discharge home from hospital can be a critical stage in the rehabilitation of patients with central neurological disorders such as stroke. The new skills and early recovery achieved in the hospital may be difficult to transfer to the home environment. This work addresses the monitoring of arm usage and proposed a new metric called Weighted Activity Counts (WAC) based on a sensing system that consists of only one inertial measurement unit (IMU). The proposed metric combines activity counts and the smoothness of the movement. This work defines Normalized Gross Energy Expenditure (NGEC) as the reference metric. WAC shows good performance under the validation protocol we designed (correlation coefficient r > 0.90). The optimal placement for the single sensor which can sufficiently and reliably describe arm usage is also explored in this work.

Paper Nr: 7
Title:

Differences between Mental and Physical Preparation of Muscular Contraction: A Pilot Study

Authors:

Yosra Saidane, Sofia Ben Jebara, Tarak Driss and Giovanni de Marco

Abstract: This paper studies some differences between mental and physical preparation of muscular contraction from a signal processing point of view. Mental preparation is a cognitive process prior to performance while physical preparation is a bodily movement produced by skeletal muscles. Two features are selected. The first indicator, called Mean Normalized Preparation Power (MNPP), represents the amount of muscular activity produced during preparation. The second feature, called Normalized Mutual Information (NMI), studies the functional connectivity between a pair of agonist/antagonist muscles entering in action. Results showed that connectivity is more important in Mental Preparation Activity (MPA) than in Physical Preparation Activity (PPA) while muscles power is more important in PPA than in MPA. When classifying the preparation according to muscular activity importance (low and large preparation), previous conclusions are valid for large preparation. In case of small preparation, there is no differences between MPA and PPA in term of MNPP while MPA has higher NMI. Finally, the study of the correlation between MNPP and NMI showed a moderate dependence with the agonist muscle and an independence with the antagonist muscle.

Paper Nr: 8
Title:

RF Pulses Modelization for EMG Signal Denoising in fRMI Environment

Authors:

Sofia Ben Jebara

Abstract: This paper deals with noise contaminating EMG signal acquired in fRMI environment. The RF pulses are particularly addressed. Their characterization in the frequency domain allows their presentation as discrete pulses repeated at frequencies multiple of RF pulses repetition. The Harmonic plus Noise Model (HNM) is then used to model these pulses in the time-domain. The parameters of the model are extracted, frame by frame, according to the principle of short time analysis. The model is validated according to two criteria: the Segmental Signal to Noise Ratio (SSNR) and its Normalized Standard Deviation (NSDSSNR). Once modeled, the estimated noise is subtracted from noisy observation of EMG signal, leading to an enhanced version. Simulation results are given, validating the approach. In absence of ground truth, realistic situations are simulated in order to calculate quantitative criteria. Furthermore, qualitative appreciation is given thanks to muscular contractions profiles. Finally, the results are compared to those obtained with spectral subtraction and comb filtering.

Paper Nr: 16
Title:

Lagged Transfer Entropy Analysis to Investigate Cardiorespiratory Regulation in Newborns during Sleep

Authors:

Nicolò Pini, Maristella Lucchini, William P. Fifer, Nina Burtchen and Maria G. Signorini

Abstract: The autonomic nervous system (ANS) acts modulating the cardiac and respiratory systems by means of the sympathetic and parasympathetic branches. In this work, we propose to employ Transfer Entropy (TE) with the aim of disambiguating the contributions of the two branches over cardiorespiratory regulation in newborns during sleep. Specifically, we computed TE on the original time series representative of the two subsystems, namely Heart Rate Variability (HRV) and Respiration (RESP). Furthermore, we employed a lagged version of the two original signals to derive a TE estimation capable of providing and insight on the short-term memory between the two systems. Results show the information transfer quantified by TERESP→RR decaying rapidly as the shift between the two time series increases. On the other hand, TERR→RESP exhibits a slower but prolonged interaction, which lasts over numerous lags. The novel approach presented in this work affords the potential to assess infants’ ANS development in terms of the quantification of cardiorespiratory control functioning.

