BIOSIGNALS 2012 Abstracts


Full Papers
Paper Nr: 25
Title:

A WEARABLE GAIT ANALYSIS SYSTEM USING INERTIAL SENSORS PART II - Evaluation in a Clinical Setting

Authors:

A. Sant'Anna, N. Wickström, H. Eklund and R. Tranberg

Abstract: The gold standard for gait analysis, in-lab 3D motion capture, is not routinely used for clinical assessment due to limitations in availability, cost and required training. Inexpensive alternatives to quantitative gait analysis are needed to increase the its adoption. Inertial sensors such as accelerometers and gyroscopes are promising tools for the development of wearable gait analysis (WGA) systems. The present study evaluates the use of a WGA system on hip-arthroplasty patients in a real clinical setting. The system provides information about gait symmetry and normality. Results show that the normality measurements are well correlated with various quantitative and qualitative measures of recovery and health status.

Paper Nr: 36
Title:

AN EFFICIENT NUMERICAL RESOLUTION FOR MRI RICIAN DENOISING

Authors:

A. Martín, J. F. Garamendi and E. Schiavi

Abstract: We consider a variational Rician denoising model for Magnetic Resonance Images (MRI) that we solve by a semi-implicit numerical scheme, which leads to the resolution of a sequence of Rudin, Osher and Fatemi (ROF) models. This allows to implement efficient numerical gradient descent schemes based on the dual formulation of the ROF model which are compared with a direct semi-implicit approach for the primal problem recently proposed for model validation. In this new framework the total variation operator is exactly solved as opposed to the approximating problems which must be considered when the primal problem is dealt with. The comparison among the above methods is performed using synthetic and real MR brain images and the results show the effectiveness of the new method in both, the accuracy and the speeding up of the algorithm.

Paper Nr: 37
Title:

TISSUE TYPE DIFFERENTIATION FOR BRAIN GLIOMA USING NON-NEGATIVE MATRIX FACTORIZATION

Authors:

Yuqian Li, Diana M. Sima, Sofie Van Cauter, Uwe Himmelreich, Yiming Pi and Sabine Van Huffel

Abstract: The purpose of this paper is to introduce a hierarchical Non-negative Matrix Factorization (NMF) approach, customized for the problem of blindly separating brain glioma tumor tissue types using short-echo time proton magnetic resonance spectroscopic imaging (1H MRSI) signals. The proposed algorithm consists of two levels of NMF, where two constituent spectra are computed in each level. The first level is able to correctly detect the spectral profile corresponding to the most predominant tissue type, i.e., normal tissue, while the second level is optimized in order to detect two ‘abnormal’ spectral profiles so that the 3 recovered spectral profiles are least correlated with each other. The two-level decomposition is followed by the reestimation of the overall spatial distribution of each tissue type via standard Non-negative Least Square (NNLS). This method is demonstrated on in vivo short-TE 1H MRSI brain data of a glioblastoma multiforme patient and a grade II-III glioma patient. The results show the possibility of differentiating normal tissue, tumor tissue and necrotic tissue in the form of recovered tissue-specific spectra with accurate spatial distributions.

Paper Nr: 53
Title:

SUPPORT VECTOR DATA DESCRIPTION FOR SPOKEN DIGIT RECOGNITION

Authors:

Amirhossein Tavanaei, Alireza Ghasemi, Mohammad Tavanaei, Hossein Sameti and Mohammad T. Manzuri

Abstract: A classifier based on Support Vector Data Description (SVDD) is proposed for spoken digit recognition. We use the Mel Frequency Discrete Wavelet Coefficients (MFDWC) and the Mel Frequency cepstral Coefficients (MFCC) as the feature vectors. The proposed classifier is compared to the HMM and results are promising and we show the HMM and SVDD classifiers have equal accuracy rates. The performance of the proposed features and SVDD classifier with several kernel functions are evaluated and compared in clean and noisy speech. Because of multi resolution and localization of the Wavelet Transform (WT) and using SVDD, experiments on the spoken digit recognition systems based on MFDWC features and SVDD with weighted polynomial kernel function give better results than the other methods.

Paper Nr: 60
Title:

REDUCING THE NUMBER OF CHANNELS AND SIGNAL-FEATURES FOR AN ACCURATE CLASSIFICATION IN AN EMG PATTERN RECOGNITION TASK

Authors:

Iker Mesa, Angel Rubio, Javier Diaz, Jon Legarda and Beatriz Sedano

Abstract: In this work 32 surface Electromyography (sEMG) electrode locations and 41 signal-features are evaluated in order to achieve an accurate classification rate in a static-hand gesture classification task. A novel implementation of the minimal Redundancy Maximal Relevance (mRMR) Variable Selection algorithm is proposed with the aim of selecting the most informative and least redundant combination of sEMG channels and signal features. The performance of the new algorithm and of the selected set of channels and signal-features are tested with a Support Vector Machine classifier.

Paper Nr: 78
Title:

REAL TIME ELECTROCARDIOGRAM SEGMENTATION FOR FINGER BASED ECG BIOMETRICS

Authors:

André Lourenço, Hugo Silva, Paulo Leite, Renato Lourenço and Ana Fred

Abstract: In biometric recognition based on Electrocardiographic (ECG) signals, there are two main approaches for feature extraction: fiducial and non-fiducial. Fiducial methods use points of interest within single heartbeat waveforms, obtained by segmenting the ECG signal using QRS complexes as a reference. In this paper we study several QRS detection algorithms, with the purpose of determining what is the best algorithm in the context of finger based ECG biometrics using fiducial approaches; our main focus is the real-time segmentation of ECG signals resulting on a set of single heart beats. We propose a method combining the adaptive characteristics of the algorithm by Christov, with the strategy of the widely adopted Engelse and Zeelenberg algorithm. Experimental results obtained for real-world data show that online approaches are competitive with offline versions, and represent a contribution for the realization of real-time biometric recognition.

Paper Nr: 83
Title:

IMPACT OF WINDOW LENGTH AND DECORRELATION STEP ON ICA ALGORITHMS FOR EEG BLIND SOURCE SEPARATION

Authors:

Gundars Korats, Steven Le Cam and Radu Ranta

Abstract: Blind Source Separation (BSS) approaches for multi-channel EEG processing are popular, and in particular Independant Component Analysis (ICA) algorithms have proven their ability for artifacts removal and source extraction for this very specific class of signals. However, the blind aspect of these techniques implies well-known drawbacks. As these methods are based on estimated statistics from the data and rely on an hypothesis of signal stationarity, the length of the window is crucial and has to be chosen carefully: large enough to get reliable estimation and short enough to respect the rather non-stationary nature of the EEG signals. In addition, another issue concerns the plausibility of the resulting separated sources. Indeed, some authors suggested that ICA algorithms give more physiologically plausible results depending on the chosen whitening/sphering step. In this paper, we address both issues by comparing three popular ICA algorithms (namely FastICA, Extended InfoMax and JADER) on EEG-like simulated data and assessing their performance by using an original correlation matrices distance measure and a separation performance index. The results are consistent and lead us to a precise idea of minimal sample size that guarantees statistically robust results regarding the number of channels.

Paper Nr: 85
Title:

HIERARCHICAL DYNAMIC MODEL FOR HUMAN DAILY ACTIVITY RECOGNITION

Authors:

Blanca Florentino-Liaño, Niamh O'Mahony and Antonio Artés-Rodríguez

Abstract: This work deals with the task of human daily activity recognition using miniature inertial sensors. The proposed method is based on the development of a hierarchical dynamic model, incorporating both inter-activity and intra-activity dynamics, thereby exploiting the inherently dynamic nature of the problem to aid the classification task. The method uses raw acceleration and angular velocity signals, directly recorded by inertial sensors, bypassing commonly used feature extraction and selection techniques and, thus, keeping all information regarding the dynamics of the signals. Classification results show a competitive performance compared to state-of-the-art methods.

Paper Nr: 86
Title:

INFERENCE ABOUT MULTIPLE PATHWAYS IN MOTOR CONTROL LIMB IN LOCUST

Authors:

C. D. Maciel, D. M. Simpson and P. L. Newland

Abstract: In locust local circuits that control limb movements, the neural signals are processed by both spiking and nonspiking interneurons that operate in parallel to process sensory information. These interneurons receive sensory inputs from leg mechanoreceptors and together project to leg motor neuron pools. The main feature of the nonspiking interneurons is their ability to communicate with other neurons without the intervention of nerve impulses, or spikes, so that they exert graded control over their postsynaptic motor neurons, while spiking local interneurons communicate by means of action potentials and are involved in the integration of sensory signals. Our work presents an investigation from different classes of neurons driven by random Gaussian excitatory movements to a proprioceptor at the knee joint. The underlying aim of this work was to use information theory in understanding connectivity in the neural network.