Paper Nr: 17
Title:

The Stress Relief Effects of Foot Warming during Mental Workload

Authors:

Masahiro Inazawa, Yuki Ban, Rui Fukui and Shin’ichi Warisawa

Abstract: Stress has become a social problem in recent years, and stress control plays a key role in daily life. Researchers have studied methods of stress detection and spontaneous stress relief such as listening to soothing music and walking in a forest. However, some people are unable to take spontaneous breaks; therefore, the development of a means of taking “nonexplicit breaks,” to relieve stress unconsciously while working, is required. In this study, we proposed a stress-relief method that did not disturb working. To relieve stress, individuals warm their feet while working, because their hands and feet often become cold when they are under stress. We examined the effect of stress relief via foot warming, and the results revealed that warming the feet caused an increase in RR intervals related to relaxation levels.

Paper Nr: 20
Title:

Predicting Response Uncertainty in Online Surveys: A Proof of Concept

Authors:

Maria C. Dias, Catia Cepeda, Dina Rindlisbacher, Edouard Battegay, Marcus Cheetham and Hugo Gamboa

Abstract: Online questionnaire-based research is growing at a fast pace. Mouse-tracking methods provide a potentially important data source for this research by enabling the capture of respondents’ online behaviour while answering questionnaire items. This behaviour can give insight into respondents’ perceptual, cognitive and affective processes. The present work focused on the potential use of mouse movements to indicate uncertainty when answering questionnaire items and used machine learning methods as a basis to model these. N=79 participants completed an online questionnaire while mouse data was tracked. Mouse movement features were extracted and selected for model training and testing. Using logistic regression and k-fold cross-validation, the model achieved an estimated performance accuracy of 89%. The findings show that uncertainty is indicated by an increase in the number of horizontal direction inversions and the distance covered by the mouse and by longer interaction times with and a higher number of revisits to questionnaire items that evoked uncertainty. Future work should validate these methods further.

Paper Nr: 24
Title:

Evaluation of Spatial-Temporal Anomalies in the Analysis of Human Movement

Authors:

Rui Varandas, Duarte Folgado and Hugo Gamboa

Abstract: In industrial contexts, the performed tasks consist of sets of predetermined movements that are continuously repeated. The execution of improper movements and the existence of events that might prejudice the productive system are regarded as anomalies. In this work, it is proposed a framework capable of detecting anomalies in generic repetitive time series, adequate to handle human motion from industrial scenarios. The proposed framework consists of (1) a new unsupervised segmentation algorithm; (2) feature extraction, selection and dimensionality reduction; (3) unsupervised classification based on Density-Based Spatial Clustering Algorithm for applications with Noise. The proposed solution was applied in four different datasets. The yielded results demonstrated that anomaly detection in human motion is possible with an accuracy of 73±19%, specificity of 74 ± 21% and sensitivity of 74 ± 35%, and also that the developed framework is generic and may be applied in general repetitive time series with little adaptation effort for different domains.

Paper Nr: 26
Title:

Ventricular Activity Signal Removal in Atrial Electrograms of Atrial Fibrillation

Authors:

Bahareh Abdi, Richard C. Hendriks, Alle-Jan van der Veen and Natasja M. S. de Groot

Abstract: Diagnosis and treatment of atrial fibrillation can benefit from various signal processing approaches employed on atrial electrograms. However, the performance and interpretation of these approaches get highly degraded by far-field ventricular activities (VAs) that distort the morphology of the pure atrial activities (AAs). In this study, we aim to remove VAs from the recorded unipolar electrogram while preserving the AA components. To do so, we have developed a framework which first removes the VA-containing segments and interpolates the remaining samples. This will also partly remove the atrial components that overlap with VA signals, e.g., during atrial fibrillation. To reconstruct the AA components, we estimate them from the removed VA-containing segments based on a low-rank and sparse matrix decomposition and add them back to the electrograms. The presented framework is of rather low complexity, preserves AA components, and requires only a single EGM recording. Instrumental comparison to template matching and subtraction and independent component analysis shows that the proposed approach leads to smoother results with better similarity to the true atrial signal.

Paper Nr: 31
Title:

An Artificial Neural Network for Hand Movement Classification using Surface Electromyography

Authors:

Paulo L. Viana, Victoria S. Fujii, Larissa M. Lima, Gabriel L. Ouriques, Gustavo C. Oliveira, Renato Varoto and Alberto Cliquet Jr.

Abstract: In this paper we present the development of an artificial neural network that uses surface EMG data from two forearm muscles to classify hand movements and gestures. We trained our network to classify three different sets of movements, using EMG data from six healthy subjects. We were able to achieve hit rates of above 99% in the training sets and hit rates of above 85% in all three test sets, with a maximum of 88.8% for the second movement set. Advantages of the proposed method include small number of electrodes, reduced complexity, computational cost and response time.

Paper Nr: 35
Title:

Automatic Nuclei Detection in Histopathological Images based on Convolutional Neural Networks

Authors:

Roaa A. Alah, Gokhan Bilgin and Abdulkadir Albayrak

Abstract: Analysis of cells in histopathological images with conventional manual methods is relatively expensive and time-consuming work for pathologists. Recently, computer aided and facilitated researches for the diagnostic algorithms have obtained a high significance to assist the pathologists to extract cellular structures. In this paper, we are compering the conventional fuzzy c-means (FCM) clustering method with the proposed automated detection system based on Tiny-Convolutional Neural Network (Tiny-CNN) to detect center of nucleus in histopathological images, Also, in this study, we are tried to find center of nucleus by combined unsupervised method (FCM) with supervised method (Tiny-CNN). Briefly, First step, nuclei centers are detected with FCM algorithm which is applied as a clustering-segmentation method to perform segmentation of nucleus cellular and nucleus non-cellular structure to find the correct center of nuclei. Second step, the deep learning method is used to detect center of nucleus based automated method. Afterward, combined each of these individual methods to evaluate our model for extracting the center of nucleus on two different data set the University of California Santa Barbara’s UCSB-58 data set and data set University of Warwick’s CRC-100 data set.

Paper Nr: 36
Title:

Progress of MRI-guided EP Interventions is Hampered by a Lack of ECG-based Patient Monitoring – An Engineering Perspective

Authors:

Johannes K. Passand and Georg Rose

Abstract: This position paper discusses the current developments and advances of electrophysiological (EP) interventions guided by magnetic resonance imaging (MRI) and the associated technological challenges and difficulties which need to be overcome in the future. MRI provides several advantages compared to other medical imaging modalities. However, performing any kind of intervention or surgery in an MRI scanner is technical challenging. EP procedures are a special case since they involve many sensitive electronic stimulation and measurement devices and also require a high quality patient monitoring. Monitoring the patient’s electrocardiogram (ECG) inside an MRI is a challenging task due to the MRI’s hazardous environment. Hence, ECG signals are highly distorted and are of limited diagnostic value. This limitation in ECG-based patient monitoring and the lack of a fully functional, MRI-conditional 12-lead ECG hampers or delays the progress of EP procedures during MRI. We review and discuss the main reasons for this limitation and give an outlook and recommendation for further research approaches.

Paper Nr: 38
Title:

Cataglyphis Ant Navigation Strategies Solve the Global Localization Problem in Robots with Binary Sensors

Authors:

Nils Rottmann, Ralf Bruder, Achim Schweikard and Elmar Rueckert

Abstract: Low cost robots, such as vacuum cleaners or lawn mowers, employ simplistic and often random navigation policies. Although a large number of sophisticated localization and planning approaches exist, they require additional sensors like LIDAR sensors, cameras or time of flight sensors. In this work, we propose a global localization method biologically inspired by simple insects, such as the ant Cataglyphis that is able to return from distant locations to its nest in the desert without any or with limited perceptual cues. Like in Cataglyphis, the underlying idea of our localization approach is to first compute a pose estimate from pro-prioceptual sensors only, using land navigation, and thereafter refine the estimate through a systematic search in a particle filter that integrates the rare visual feedback. In simulation experiments in multiple environments, we demonstrated that this bioinspired principle can be used to compute accurate pose estimates from binary visual cues only. Such intelligent localization strategies can improve the performance of any robot with limited sensing capabilities such as household robots or toys.

Paper Nr: 42
Title:

Adaptive Method for Detecting Zero-Velocity Regions to Quantify Stride-to-Stride Spatial Gait Parameters using Inertial Sensors

Authors:

Mohamed Boutaayamou, Cédric Schwartz, Laura Joris, Bénédicte Forthomme, Vincent Denoël, Jean-Louis Croisier, Jacques G. Verly, Gaëtan Garraux and Olivier Brüls