Paper Nr: 87
Title:

AN EFFICIENT STOCHASTIC BASED MODEL FOR SIMULATING MICROELECTRODE RECORDINGS OF THE DEEP BRAIN - Modelling and Analysis

Authors:

K. J. Weegink, J. J. Varghese, P. A. Bellette, T. Coyne, P. A. Silburn and P. A. Meehan

Abstract: We have developed a computationally efficient stochastic model for simulating microelectrode recordings, including electronic noise and neuronal noise from the local field of 3000 neurons. From this we have shown that for a neuron network model spiking with a stationary Weibull distribution the power spectrum can change from exhibiting periodic behaviour to non-stationary behaviour as the distribution shape is changed. It is shown that the windowed power spectrum of the model follows an analytical result prediction in the range of 100-5000 Hz. The analysis of the simulation is compared to the analysis of real patient interoperative sub-thalamic nucleus microelectrode recordings. The model runs approximately 200 times faster compared to existing models that can reproduce power spectral behaviour. The results indicate that a spectrogram of the real patient recordings can exhibit non-stationary behaviour that can be re-created using this efficient model in real time.

Paper Nr: 90
Title:

DYNAMIC AUTOREGRESSIVE MODELLING OF CRITICAL CARE PATIENTS AS A BASIS FOR HEALTH MONITORING

Authors:

K. Van Loon, G. Meyfroidt, T. Tambuyzer, G. Van den Berghe, D. Berckmans and J.-M. Aerts

Abstract: Real-time modelling techniques could be valuable to continuously evaluate individual critically ill patients and to help the medical staff with estimation of prognosis. This preliminary study examines the possibilities to distinguish survivors from non-survivors on the basis of instabilities in the dynamics of daily measured variables. A data set, containing 140 patients, was generated in the intensive care unit (ICU) of the university hospital of Leuven. First and second order dynamic auto-regression (DAR) models were used to quantify the stability of time series of three physiological variables as a criterion to distinguish survivors from non-survivors. The best results were found for blood urea concentration with true negative fractions of 45/72 (63%) and true positive fractions of 43/68 (62%). The results indicate that the dynamics of time series of laboratory parameters from critically ill patients are indicative for their clinical condition and outcome.

Paper Nr: 92
Title:

TOWARDS A SILENT SPEECH INTERFACE FOR PORTUGUESE - Surface Electromyography and the Nasality Challenge

Authors:

João Freitas, António Teixeira and Miguel Sales Dias

Abstract: A Silent Speech Interface (SSI) aims at performing Automatic Speech Recognition (ASR) in the absence of an intelligible acoustic signal. It can be used as a human-computer interaction modality in high-background-noise environments, such as living rooms, or in aiding speech-impaired individuals, increasing in prevalence with ageing. If this interaction modality is made available for users own native language, with adequate performance, and since it does not rely on acoustic information, it will be less susceptible to problems related to environmental noise, privacy, information disclosure and exclusion of speech impaired persons. To contribute to the existence of this promising modality for Portuguese, for which no SSI implementation is known, we are exploring and evaluating the potential of state-of-the-art approaches. One of the major challenges we face in SSI for European Portuguese is recognition of nasality, a core characteristic of this language Phonetics and Phonology. In this paper a silent speech recognition experiment based on Surface Electromyography is presented. Results confirmed recognition problems between minimal pairs of words that only differ on nasality of one of the phones, causing 50\% of the total error and evidencing accuracy performance degradation, which correlates well with the exiting knowledge.

Paper Nr: 95
Title:

DECISION-TREE BASED ANALYSIS OF SPEAKING MODE DISCREPANCIES IN EMG-BASED SPEECH RECOGNITION

Authors:

Michael Wand, Matthias Janke and Tanja Schultz

Abstract: This study is concerned with the impact of speaking mode variabilities on speech recognition by surface electromyography (EMG). In EMG-based speech recognition, we capture the electric potentials of the human articulatory muscles by surface electrodes, so that the resulting signal can be used for speech processing. This enables the user to communicate silently, without uttering any sound. Previous studies have shown that the processing of silent speech creates a new challenge, namely that EMG signals of audible and silent speech are quite distinct. In this study we consider EMG signals of three speaking modes: audibly spoken speech, whispered speech, and silently mouthed speech. We present an approach to quantify the differences between these speaking modes by means of phonetic decision trees and show that this measure correlates highly with differences in the performance of a recognizer on the different speaking modes. We furthermore reinvestigate the spectral mapping algorithm, which reduces the discrepancy between different speaking modes, and give an evaluation of its effectiveness.

Paper Nr: 96
Title:

GRADIENT ARTEFACT CORRECTION IN THE EEG SIGNAL RECORDED WITHIN THE fMRI SCANNER

Authors:

José L. Ferreira, Pierre J. M. Cluitmans and Ronald M. Aarts

Abstract: In recent years, combined EEG-fMRI has become a powerful brain imaging technique which is largely employed in clinical and neuroscience research. Parallel to the achievements reached in this area, a number of challenges remain to be overcome in order to consolidate such technique as an independent and effective method for brain imaging. In particular, the occurrence of gradient artefacts in the EEG signal due to the magnetic field of the fMRI magnetic scanner. This paper presents a proposal for modelling the variability of the gradient artefact template which makes use of the standard deviation and the slope differentiator between consecutive samples of the signals. Combination of such a model with the average artefact subtraction method achieves a reasonable elimination of the gradient artefact from EEG recordings.

Paper Nr: 102
Title:

BAYESIAN-BASED EARLY DETECTION OF COGNITIVE IMPAIRMENT IN ELDERLY USING fNIRS SIGNALS DURING COGNITIVE TESTS

Authors:

Shohei Kato, Hidetoshi Endo and Yuta Suzuki

Abstract: This paper presents a new trial approach to early detection of dementia in the elderly with the use of functional brain imaging during cognitive tests. We have developed a non-invasive screening system of the elderly with cognitive impairment. In addition of our previous research of speech-prosody based data-mining approach, we had started the measurement of functional brain imaging for patient having a cognitive test by using functional near-infrared spectroscopy (fNIRS). We had collected 42 CHs fNIRS signals on frontal and right and left temporal areas from 50 elderly participants (18 males and 32 females between ages of 64 to 92) during cognitive tests in a specialized medical institute. We propose a Bayesian classifier, which can discriminate among elderly individuals with three clinical groups: normal cognitive abilities (NL), patients with mild cognitive impairment (MCI), and Alzheimer’s disease (AD). The Bayesian classifier has two phases on the assumption of screening process, that firstly checks whether a suspicion of the cognitive impairment (CI) or not (NL) from given fNIRS signals; if any, and then secondly judges the degree of the impairment: MCI or AD. This paper also reports the examination of the detection performance by cross-validation, and discusses the effectiveness of this study for early detection of cognitive impairment in elderly subjects. Consequently, empirical results that both the accuracy rate of AD and the predictive value of NL are equal to or more than 90%. This suggests that proposed approach is adequate practical to screen the elderly with cognitive impairment.

Paper Nr: 103
Title:

CHARACTERIZATION OF THE ENCAPSULATION PROCESS OF DEEP BRAIN STIMULATION ELECTRODES USING IMPEDANCE SPECTROSCOPY IN A RODENT MODEL

Authors:

K. Badstübner, T. Kröger, E. Mix, U. Gimsa, R. Benecke and J. Gimsa

Abstract: Deep brain stimulation (DBS) is effective for the treatment of patients with Parkinson’s disease (PD), especially in advanced stages which are refractory to conventional therapy. Despite of the regular use in clinical therapy, rodent models for basic research into DBS are not routinely available. The main reason is the geometry difference from rodents to humans, imposing larger problems in the transfer of the stimulation conditions than from primates to humans. For rodents, the development of miniaturized mobile stimulators and stimulation parameters, as well as improved electrode materials and geometry are desirable. The impedance of custom made, cylindrical (contact diameter 200 µm, length 100 µm), platinum/iridium electrodes has been measured in vivo for two weeks to characterize the influence of electrochemical processes and of the adherent cell growth at the electrode surface. During the encapsulation process, the real part of the electrode impedance at 10 kHz doubled with respect to its initial value after a characteristic decrease by approximately one third at the second day. An outlook is given on further investigations with different electrode designs for long-term DBS.

Paper Nr: 139
Title:

SPEECH EMOTIONAL FEATURES MEASURED BY POWER-LAW DISTRIBUTION BASED ON ELECTROGLOTTOGRAPHY

Authors:

Lijiang Chen, Xia Mao, Yuli Xue and Mitsuru Ishizuka

Abstract: This study was designed to introduce a kind of novel speech emotional features extracted from Electroglottography (EGG). These features were obtained from the power-law distribution coefficient (PLDC) of fundamental frequency (F0) and duration parameters. First, the segments of silence, voiced and unvoiced (SUV) were distinguished by combining the EGG and speech information. Second, the F0 of voiced segment and the first-order differential of F0 was obtained by a cepstrum method. Third, PLDC of voiced segment as well as the pitch rise and pitch down duration were calculated. Simulation results show that the proposed features are closely connected with emotions. Experiments based on Support Vector Machine (SVM) are carried out. The results show that proposed features are better than those commonly used in the case of speaker independent emotion recognition.