Abstract: We present a new adaptive method that robustly detects zero-velocity regions to accurately and precisely quantify (1) individual stride lengths (SLs), (2) individual stride velocities (SVs), (3) the average of SL, (4) the average of SV, and (5) the cadence during slow, normal, and fast overground walking conditions in young and healthy people. The measurements involved in the estimation of these spatial gait parameters are obtained using only one inertial measurement unit attached on a regular shoe at the level of the heel. This adaptive method reduced the integration drifts across consecutive strides and improved the accuracy and precision in the spatial gait parameter estimation. The validation of the proposed algorithm has been carried out using reference spatial gait parameters obtained from a kinematic reference system. The accuracy ± precision results were for SLs: 0.0 ± 4.7 cm, −0.7 ± 4.4 cm, and −5.8 ± 5.8 cm, during slow, normal, and fast walking conditions, respectively, corresponding to −0.1 ± 4.2 %, −0.5 ± 3.2 %, and −3.3 ± 3.0 % of the respective mean SL. The accuracy ± precision results were for SVs: 0.0 ± 2.9 cm/s, −0.7 ± 3.8 cm/s, and −6.7 ± 6.7 cm/s, during slow, normal, and fast walking conditions, respectively, corresponding to −0.6 ± 3.3 %, −0.1 ± 4.5 %, and −3.5 ± 3.1 % of the respective mean SV. These validation results show a good agreement between the proposed method and the reference, and demonstrate a fairly accurate and precise estimation of these spatial gait parameters. The proposed method paves the way for an objective quantification of spatial gait parameters in routine clinical practice.

Paper Nr: 43
Title:

Predicting Respiratory Depression in Neonates using Intra-arterial Pressure Measurements

Authors:

Aleksandar Jeremic and Dejan Nikolic

Abstract: Respiratory problems are one of the most common reasons for neonatal intensive care unit (NICU) admission of newborns. It has been estimated that as much as 29% of late preterm infants develop high respiratory morbidity. To this purpose invasive ventilation is often necessary for their treatment in NICU. These patients usually have underdeveloped respiratory system with deficiencies such as small airway caliber, few collateral airways, compliant chest wall, poor airway stability, and low functional residual capacity. Consequently ventilation control has been subject of considerable research interest. In this paper we propose an algorithm for detection of respiratory depression by predicting the onset of pO2 depressions using intra-arterial pressure measurements and second order statistical properties of these signals. We calculate the average covariance matrix of intra-arterial pressure measurements in the absence of respiratory depression. We then use this matrix as a reference measure and monitor the changes in the actual covariance matrix measurements. We predict the onset of respiratory depression once the distance is larger than empirically determined threshold. We demonstrate the applicability of our results using a real data set.

Paper Nr: 44
Title:

Predicting Functional Recovery of Stroke Patients using Age Dependent Model

Authors:

Aleksandar Jeremic, Milan Savic, Ljubica Nikcevic, Dejan Nikolic and Natasa Kovacevic-Kostic

Abstract: Predicting functional recovery of stroke patients is important from both clinical and academic points of view. From the clinical point of view it is important to patients, families and clinical workers. Most importantly, an accurate prediction enables us to provide more accurate prognoses, set goals, manage therapies and improve management of healthcare resources through optimal discharge procedures. For example, being able to predict recovery of particular limbs we could potentially improve advanced planning of safe transfer in an optimally determined time frame. Functional recovery is usually evaluated using various functional indices that evaluate patients’ ability to perform daily living tasks. In this paper we propose to predict functional recovery using two well established functional indices: functional independence measure and Barthels index. We model those indices as a age dependent polynomial functions with unknown coefficients and estimate the unknown parameters. In order to demonstrate applicability of the propose technique we compare the performance of our non-linear polynomial model with the performance on linear MANOVA model.

Paper Nr: 46
Title:

From a Swarm to a Biological Computer

Authors:

Andrew Schumann

Abstract: According to behaviourism, any swarm behaviour can be managed by outer stimuli: attractants (motivational reinforcement) and repellents (motivational punishment). In the meanwhile, there are the following two main stages in reactions to stimuli: (i) sensing (perceiving signals) and (ii) motoring (appropriate direct reactions to signals). Hence, by placing attractants and repellents at different sites we can manage and program the swarm behaviour. This opportunity allows us to design a biological computer – an abstract machine (i) with inputs presented by stimuli coming from attractants and repellents and (ii) with outputs presented by the swarm reactions to appropriate stimuli. This computer can be realized on different swarms differently. The point is that different matters are attractants and repellents for different animals. They differ a lot even for microorganisms. Nevertheless, their logic and mathematics are the same. Behaviourism means that (i) the complex of swarm behavioural patterns can be reduced to a composition of some elementary swarm patterns, (ii) if we know an appropriate attractant or repellent for each elementary pattern, then from a complex of attractants and repellents we can deduce a complex of patterns. Nevertheless, it can be shown that both assumptions are false. The point is that swarms are populations which behave as a distributed network, capable of responding to a wide range of spatially represented stimuli so that in their behaviours we can observe effects of neural networks with lateral activation and lateral inhibition mechanisms. As a result, behavioral patterns cannot be additive. In the paper it is discussed what we can do with this feature of swarm behaviour to program swarms.

Paper Nr: 47
Title:

Behavioural Data Modeling: A Case Study in IoT

Authors:

Jiri Petnik, Lenka Lhotska, Jaromir Dolezal and Jindrich Adolf

Abstract: Modeling and analysis of behaviour by using data extracted from Internet of Things (IoT) sensors is an open area. We take Behaviour Informatics (BI) as a formal representation into account and describe the case study of the apartment monitored by IoT sensors. The case study targets persons who live home alone (e.g., elderly people) without assistants (nurses), or any roommates. We present the apartment as a directed multigraph and propose the model to deal with the conversion of transactional data coming from IoT sensors into behavioural feature space represented by behavioural vectors. Further, the article describes a few use cases which can occur in the apartment with installed sensors and explains how behavioural vectors are created. Last but not least, we present the high-level overview of the complex system for detection and evaluation of behaviour identified from data of IoT sensors.

Paper Nr: 9
Title:

Glottal Flow Analysis in Parkinsonian Speech

Authors:

Patrick Corcoran, Arnold Hensman and Barry Kirkpatrick

Abstract: Speech and vocal impairments are one of the earliest symptoms of Parkinson’s disease (PD). Laryngoscope examinations have identified that patients with the disease show pathological behaviour of the vocal folds. The behaviour of the vocal folds is investigated by analysing the glottal flow waveform in Parkinsonian speech in this study. This study aims to determine the appropriate method for estimating the glottal source in PD speech and to identify glottal parameters that could be indicative of PD. An experiment was conducted to analyse a selection of glottal parameters (2 time-domain and 3 frequency-domain) measured from the glottal flow waveform estimated from speech recordings. A database of 52 healthy speakers and 44 speakers with Parkinson’s disease was considered for this experiment. Two glottal estimation techniques are considered in the experiment: iterative and adaptive inverse filtering (IAIF) and quasi-closed phase (QCP) inverse filtering. The results showed that 2 of the 5 glottal parameters (1 time domain and 1 frequency domain) produced values indicating a difference between healthy and PD speech files in the database. The results also indicate that glottal estimates from the IAIF method resulted in parameters discriminating between healthy and PD higher than glottal estimates from the QCP method.

Paper Nr: 10
Title:

System Identification Algorithms Applied to Glottal Model Fitting

Authors:

Piotr Barycki, Irene Murtagh and Barry Kirkpatrick

Abstract: This study proposes a new method of fitting a glottal model to the glottal flow estimate using system identification (SI) algorithms. Each period of the glottal estimate is split into open and closed phases and each phase is modelled as the output of a linear filter. This approach allows the parametric model fitting task to be cast as a system identification problem and sidesteps issues encountered with standard glottal parametrisation algorithms. The study compares the performance of two SI methods: Steiglitz-McBride and Prony. The tests were performed on synthetic glottal signals (n=121) and real speech (n=50 healthy, n=23 pathological). The effectiveness of the techniques is quantified by calculating the Normalised Root Mean Squared Error (NRMSE) between the estimated glottal fit and the glottal estimate. Tests on synthetic glottal signals show that the average performance of the Steiglitz-McBride method (97.25%) was better than the Prony method (70.41%). Real speech tests produced results of 64.29% and 51.57% for healthy and pathological speech respectively. The results show that system identification techniques can produce robust parametric model estimates of the glottal waveform and that the Steiglitz-McBride method is superior to the Prony method for this task.

Paper Nr: 12
Title:

Comparative Study of Compression Techniques Applied in Different Biomedical Signals

Authors:

A. A. Saraiva, F. J. Castro, Nator C. Costa, Jose M. Sousa, N. F. Ferreira, Antonio Valente and Salviano Soares

Abstract: Is work aims to compare the compression of electro-oculographic signals, based on the (EOG) from MIT / BIH database, and the electromyographic signals, based on the (EMG) from MIT / BIH database, for that purpose, two compression techniques that can be used in electro-oculograms and electromyograms was approached, the two techniques mentioned above, were, the discrete cosine transform and Fast Walsh Hadamard Transform. For statistic the methods used was, the Mean squared error, mean absolute error, signal-to-noise ratio and peak signal-to-noise ratio as well, and for results, the techniques and they performance on each tested signal.