Paper Nr: 141
Title:

NOISE ROBUST SPEAKER VERIFICATION BASED ON THE MFCC AND pH FEATURES FUSION AND MULTICONDITION TRAINING

Authors:

L. Zão and R. Coelho

Abstract: This paper investigates the fusion of Mel-frequency cepstral coefficients (MFCC) and pH features, combined with the multicondition training (MT) technique based on artificial colored spectra noises, for noise robust speaker verification. The a-integrated Gaussian mixture models (a-GMM), an extension of the conventional GMM, are used in the speaker verification experiments. Five real acoustic noises are used to corrupt the speech signals in different signal-to-noise ratios (SNR) for tests. The experiments results show that the use of MFCC + pH feature vectors improves the accuracy of speaker verification systems based on single MFCC. It is also shown that the speaker verification system with the MFCC + pH fusion and the a-GMM with the MT technique achieves the best performance for the speaker verification task in noisy environments.

Short Papers
Paper Nr: 7
Title:

ELECTROCARDIOGRAM DERIVED RESPIRATION USING AN EVOLUTIONARY ALGORITHM

Authors:

Jakob Juul Larsen, Lars G. Johansen and Henrik Karstoft

Abstract: In this work we present a method to extract the respiratory signal from single lead ECG measurements, electrocardiogram derived respiration (EDR). The method is based on adaptive ECG modeling and respiratory signal estimation using an evolutionary algorithm fed with the model parameters. The evolutionary algorithm, which is allowed to employ a large constellation of functions, comes up with a set of relatively simple expressions (3-4 terms) describing valid relationships between ECG model parameters and the respiratory signal. In fact, the expressions mainly turn out to be linear combinations of the model parameters. Our preliminary experiments indicate that this method yields a robust EDR, and that this EDR correlates very well with a reference respiratory signal measurement. Correlation coefficients for the derived expressions lie around 0.95.

Paper Nr: 9
Title:

A PERFECTLY INVERTIBLE RANK ORDER CODER

Authors:

Khaled Masmoudi, Marc Antonini and Pierre Kornprobst

Abstract: Our goal is to revisit rank order coding by proposing an original exact decoding procedure for it. Rank order coding was proposed by Simon Thorpe et al. to explain the impressive performance of the human visual system in recognizing objects. It is based on the hypothesis that the retina represents the visual stimulus by the order in which its cells are activated. A classical rank order coder/decoder was then designed (Van Rullen and Thorpe, 2001) involving three stages: (i) A model of the stimulus transform in the retina consisting in a redundant filter bank analysis; (ii) A sorting stage of the filters according to their activation degree; (iii) A straightforward decoding procedure that consists in a weighted sum of the most activated filters. Focusing on this last stage, it appears that the decoding procedure employed yields reconstruction errors that limit the model Rate/Quality performances when used as an image codec. Attempts made in the literature to overcome this issue are time consuming and alter the coding procedure or are infeasible for standard size images and lacking mathematical support. Here we solve this problem in an original fashion by using the frames theory, where a frame of a vector space designates an extension for the notion of basis. Our contribution is threefold. First we add an adequate scaling function to the filter bank under study that has both a mathematical and a biological justification. Second, we prove that the analyzing filter bank considered is a frame, and then we define the corresponding dual frame that is necessary for the exact image reconstruction. Finally, to deal with the problem of memory overhead, we design an original recursive out-of-core blockwise algorithm for the computation of this dual frame. Our work provides a mathematical formalism for the retinal model under study and specifies a simple and exact reverse transform for it. Furthermore, the framework presented here can be extended to several models of the visual cortical areas using redundant representations.

Paper Nr: 14
Title:

PHASE-RECTIFIED SIGNAL AVERAGING FOR THE QUANTIFICATION OF THE INFLUENCE OF PRENATAL ANXIETY ON HEART RATE VARIABILITY OF BABIES

Authors:

Hannelore Eykens, Devy Widjaja, Katrien Vanderperren, Joachim Taelman, Marijke A. K. A. Braeken, Renée A. Otte, Bea R. A. Van den Bergh and Sabine Van Huffel

Abstract: The autonomic nervous system (ANS) modulates heartbeat intervals responding to inputs from its different branches, resulting in periodicities that occur on different time scales. Internal and external perturbations are continuously interrupting the periodic behavior, making the heartbeat intervals quasi-periodic. Phase-rectified signal averaging (PRSA) is a technique to detect those quasi-periodicities in noisy, non-stationary signals, like tachograms. The method compresses the tachogram in shorter curves based on internal information, and provides information on the deceleration and acceleration capacity of the heart. In this study, the PRSA technique is investigated as a novel signal processing technique for the analysis of heart rate variability (HRV) of babies. In this way, the effect of stress and anxiety during pregnancy on the ANS of the baby is analyzed. First, the PRSA curves are obtained for each baby and different measures that characterize these curves are defined. Next, these measures are linked to the anxiety level of their mothers during pregnancy. Only little influence of the anxiety level of the mother on the HRV of the baby is found.

Paper Nr: 15
Title:

RELIABILITY OF STATISTICAL FEATURES DESCRIBING NEURAL SPIKE TRAINS IN THE PRESENCE OF CLASSIFICATION ERRORS

Authors:

Ninah Koolen, Ivan Gligorijevic and Sabine Van Huffel

Abstract: In order to investigate functioning of the brain processes, it is important to have reliable processing of neural activity. For precise tracking of local neural network processes, reliable clustering of single neurons’ action potentials (spikes) is necessary. So far, it was common to keep the signals of high quality and discard the others. This work examines the possibility of extracting reliable information from bad quality signals, in the presence of spike classification errors. We tested the robustness and information capacity of several statistical parameters used to describe firing patterns of spike trains using simulated signals mimicking most common cases in nature. Although complete reconstruction of firing patterns is not always possible, we show that the approximation of the mean firing frequency as well as the detection of bursting processes can still be quantified successfully, thereby paving the way for future applications.

Paper Nr: 22
Title:

SINGLE TRAIL P300 CLASSIFICATION VIA PROBABILISTIC FUZZY CLASSIFIER AND GENETIC ALGORITHM

Authors:

Amirhossein Jafarian, Mohammadhassan Moradi, Vahid Abootalebi and Mostafa Jafarian

Abstract: P300 is an endogenous brain response to meaningful stimuli in oddball paradigm. Here the aim is to estimate whether this component exists in the recorded electroencephalogram (EEG) segment. A Probabilistic Fuzzy Classifier (PFC) followed by Genetic Algorithm (GA) has been developed in this paper. The main motivation of using PFC is that it not merely has the advantages of fuzzy systems, but also can exploit the stochastic properties of the underlying data. Moreover, by selecting the best set of time-frequency features utilizing GA the classification accuracy is enhanced. A comparison between the performance of the classifier and those based on stochastic properties of the data, like LDA (Linear Discriminate Analysis) and conventional fuzzy classifier verifies the superior performance of using this system.

Paper Nr: 24
Title:

A WEARABLE GAIT ANALYSIS SYSTEM USING INERTIAL SENSORS PART I - Evaluation of Measures of Gait Symmetry and Normality against 3D Kinematic Data

Authors:

A. Sant'Anna, N. Wickstrom, R. Zügner and R. Tranberg

Abstract: Gait analysis (GA) is an important tool in the assessment of several physical and cognitive conditions. The lack of simple and economically viable quantitative GA systems has hindered the routine clinical use of GA in many areas. As a result, patients may be receiving sub-optimal treatment. The present study introduces and evaluates measures of gait symmetry and gait normality calculated from inertial sensor data. These indices support the creation of mobile, cheap and easy to use quantitative GA systems. The proposed method was compared to measures of symmetry and normality derived from 3D kinematic data. Results show that the proposed method is well correlated to the kinematic analysis in both symmetry (r=0.84, p<0.0001) and normality (r=0.81, p<0.0001). In addition, the proposed indices can be used to classify normal from abnormal gait.

Paper Nr: 26
Title:

DEVELOPMENT OF A SIT-TO-STAND ASSISTANCE SYSTEM FOR PARKINSON'S DISEASE SUFFERERS - (Intellectual Handrail)

Authors:

Osamu Nitta, Yasunari Fujimoto, John Surya and Yoshiyuki Takahashi

Abstract: A Sit-to-Stand Assistance System that can provide functional assistance in standing was developed. Two 650 mm stroke AC servo motor driven linear actuators were squarely combined. The handrail was installed at the intersection of these actuators. When the user stands from a chair, the handrail leads the user’s motion. A personal computer (PC) is used to control the handrail motion, and force plates are placed under the feet. In this experiment the subjects were Parkinson's disease (PD) sufferers. Subjects were not able to stand up with a fixed handrail; however, they were able to stand up using this system without help.

Paper Nr: 27
Title:

WHY COLOR CONSTANCY IMPROVES FOR MOVING OBJECTS

Authors:

Marc Ebner

Abstract: Light which is measured by retinal receptors varies with the illuminant. However, a human observer is able to discount the illuminant and to accurately determine the color of objects. The human brain computes a color constant descriptor which is approximately independent of the illuminant. This ability is called color constancy. Recently, it has been shown that color constancy improves for a moving stimulus. It has been argued that high level motion areas may have an influence on the computation of a color constant descriptor. We have developed a computational model for color perception which can be mapped to the different stages of the human visual system. We test our model with two types of stimuli: stationary and moving. In our model, color constancy is computed purely bottom up. Our model also shows better color constancy for a moving stimulus. This indicates that an influence from high level motion areas is not required.

Paper Nr: 28
Title:

KINEMATIC FEATURES OF REACH AND GRASP MOVEMENTS IN STROKE REHABILITATION USING ACCELEROMETERS

Authors:

Julien Stamatakis, Adriana Gonzalez, Benoit Caby, Stephanie Lefebvre, Yves Vandermeeren and Benoit Macq

Abstract: Rehabilitation is an essential process to recover impaired motor functions after stroke. Typically, visual marker-based systems such as the Codamotion are used, as kinematic analyses seem to be an excellent tool to quantify objectively the effects of rehabilitation processes. However, this solution remains expensive. A low-cost accelerometer-based system has been developed and its performances were compared to those of the Codamotion system, used as a gold standard. Thanks to a model for prediction and an error model Kalman filter, the recorded signals were broken up into gravity and dynamic accelerations components that were placed in a global frame and compared to the Codamotion signals. The vertical z-axis was well reconstructed and used as a basis for kinematic analyses. Different features expressing movement speed, control strategy or movement smoothness have been computed from both systems and compared. Despite the fact that some of them showed differences between both systems, the accelerometer-based system computed features with a discriminant power comparable to the ones derived from the Codamotion. In conclusion, this accelerometer-based system is a low-cost alternative to expensive visual marker-based systems that could be extensively used for rehabilitation processes in routine clinical practice or even at home.

Paper Nr: 32
Title:

AN OPEN SOURCE TOOL FOR HEART RATE VARIABILITY WAVELET-BASED SPECTRAL ANALYSIS

Authors:

Constantino A. García, Abraham Otero, Xosé Vila and Maria J. Lado

Abstract: Heart rate variability (HRV) power spectrum analysis is a well-known technique used to study the activity of the autonomic nervous system. It is performed by calculating the spectral power of certain bands of the RR time series. There are several tools that perform this type of analysis: Kubios HRV, PhysioNet’s HRV toolkit for MatLab and aHRV, among others. All these tools use the Short Fourier Transform to estimate spectral power. However, the RR time series is a non-stationary signal. The Wavelet transform is often a more suitable tool for analyzing non-stationary signals than the Short Time Fourier Transform. Its usefulness in HRV analysis has already been proven in the literature. However, the lack of HRV analysis tools that support it has made this technique underutilized in HRV studies. In this paper we present an extension to the RHRV opensource package that enablesWavelet-based HRV spectral analysis. Until now this package only supported HRV spectral analysis based on the Fourier transform.

Paper Nr: 38
Title:

SIMULATION STUDY OF TISSUE TYPE DIFFERENTIATION USING NON-NEGATIVE MATRIX FACTORIZATION

Authors:

Yuqian Li, Diana M. Sima, Sofie Van Cauter, Uwe Himmelreich, Yiming Pi and Sabine Van Huffel

Abstract: Finding the brain tumor tissue-specific magnetic resonance spectra and their corresponding spatial distribution is a typical Blind Source Separation (BSS) problem. Non-negative Matrix Factorization (NMF), which only requires non-negativity constraints, has become popular because of its advantages compared to other BSS methods. A variety of algorithms based on traditional NMF have been recently proposed. This study focuses on the performance comparison of several NMF implementations, including some newly released methods, in brain glioma tissue differentiation using simulated magnetic resonance spectroscopic imaging (MRSI) signals. Experimental results demonstrate the possibility of finding typical tissue types and their distributions using NMF algorithms. The (accelerated) hierarchical alternating least squares algorithm was found to be the most accurate.

Paper Nr: 62
Title:

CLASSIFYING EVENT RELATED POTENTIALS FOR VALID AND PARADOX REASONING

Authors:

Solomon Zannos, Fotios Giannopoulos, Dimitrios Arabadjis, Panayiotis Rousopoulos, Panos Papageorgiou, Elias Koukoutsis, Constantin Papaodysseus and Charalabos Papageorgiou

Abstract: In this paper, a new methodology is presented for comparing the ERPs of Aristotle's "valid reasoning" and Zeno's "paradoxes". To achieve that, the ERPs of each such syllogism are grouped, by means of a new care-fitting approach. This consists of a) application of time-domain and amplitude scaling to one ERP and b) optimal fit of two ERPs via minimization of a properly defined error function. Next, the optimally fit ERPs, which form a group, are averaged to obtain an ideal representative for the valid and paradoxes reasoning separately. These ideal representatives manifest essential statistical differences per subject for a considerable number of electrodes (18 electrodes). The latter supports the assumption that the underlying mental processes of the valid and paradoxes reasoning are, indeed, different and this difference reflects upon the corresponding ERPs and, in particular, upon the introduced ideal representatives.

Paper Nr: 64
Title:

KINECT AND SHIMMER SENSORS IN MOTION ANALYSIS IN HEALTH APPLICATIONS

Authors:

Katja Orlowski, Harald Loose, Karen Otte, Sebastian Mansow-Model and Angelina Thiers

Abstract: Motion capture systems based on different physical principles and sensor elements become moderate in price and mobile in application. New application fields have been developed. Motion capture and analysis is used in physiotherapy, geriatrics or public sports. In this paper three different types of motion capture systems are investigated with respect to their adaptability to health applications: the OptiTrack-Motion Capture System, the Kinect of Microsoft and SHIMMER sensors. Human gait was measured using all three systems and the obtained data were compared. In further tests the gait of volunteers and patients were captured with Kinect and two scores evaluating their motion capability were calculated.

Paper Nr: 73
Title:

NONINVASIVE CARDIOVASCULAR SYSTEM IDENTIFICATION USING PULSE WAVE TRANSIT TIME

Authors:

Sérgio Okida, Pedro Giassi Júnior, João Fernando Refosco Baggio, Raimes Moraes, Maurício Gonçalves de Oliveira and Gastão Fernandes Duval Neto

Abstract: This work shows that it is possible to model the heart rate autonomic control from samples of ECG, PPG and respiratory flow waveform (RFW). Usually, such modelling is carried out with physiological signals that are more difficult to acquire during the clinical exams: ECG, arterial blood pressure and instantaneous lung volume. In this work, the ECG, PPG and RFW were recorded with a portable system from volunteers at two different postures: supine and standing. The ECG, PPG and RFW were processed off line in order to obtain the RR, the inverse of the pulse wave transit time (IPWTT) and the RFW series. These series were used as input for ARMA models and the obtained results were compared to the ones available in the literature. The qualitative and quantitative comparisons of the results reveal very similar performance.

Paper Nr: 74
Title:

PATHOLOGICAL VOICE DETECTION USING TURBULENT SPEECH SEGMENTS

Authors:

Fernando Perdigão, Cláudio Neves and Luís Sá

Abstract: Identification of voice pathologies using only the voice signal has a great advantage over the conventional methods, such as laryngoscopy, since they enable a non-invasive diagnosis. The first studies in this area were based on the analysis of sustained vowel sounds. More recently, there are studies that extend the analysis to continuous speech, achieving similar or better results. All these studies use of a pitch detector algorithm to select only the voiced parts of the acoustic signal. However, the existence of a pathology affecting the speaker’s vocal folds produces a more irregular vibration pattern and, consequently, a degradation of the voice quality with less voiced segments. Thus, by selecting only clear voiced segments for the classifier, useful pathological information may be disregarded. In this study we propose a new approach that enables the classification of voice pathology by also analyzing the unvoiced information of continuous speech. The signal frames are divided in turbulent/non-turbulent, instead of voice/non-voiced. The results show that useful information is indeed present in turbulent or near unvoiced segments. A comparison with systems that use the entire signal or only the non-turbulent frames shows that the unvoiced or highly turbulent speech segments contain useful pathological information.

Paper Nr: 76
Title:

NON-LINEAR ANALYSIS OF FETAL HEART RATE IN CARDIOTOCOGRAPHY USING SAMPLE ENTROPY

Authors:

J. A. L. Marques, P. C. Cortez, J. P. V. Madeiro and F. S. Schlindwein

Abstract: The complex system of mother and foetus interacting during pregnancy contains both dependent and independent subsystems and it is unlikely that it can be studied using only linear techniques. Considering this, the conventional medical analysis of Fetal Heart Rate (FHR) based on Cardiotocography (CTG) traces can be expanded considering nonlinear approaches. This work presents the use of Sample Entropy (SampEn) as a measure of system complexity, using a 5 minutes window of FHR signal (1200 samples), using values for parameters m and r based on literature to analyse the signal complexity behaviour in time. The database is comprised of 22 pre-classified intrapartum exams, expected to have a high degree of time domain dynamics. The analysis shows that severe FHR decelerations result in small values of SampEn, reflecting a low level of complexity. On the other hand, a set of high level transient FHR accelerations also causes the same effect. The occurrences of repetitive patterns (similar to sinusoidal waves, which are pathological) cause a drop of SampEn values. The results encourage us to consider SampEn as one viable parameter for nonlinear FHR signal analysis.