Paper Nr: 25
Title:

Detailed Human Activity Recognition based on Multiple HMM

Authors:

Mariana Abreu, Marília Barandas, Ricardo Leonardo and Hugo Gamboa

Abstract: A wide array of activities is performed by humans, everyday. In healthcare, precocious detection of movement changes in daily activities and their monitoring, are important contributors to assess the patient general well-being. Several previous studies are successful in activity recognition, but few of them provide a meticulous discrimination. Hereby, we created a novel framework specialized in detailed human activities, where signals from four sensors were used: accelerometer, gyroscope, magnetometer and microphone. A new dataset was created, with 10 complex activities, suchlike opening a door, brushing the teeth and typing on the keyboard. The classifier was based on multiple hidden Markov models, one per activity. The developed solution was evaluated in the offline context, where it achieved an accuracy of 84±4.8%. It also showed a solid performance in other performed tests, where it was tested with different detailed activities, and in simulations of real time recognition. This solution can be applied in elderly monitoring to access their well-being and also in the early detection of degenerative diseases.

Paper Nr: 37
Title:

Can Sit-to-walk Assessment Maximize Instrumented Timed Up & Go Test Output?

Authors:

Slavka Viteckova, Radim Krupicka, Petr Dusek, Patrik Kutilek, Zoltan Szabo and Evžen Růžička

Abstract: Daily human activities commonly include standing from a seated position. In research this transition is investigated, among others, as a part of a functional Timed Up & Go test. Spatio-temporal parameters are widely used to assess the sit-to-walk transition. Usually, the parameters calculated for the sit-to-walk signal is in its entirety. Another approach primarily splits the transition into phases and then calculates parameters for individual phases separately. The objective of this work is to examine whether splitting the Timed Up & Go test into subphases provides additional value for transition assessment. In order to compare both approaches, we utilized angular rate parameters (duration, peak value, mean, variance) and analyzed their reliability. The reliability proved to be dependent on the subject group and transition phase. In addition, we compared transition parameters from the entire transition and individual phases between the two subject groups. The mean only differentiated between the subject groups in individual phases, but not is entire transition. To summarize, splitting the transition into phases turned out to be beneficial for sit-to-walk transition assessment.

Paper Nr: 41
Title:

An IOT based Wearable Smart Glove for Remote Monitoring of Rheumatoid Arthritis Patients

Authors:

M. Raad, M. Deriche, A. B. Hafeedh, H. Almasawa, K. B. Jofan, H. Alsakkaf, A. Bahumran and M. Salem

Abstract: Rheumatoid Arthritis is a disabling and painful disease of finger joints affecting mainly the elderly people and requiring continuous medications and physical therapy. Traditional arthritis measurements require labor intensive examination by clinical staff. These manual measurements are inaccurate and subject to observer variations. Nowadays the use of wearable technologies is spreading and rapidly becoming a trend promising to offer key benefits in the management of chronic diseases especially at home. Here, we propose an affordable Smart Glove instrument for assisting physiotherapists in remotely analysing patients finger flexions when performing diverse activities at home. An E-textile based glove uses flex and force sensors, and an Arduino platform to transmit motion data to the physiotherapists using a smart phone using a dedicated App. The flex sensor on the index finger detects and estimates the motion, a BLE (Bluetooth Low Energy) Nano is used for processing and wireless transmission. The wearable smart glove uses a lithium ion 3.3V rechargeable 400mAH battery for consuming power. This Smart Glove also helps in monitoring the patient’s response to either medication and/or diverse recommended movements. The data collected can be used to analyse the status of the patient with time and also in assisting care givers to change planned activities or exercises when needed.

Paper Nr: 45
Title:

Detecting Neonatal Seizures using Sample Covariance Estimation

Authors:

Aleksandar Jeremic and Dejan Nikolic

Abstract: One of the most frequent of neurological dysfunctions in prematurely born infants is the presence of frequent seizures. As they may be related to serious neurological problems they require immediate detection which is most commonly done using electroencephalography (EEG) systems that enable trained physicians to detect them in the real time. Due to the length of neonatal period (first 28 days) it would be extremely beneficial to have an automated system that is able to detect seizures as it would enable more efficient use of expert time. In this paper we propose a new multichannel technique for detecting seizure in neonates that calculates distance measure using second order statistical properties and Frechet mean. We have demonstrated previously that Frechet mean in certain cases can outperform clustering/detection algorithms that are based on first order distances.