Paper Nr: 84
Title:

EEG/SEEG SIGNAL MODELLING USING FREQUENCY AND FRACTAL ANALYSIS

Authors:

Vairis Caune, Juris Zagars and Radu Ranta

Abstract: EEG (Electroencephalography) is used to measure the electrical activity of a human brain. It is widely used to analyse both normal and pathological data, because of its very high temporal resolution. Different algorithms were proposed in the literature for EEG signal processing, but a difficult issue is their validation on real signals. An important goal is thus to realistically simulate EEG data. The starting point of this research was the model proposed by Rankine et al. for the surface newborn EEG signal generation. The model, based on both statistical, fractal and classical frequency modelling, has parameters estimated from the real data. A first objective is to validate and parametrize this model on adult surface EEG. A second and more important goal is to parametrize it and to apply it to depth EEG measurements (SEEG). The first results presented in this communication show that the proposed model can be applied in both cases (surface and depth adult EEG), although the parameters are slightly different. As expected, seizures cannot be modelled using this approach.

Paper Nr: 91
Title:

VISUALIZATION OF NONLINEAR CLASSIFICATION MODELS IN NEUROIMAGING - Signed Sensitivity Maps

Authors:

Peter M. Rasmussen, Tanya Schmah, Kristoffer H. Madsen, Torben E. Lund, Grigori Yourganov, Stephen C. Strother and Lars K. Hansen

Abstract: Classification models are becoming increasing popular tools in the analysis of neuroimaging data sets. Besides obtaining good prediction accuracy, a competing goal is to interpret how the classifier works. From a neuroscientific perspective, we are interested in the brain pattern reflecting the underlying neural encoding of an experiment defining multiple brain states. In this relation there is a great desire for the researcher to generate brain maps, that highlight brain locations of importance to the classifiers decisions. Based on sensitivity analysis, we develop further procedures for model visualization. Specifically we focus on the generation of summary maps of a nonlinear classifier, that reveal how the classifier works in different parts of the input domain. Each of the maps includes sign information, unlike earlier related methods. The sign information allows the researcher to assess in which direction the individual locations influence the classification. We illustrate the visualization procedure on a real data from a simple functional magnetic resonance imaging experiment.

Paper Nr: 100
Title:

NONINVASIVE MEASUREMENT OF BLOOD ACID-BASE (pH) USING CONCENTRATIONS OF EXHALED GASES

Authors:

A. S. Altaan, O. Abdallah, Mohammad T. Othman, Nasser Musaab and A. Bolz

Abstract: An important property of blood is its degree of acidity and alkalinity which is referred to as acid-base balance. The acidity or alkalinity of the blood is indicated on the pH scale. The blood pH has a serious effect on all of the body’s systems and the body uses different mechanisms to control the blood’s acid-base balance. Acid-base imbalances result primarily from metabolic or respiratory failures, both imbalances cause changing in the normal range of CO2 in the blood. The concentrations of oxygen and carbon dioxide from the exhaled breath were used to evaluate the pH of the blood. The results show the relation between concentration of the exhaled CO2 and the blood acid-base pH; decreasing CO2 causes the blood to be alkaline, while increasing CO2 leads the blood to become acidic.

Paper Nr: 104
Title:

LEAST-SQUARES ESTIMATION OF NANOPORE CHANNEL CONDUCTANCE IN VOLTAGE-VARYING EXPERIMENTS

Authors:

Christopher R. O'Donnell and William B. Dunbar

Abstract: Step-changing and sinusoidal voltage patterns have expanded the capabilities of the nanopore instrument for single molecule manipulation and measurement. A challenge with voltage-varying experiments is that capacitance in the system is excited and masks the contribution of the nanopore channel conductance in the measured current. The conductance is the parameter that can be used to infer the dynamics of the complex (e.g., DNA, or DNA-protein) in the pore. We present a least-squares parameter estimation (LSPE) algorithm for estimating the channel conductance under voltage-varying conditions, including step and sinusoidal voltages, with the objective of inferring the channel conductance parameter as continuously as possible. The algorithm is shown to recover the conductance faster than by waiting for capacitive transients to settle in step-voltage experiments, and provides accurate continuous conductance estimates in sinusoidal voltage experiments, with realistic noise levels superimposed on the measurements.

Paper Nr: 106
Title:

ARTIFICIAL NEURAL NETWORKS IN THE ESTIMATION OF MEASURES OF TEMPERATURE AND HUMIDITY INSIDE A NEONATAL INCUBATOR

Authors:

Alberto A. M. Albuquerque, Arthur P. S. Braga, Bismark C. Torrico and Otacílio M. Almeida

Abstract: This paper seeks to estimate through Artificial Neural Networks the future behavior of temperature and humidity inside an incubator. This goal is motivated by the observation that the model-based predictive control is an interesting alternative for the generation of control signals of a neonatal incubator since: (i) it seeks to optimize a performance criterion that considers the future behavior of this controller, and (ii) restrictions may be imposed on future control signals. These two features can make more safe and comfortable the microclimate inside the device for the newborn: variables such as temperature and humidity can be better kept within the limits of technical standards such as the NBR IEC 601-2-19 and its amendment No. 1, NBR IEC 60601-2-19-2000. However, one predictive model of the process to be controlled must first be obtained. The obtained neural model has accuracy in predicting the incubator behavior one time step forward compatible with the technical standard, and it is ready to be applied in a predictive control structure.

Paper Nr: 111
Title:

IMPROVED ROBUST PERFORMANCE IN A SYSTEM FOR AUTOMATIC ADMINISTRATION OF A VASOACTIVE DRUG

Authors:

Nicolò Malagutti, Arvin Dehghani and Rodney A. Kennedy

Abstract: The problem of automatic administration of vasoactive drugs to regulate mean arterial pressure in surgical and postsurgical patients can be considered as a setpoint tracking problem involving a system which is characterised by significant modelling uncertainty in the form of uncertain parameters, unmodelled dynamics and disturbances. Yet, specific levels of performance are required and patient safety must be guaranteed. As part of the development process of a novel Multiple-Model Adaptive Control (MMAC) architecture for this application, we have adopted a mixed-ì synthesis approach to controller design. Simulation results show that the new controllers are capable of improved disturbance rejection and robustness even in the face of large system delays and parametric uncertainty.

Paper Nr: 118
Title:

OPTICAL SPECTROSCOPY AND OBSTACLES BY NON-INVASIVE DETECTION OF GLUCOSE CONCENTRATION BY HOME MONITORING

Authors:

O. Abdallah, Q. Qananwah, A. Bolz, J. Hansmann, H. Walles and T. Hirth

Abstract: Tight glycemic monitoring and control is the main goal in successful diabetes management to avoid its complications. Frequent blood glucose measurements with a combination of regimented diet, exercise and insulin administration can accomplish this task. Different methods are applied for non-invasive measurement of blood glucose concentration. Despite the great interest and the intensive research in this field since 1980s, there is no convenient device at the market that can measure the glucose concentration non-invasively in an easy manner. This paper discusses the different methods for detecting the glucose concentration. Elastic and inelastic (Raman) scattering as well as fluorescence and IR Spectroscopy measurements well be shown and discussed for the development of a compact non-invasive device for home monitoring. In conclusions, an optical multi-sensor measuring the fluorescence and light scattering in the tissue optical window in and around visible range (360 nm – 1200 nm) taking the perturbation factors into account is promising and under development.

Paper Nr: 120
Title:

COMPENSATORY MOVEMENT DETECTION THROUGH INERTIAL SENSOR POSITIONING FOR POST-STROKE REHABILITATION

Authors:

Carla M. Borges, Claudia Silva, Antonio J. Salazar, Ana S. Silva, Miguel V. Correia, Rubim S. Santos and João P. Vilas-Boas

Abstract: An increasing ageing society and consequently rising number of post-stroke related neurological dysfunction patients are forcing the rehabilitation field to adapt to ever-growing demands. In parallel, an unprecedented number of research efforts and technological solutions meant for human monitoring are continuously influencing traditional methodologies, causing paradigm shifts; extending the therapist patient dynamics. Compensatory movements can be observed in post-stroke patient when performing functional tasks. Although some controversy remains regarding the functional benefits of compensatory movement as a way of accomplish a given task, even in the presence of a motor deficit; studies suggest that such maladaptive strategies may limit the plasticity of the nervous system to enhance neuro-motor recovery. This preliminary study intends to aid in the development of a system for compensatory movement detection in stroke patients through the use of accelerometry data. A post-stroke patients group is presented and discussed, instructed to perform reach and press movements while sensors were positioned at different location on the arm, forearm and trunk, in order to assess sensor positioning influence. Results suggest that P1 is advantageous for compensatory elevation movement detection at the shoulder; P4 seems the most appropriate for detecting the abduction; and P5 presents a reasonable sensitivity for detection of anteriorization and rotation of the trunk.

Paper Nr: 126
Title:

APPLICATION OF THE MP THEORY TO SYSTEMS BIOLOGY

Authors:

Vincenzo Manca and Luca Marchetti

Abstract: The main framework analysis for the most part of biological dynamics remains the theory of ordinary differential equations (ODEs). However, ODEs present some intrinsic limitations in the evaluation of the kinetic reaction rates. In contrast to ODEs, Metabolic P systems (MP systems), based on P˘aun’s P systems, were introduced for modelling metabolic systems by means of suitable multiset rewriting grammars. In this work three applications of MP systems are presented, for discovering the internal regulation logic of three phenomena relevant in systems biology: i) the Goldbeter’s mitotic oscillator; ii) the glucose/insulin dynamics in the Intravenous Glucose Tolerance Test; iii) the HER-2 oncogene-regulated transcriptome in human SUM-225 cells. Despite the differences between the considered phenomena, in all the cases a model was found that exhibits good approximation of the observed time series and highlights results which are new or that have been only theorized before.

Paper Nr: 131
Title:

EEG AND HUMAN LOCOMOTION - Descending Commands and Sensory Feedback should be Disentangled from Artifacts Thanks to New Experimental Protocols Position Paper

Authors:

Thierry Castermans, Matthieu Duvinage, Guy Cheron and Thierry Dutoit

Abstract: The main challenge when studying EEG signals related to human walk control comes from the fact that signals of many different origins are mixed up. Indeed, descending commands from the brain are generated, while ascending sensorimotor information coming from the feet is sent to the brain. In addition to the inherent complexity of the human control mechanism, experimental investigation of the cerebral activity elicited during walk is highly challenging: electrode movements are produced by movements of the head, but also by the shocks undergone by the whole body at each step, which – albeit significantly attenuated – are transmitted to the head and degrade the quality of EEG signals. Recently, different EEG studies of human locomotion have been published. These are based on different hypotheses and/or produce results that are contradictory. After reviewing and describing the discrepancies between the different approaches, we propose new experimental protocols which should help to solve important issues.

Posters
Paper Nr: 8
Title:

SURFACE EMG CLASSIFICATION FOR PROSTHESIS CONTROL - Fuzzy Logic vs. Artificial Neural Network

Authors:

Siti Anom Ahmad, Mohd Asyraf Khalid, Asnor J. Ishak and Sawal H. M. Ali

Abstract: Electromyography control system (ECS) is a well-known technique for prosthesis control application. It consists of two main modules namely feature extraction and classification. This paper presents the investigation of the classification module in the ECS. The surface electromyographic (EMG) signals were recorded from flexor and extensor muscles of the forearm during wrist flexion and extension. Standard deviation and mean absolute value were used to extract information from the raw EMG signals. Two different classifiers, fuzzy logic and artificial neural network were used in investigating the surface EMG signals. The classifier is responsible to determine the movement of the subject’s limb during specific moment. The two classifiers were compared in terms of their performance.

Paper Nr: 19
Title:

SEGMENTATION AND ANALYSIS OF RETINAL VASCULAR TREE FROM FUNDUS IMAGES PROCESSING

Authors:

Sandra Morales, Valery Naranjo, Jesús Angulo, Juan José Fuertes and Mariano Alcañiz

Abstract: From a fundus image, the system proposed in this paper automatically detects retinal vessels and measures some geometrical properties on them such as caliber and bifurcation angles. Its goal is to establish objective relations between different vessels, thus being able to determine cardiovascular risk or other diseases, as well as to monitor their progression and response to different treatments. The proposed approach is mainly based on mathematical morphology although also incorporates curvature evaluation for the detection of retinal vascular tree. This method has been validated on a public database improving the results of previous published methods.

Paper Nr: 21
Title:

CARDIAC CYCLE ARTEFACT REMOVAL IN MAGNETOENCEPHALOGRAPHIC DATA OF PATIENTS WITH DEEP BRAIN ELECTRODES - Implementation of Simultaneous Magnetoencephalographic and Local Field Potential Recordings

Authors:

Antje Bock, Andrea A. Kühn, Lutz Trahms and Tilmann H. Sander

Abstract: Simultaneous magnetoencephalography (MEG) and local field potential (LFP) recordings in patients undergoing deep brain stimulation (DBS) for severe movement disorders is a promising technique both for clinical applications and basic research. Recordings can be accomplished during the time interval between electrode insertion and implantation of the stimulator while electrodes are externalised. At present, strong cardiac cycle artefacts (CCA) are observed in theMEG signals around the area, where the disposable stainless steel electrode wires leave the skull. The CCA refers to the remanent magnetic field of those wires underneath the sensors, which are moved by local pulsations of the blood vessels. Here, we demonstrate a new approach to partially remove the CCA by applying principal component analysis (PCA) to an averaged CCA and subsequent signal space projection (SSP) method. Further steps of analysis such as coherence calculations are less distorted after SSP.

Paper Nr: 34
Title:

ECG ARTEFACT DETECTION ALGORITHM - An Algorithm to Improve Long-term ECG Analysis

Authors:

Susana Brás, Nuno Ferreira and João Paulo Silva Cunha

Abstract: Newly devices allow the analysis and collection of very long-term electrocardiogram (ECG). However, associated with this devices and long-term signal, are artefacts that conduce to misleading interpretations and diagnosis. So, new developments over automatic ECG classification are needed for a reliable interpretation. The feasibility of the cardiac systems is one of the main concerns, once they are currently used as diagnosis or help systems. In this project, an artefact detection algorithm is developed, dividing the time-series in intervals of signal and artefact. The algorithm is based on the assumption that, if the analysed frame is signal, there is not an abrupt alteration over consecutive short windows. So, the time-series is divided in consecutive nonoverlapped short windows. Over these windows, it is calculated the time-series standard deviation, the maximum and minimum slope. A threshold-based rule is applied, and the algorithm reveals that, in mean, it is verified a 99.29% of correctly classified signal and only 0.71% of signal erroneously classified. Over the results obtained, the algorithm seems to present good results, however it is needed its validation in a wider and representative sample with segments marked as artefact by multiple specialists.

Paper Nr: 39
Title:

HEART RATE VARIABILITY IN SIESTA POLYSOMNOGRAMS - A Preliminary Study

Authors:

Xosé A. Vila, Arturo J. Méndez, Abraham Otero, Leandro Rodríguez-Liñares and María J. Lado

Abstract: Nowadays, sleep apnea is a disease with a high prevalence. Its diagnosis requires to admit the patient in a hospital sleep unit and to conduct a polysomnography during the night. For this reason, many efforts have been devoted to alternative techniques to diagnose apnea from other signals, such as ECG or oxygen saturation, easier to obtain outside a hospital. The aim of this work is to investigate if these recordings behave similar to overnight ECGs.This paper presents the results of a small study (only 7 patients) conducted on short naps using heart rate variability (HRV) parameters. The results indicate that the spectral parameters are different for obstructive sleep apnea (OSA) and healthy patients. Relationship with the apnea/hypoapnea index (AHI) was also different. This is a promising starting point for more extensive studies in the future.

Paper Nr: 41
Title:

AN INDOOR LOCALIZATION ALGORITHM BASED IN WEIBULL DISTRIBUTION AND BAYESIAN CLASSIFIER

Authors:

S. Orozco-Ochoa, X. A. Vila-Sobrino, I. Gómez-Conde and M. J. Lado

Abstract: The location of objects and people by the use of Global Positioning System (GPS) or Global System for Mobile Communications (GSM) network is increasingly used to provide location-based services. These technologies work well outdoors and when the required accuracy is not very high, up to 10 meters. This paper describes an algorithm for monitoring elderly people at home, by continuously taking its position, which uses the RSS information exchanged between Bluetooth devices, Weibull distributions and Bayesian classifiers. This algorithm has been validated in a real environment, an area of 13 x 12 meters, with several rooms and corridors, where zones of approximately 6 square meters have been delimited. Our algorithm achieved a rate of correct detections of 91.875%.

Paper Nr: 42
Title:

AUTOMATIC VIDEO DETECTION OF NOCTURNAL EPILEPTIC MOVEMENT BASED ON MOTION TRACKS

Authors:

Kris Cuppens, Bert Bonroy, Anouk Van de Vel, Berten Ceulemans, Lieven Lagae, Tinne Tuytelaars, Sabine Van Huffel and Bart Vanrumste

Abstract: Epileptic seizure detection in a home situation is often not feasible due to the complicated attachment of the EEG-electrodes on the scalp. We propose to detect nocturnal seizures with a motor component in patients by means of a single video camera. To this end we use a combination of optical flow and mean shift clustering to register moving body parts. After extraction of seven features, related to amplitude, duration and direction of the motion, we carry out a first validation with a linear support vector machine classifier. This resulted in a sensitivity of 80.60% and a positive predictive value of 62.07%.

Paper Nr: 43
Title:

STUDY OF TWO FEATURE EXTRACTION METHODS TO DISTINGUISH BETWEEN THE FIRST AND THE SECOND HEART SOUNDS

Authors:

Ali Moukadem, Alain Dieterlen and Christian Brandt

Abstract: Most of the existing methods for the segmentation of heart sounds use the feature of systole and diastole duration to classify the first heart sound (S1) and the second heart sound (S2). These time intervals can become problematic and useless in several clinical real life settings which are particularly represented by severe tachycardia or in tachyarrhythmia. Consequently with the objective of development of a robust generic module for heart sound segmentation we propose to study two methods of extraction based on Singular Value Decomposition (SVD) technique to distinguish S1 from S2. A K-Neirest Neighbor (KNN) classifier is used to estimate the performance of each feature extraction method. The study uses a database with 80 subjects, including 40 cardiac pathologic sounds which contain different systolic murmurs and tachycardia cases. The first and the second proposed method reached 96 % and 95% correct classification rates, respectively.

Paper Nr: 51
Title:

LINEAR DISCRIMINANT ANALYSIS VERSUS ARTIFICIAL NEURAL NETWORK AS CLASSIFIERS FOR ELBOW ANGULAR POSITION RECOGNITION PURPOSES

Authors:

Maria Claudia F. Castro

Abstract: The increasing popularity of an Artificial Neural Network for pattern recognition and the absence of comparative studies showing its real superiority over Discriminant Analysis Methods motivated the present study, aiming at comparing the accuracy levels achieved for a Feed-Forward Multilayer Perceptron (MLP) and a Linear Discriminant Analysis (RLDA) applied to myoelectric signals to classify elbow angular positions. The results showed that there were no significant differences (t-student test p<0.05) between the average classification accuracies achieved for both methods even with the search of configuration parameters more appropriate to each situation. Both methods achieved average classification accuracies above 80% for a number of classes up to 4. However, 5 subjects achieved good results in a 5-class setup, which means a 20o shift between consecutive classes. Considering that for MLP there is an effort to define the architecture parameters and also learning parameters, its use is only justified if there is a need of generalization that cannot be achieved by the RLDA that does not require the predefinition of parameters, it is practical and fast, and performs very well.

Paper Nr: 55
Title:

OPTIC DISC DETECTION IN RETINAL IMAGES BY PATTERN DISTANCE MINIMIZATION

Authors:

Marcy A. Dias and Fernando C. Monteiro

Abstract: The retinal fundus photograph is widely used in the diagnosis and treatment of various eye diseases such as diabetic retinopathy and glaucoma. On the research work leading to automatic analysis of retinal images, the knowledge of the optic disc (OD) location is essential, and a new method to locate the optic disc automatically is proposed. We propose an algorithm for the detection of OD in the retina which takes advantage of the powerful preprocessing techniques such as the contrast enhancement, Gabor wavelet transform, mathematical morphology and Earth Mover’s distance as the matching process. Forty images of the retina from the DRIVE database were used to evaluate the performance of the method.

Paper Nr: 56
Title:

CARDIAC PATHOLOGIES DETECTION OVER FPGA USING ELECTROCARDIOGRAM

Authors:

Ariadna Vázquez-Sedano, Santiago T. Pérez-Suárez, Carlos M. Travieso-González and Jesús B. Alonso-Hernández

Abstract: In this work is presented an implementation of an automatic detection system for the heart diseases over a field programmable gate array (FPGA). The system is able to process, analyse and classify the cardiac pathologies in real time from electrocardiogram (ECG). These algorithms principally are based on Digital Wavelet Transform (DWT) techniques, and Principal Component Analysis (PCA). Finally, cardiac pulse detection and classification algorithms have been implemented in an Artificial Neural Network (ANN). Furthermore, the subjectivity problem in the heart disease diagnosis is solved, and the task of heart specialist is facilitated.

Paper Nr: 61
Title:

FLYBOW IMAGE SEGMENTATION - For Tracing Neuron Circuits in Drosophila Brain

Authors:

Hao-Chiang Shao, Wei-Yun Cheng, Yung-Chang Chen and Wen-Liang Hwang

Abstract: Recently developed were the Brainbow and Flybow techniques that can image and visualize a large number of neurons at a time. These techniques provide a way for imaging multiple neurons at the same time, and ideally, neurons can then be differentiated from each other according to their color information. However, due to dozens of neuron fibers spreading spatially in a very intricate structure, it is time-consuming to label them by hand and also difficult to trace them by using existing algorithms designed for tracing a single neuron. We proposed a prototype scheme based on grayscale morphological operations for segmenting Flybow imagery. The proposed method can provide segmentation results semi-automatically, and thus it would be useful for biologists to identify the neuro-circuits and develop the ground truth as well.

Paper Nr: 65
Title:

TOWARD A SILENT SPEECH INTERFACE BASED ON UNSPOKEN SPEECH

Authors:

Alejandro Antonio Torres García, Carlos Alberto Reyes García and Luis Villaseñor Pineda

Abstract: This work aims to interpret the EEG signals associated with actions to imagine the pronunciation of words that belong to a reduced vocabulary without moving the articulatory muscles and without uttering any audible sound (unspoken speech). Specifically, the vocabulary reflects movements to control the cursor on the computer. We have recorded EEG signals from 21 subjects using a markers based basic protocol. The discrete wavelet transform (DWT) is used to extract features from the delimited windows, and a subset of them with frequency ranges below 32 Hz is further selected. These subsets are used to train four classifiers: Naive Bayes (NB), Random Forests (RF), support vector machine (SVM), and Bagging-RF. The results are still preliminary but encouraging because the accuracy rates are above 20%, i.e. up to chance for five classes. The implementation process as well as some experiments with their corresponding results are shown.

Paper Nr: 66
Title:

BIOMETRY BASED ON EEG SIGNALS USING NEURAL NETWORK AND SUPPORT VECTOR MACHINE

Authors:

Hamid Bagherzadeh Rafsanjani, Mozafar Iqbal, Morteza Zabihi and Hideaki Touyama

Abstract: The use of EEG as a unique character to identify individuals has been considered in recent years. Biometric systems are generally operated into Identification mode and Verification mode. In this paper the feasibility of the personal recognition in verification mode were investigated, by using EEG signals based on P300, and also, the people’s identifying quality, in identification mode and especially in single trial, was improved with Neural Network (NN) and Support Vector Machine (SVM) as classifier. Nine different pictures have been shown to five participants randomly; before the test was examined, each subject had already chosen one or some pictures in order to P300 occurrence took place in examination. Results in the single trial were increased from 56.2\% in the previous study, to 75\% and 81.4\% by using SVM and NN, respectively. Meanwhile in a maximum state, 100% correctly classified was performed by only 5 times averaging of EEG. Also it was observed that using support vector machine has more sustainable results as a classifier for EEG signals that contain P300 occurrence.

Paper Nr: 67
Title:

1-D MATHEMATICAL MORPHOLOGY FOR WATER REMOVAL IN 1H MR SPECTROSCOPY TOOL

Authors:

Juan José Fuertes, Valery Naranjo, Jesús Angulo and Mariano Alcañiz

Abstract: This work shows the basics and performance of a new morphological signal method for 1-D water signal removal included in a simple and interactive multivoxel spectroscopy tool to help surgeons detect brain cancer. It consists of mathematical morphology usually applied in 2D images to filter 1D spectroscopic signals. 1D water signal reconstruction from the original data is performed in frequency domain through the use of an elementary operation: geodesic dilation. Then, the water signal is subtracted from the original signals due to the large amount of water which exists in the brain compared to the rest of molecules, making possible quantitation procces. The goal of this paper is to present this new morphological method commonly used in 2D domain for 1D water removal, spreading its use to several processing methods as quantitation.

Paper Nr: 68
Title:

BIOSIGNALS EVENTS DETECTION - A Morphological Signal-independent Approach

Authors:

Rui Santos, Joana Sousa, Borja Sañudo, Carlos J. Marques and Hugo Gamboa

Abstract: This study presents a signal-independent algorithm, which detects significant events in a biosignal, without previous knowledge or specific pre-processing steps. From a morphological analysis, the algorithm computes the instants when the most significant standard deviation discontinuities occur. An iterative optimization step is then applied. This assures that a minimal error is achieved when modeling the signal segments (between the detected instants) with a polynomial regression. The detection scale can be modified by an optional input scale factor. An objective algorithm performance evaluation procedure was designed, and applied on two types of synthetic signals, for which the events instants were previously known. An overall mean error of 20.32 (+/-16.01) samples between the detected and the real events show the high accuracy of the proposed algorithm. The algorithm was also applied on accelerometry and electromyography raw signals collected in different experimental scenarios. The fact that this approach does not require any previous knowledge and the good level of accuracy represents a relevant contribution in events detection and biosignal analysis.

Paper Nr: 71
Title:

PLUX REAL-TIME SPORTS EVALUATION - A New Real-time Tool for Sports Evaluation

Authors:

João Santinha, Rui Santos, Joana Sousa and Hugo Gamboa

Abstract: In this paper we present a new tool for athletes performance evaluation in real-time using two different biosignals (ECG and accelerometry). From the accelerometer signal, the level of activity was extracted based on a validation protocol, in which metabolic equivalent tasks (METs) values were calculated in real-time from a body-acceleration signal. METs values were obtained from various lifestyle and sporting activities, and compared with the results from the reference work (Crouter et al., 2006b) for the same activities. The results obtained showed correlation with the Crouter model. With the results we can conclude that present tool allows the assessment of the athlete performance based on ECG and accelerometer signals, being a versatile tool, which can be used by sports professionals and non-professionals.

Paper Nr: 77
Title:

INTERPRETATION OF EOG DATA IN ORDER TO OBSERVE EYE MOVEMENTS

Authors:

Tina Meißner and Harald Loose

Abstract: This paper deals with the possibilities of observing eye movements from EOG recordings. First the setup for the recording of eye movements, then the EOG method, which is used to obtain eye movements in the experimental context, are explained. The first recordings resulted in the detection of eye movements within the EOG data and the results were put into relation to the gaze points gained from the eye tracking system. In addition to these subjective observations a first attempt at quantification of the dependency between EOG signal and gaze points is presented. Though faced with a few problems, it was possible to put numerical values for distances in relation to signal amplitudes of the EOG.

Paper Nr: 88
Title:

NEURAL ADAPTATION IN LOCAL REFLEX CONTROL OF LIMB MOVEMENTS

Authors:

Oliver P. Dewhirst, Natalia Angarita-Jaimes, David M. Simpson, Robert Allen, Carlos D. Maciel and Philip L. Newland

Abstract: Neural adaptation, a change in the response of a neuron to repetitive stimulation, is a widespread property of neurons in many networks, including those controlling local reflex limb movements. The majority of previous studies have investigated the steady state properties of neurons rather than considering those of their adapting (transient) response. Bandlimited Gaussian White Noise, sinusoidal and walking stimulation signals have therefore, for the first time, been used to investigate neural adaptation in flexor and extensor motor neurons in the locusts local hind limb control system. Our results show that the adaptation rate of the response of two extensor and one flexor motor neuron are the same. We also show that the adaptation rate of the Fast Extensor Tibia motor neuron is affected by the properties of the stimulation signal.

Paper Nr: 88
Title:

NEURAL ADAPTATION IN LOCAL REFLEX CONTROL OF LIMB MOVEMENTS

Authors:

Oliver P. Dewhirst, Natalia Angarita-Jaimes, David M. Simpson, Robert Allen, Carlos D. Maciel and Philip L. Newland

Abstract: Neural adaptation, a change in the response of a neuron to repetitive stimulation, is a widespread property of neurons in many networks, including those controlling local reflex limb movements. The majority of previous studies have investigated the steady state properties of neurons rather than considering those of their adapting (transient) response. Bandlimited Gaussian White Noise, sinusoidal and walking stimulation signals have therefore, for the first time, been used to investigate neural adaptation in flexor and extensor motor neurons in the locusts local hind limb control system. Our results show that the adaptation rate of the response of two extensor and one flexor motor neuron are the same. We also show that the adaptation rate of the Fast Extensor Tibia motor neuron is affected by the properties of the stimulation signal.

Paper Nr: 89
Title:

LONG TERM BIOSIGNALS VISUALIZATION AND PROCESSING

Authors:

Ricardo Gomes, Neuza Nunes, Joana Sousa and Hugo Gamboa

Abstract: Long term acquisitions of biosignals are an important source of information about the patients’ state and its evolution, but involves managing very large datasets, which make signal visualization and processing a complex task. To overcome these problems, we introduce a new data structure to manage long term biosignals. A fast and non-specific multilevel biosignal visualization tool based on the concept of subsampling is presented, with focus on the representative signal parameters (mean, maximum, minimum and standard deviation error). The visualization tool enables an overview of the entire signal and a more detailed visualization in specific parts which we want to highlight. The ”Split and Merge” concept is exposed for long term biosignals processing. A processing tool (ECG peak detection) was adapted for long term biosignals. Several long term biosignals were used to test the developed algorithms. The visualization tool has proven to be faster than the standard methods and the developed processing algorithm detected the peaks of long term ECG signals fast and efficiently. The non-specific character of the new data structure and visualization tool, and the speed improvement in signal processing introduced by these algorithms makes them useful tools for long term biosignals visualization and processing.

Paper Nr: 97
Title:

FINDING NEW EASI ECG COEFFICIENTS - Improving EASI ECG Model using Various Regression Techniques

Authors:

Wojciech Oleksy and Ewaryst Tkacz

Abstract: Main idea of this study was to increase efficiency of the EASI ECG method introduced by Dover in 1988 using various regression techniques. EASI was proven to have high correlation with standard 12 lead ECG. Apart from that it is less susceptible to artefacts, increase mobility of patients and is easier to use because of smaller number of electrodes. Multilayer Perceptron (Artificial Neural Network), Support Vector Machine Regression (with Sequential Minimal Optimization algorithm), Linear Regression and Pace Regression methods were used to improve the quality of the 12-lead electrocardiogram derived from four (EASI) electrodes. Hundreds of ANNs with different learning rates and number of hidden layers were built and tested using data from PhysioNet and also data that were artificially generated. Next SMO Regression method with few different kernels (polynomial, normalized polynomial and RBF), Linear Regression and Pace Regression method were tested on the same dataset. All computed results were compared with those obtained using classic EASI ECG method described by Dover. Computation of Root Mean Squared Error and Correlation Coefficient was performed to measure the overall result of a given method. Obtained results show that various regression methods could be used to increase the performance of EASI ECG method and thus make it more popular.

Paper Nr: 99
Title:

VITAL-SIGN DATA FUSION MODELS FOR POST-OPERATIVE PATIENTS

Authors:

Marco A. F. Pimentel, David A. Clifton, Lei Clifton, Peter J. Watkinson and Lionel Tarassenko

Abstract: Deterioration in Patients who undergo upper-gastrointestinal surgery may be evident in the vital signs prior to adverse events. A dataset comprising observational vital-sign data from 128 post-operative patients was used to explore the trajectory of patients vital-sign changes during their stay in the post-operative ward. A model of normality based on pre-discharge data from patients who had a “normal” recovery was constructed using kernel density estimates, and tested with “abnormal” data from patients who deteriorate sufficiently to be re-admitted to the Intensive Care Unit. The results suggest that the criticality of post-operative patients can be evaluated by assessment of the distributions of their vital signs after their admission to the post-operative ward.

Paper Nr: 101
Title:

A COMPUTER AIDED DETECTION SYSTEM FOR MICROCALCIFICATIONS IN BREAST PHANTOM IMAGES

Authors:

Bruno Barufaldi, Sarah Soares de Oliveira, Leonardo Vidal Batista, Homero Schiabel and Manuella Santos Carneiro Almeida

Abstract: Breast cancer control represents one of the greatest challenges that public health service faces nowadays. In order to decrease the death rate from cancer in women, the AGEVISA-PB implemented a Mammography Quality Control Programme to improve the performance of mammographic equipment in Paraiba - Brazil. The evaluation method of these devices is accomplished through breast phantoms that simulate structures found on a mammogram in order to assure the quality of radiographic images. Even so, evaluation by technicians still suffers limitations caused by the visual inspections by individuals, such as long-time benchmarking and subjectivity. The main purpose of this research is to develop a computerised system that analyses radiological images of phantom MAMA-CDM and correlates with human visual perception. The results indicate that the system developed can be used as a second opinion, thus becoming a tool of great utility in aiding medical diagnosis.

Paper Nr: 105
Title:

BRAKE RESPONSE TIME BEFORE AND AFTER TOTAL KNEE ARTHROPLASTY - Tracking Possible Effects of the Surgery Technique on Motor Performance: Report of Two Cases

Authors:

Carlos J. Marques, Rui Santos, Hugo Gamboa, Frank Lampe, João Barreiros and Jan Cabri

Abstract: After total knee arthroplasty (TKA) patients often ask when they can resume car driving. This question was the aim of some studies in the past, however no study was found on the possible effects of different surgery techniques on brake response time (BRT). A randomized controlled trial on the effects of two surgery techniques (minimally invasive vs. standard approach) on BRT was designed. In this paper the motor performance of two female patients was compared. Surgery had different effects on the mean BRT of both Patients. The mean BRT of the MIS Patient wasn’t increased 7 days after surgery, while the BRT of the Patient undergoing standard surgery was increased by 46.8% at the same time.

Paper Nr: 109
Title:

NEUROPHYSIOLOGIC AND STATISTICAL ANALYSIS OF FAILURES IN AUTOMATIC SLEEP STAGE CLASSIFICATION

Authors:

Teresa Sousa, Dulce Oliveira, Sirvan Khalighi, Gabriel Pires and Urbano Nunes

Abstract: This paper analyses some of the challenges in automatic multiclass sleep stage classification. Six electroencephalographic (EEG) and two electrooculographic (EOG) channels were used in this study. A set of significant features are selected by a minimum-redundancy maximum-relevance (mRMR) criterion and then classified using support vector machine (SVM). The system is tested on 14 subjects suspected of having sleep apnea. The automatic sleep staging showed a 77.70% (±15.8) sensitivity and 95.49% (±2.68) specificity. From the analysis comparing EEG records with visual and automatic classification, we found that the main cause of failures are the similarities between adjacent phases of sleep, in particular in discriminating N1 and N2. Based on the variation of the values of the features it is possible to implement some thresholds and to apply some heuristic rules to improve the performance.

Paper Nr: 110
Title:

FACE AND EYE TRACKING FOR PARAMETERIZATION OF COCHLEAR IMPLANTS

Authors:

M. Cabeleira, S. Ferreira, L. F. Silva, C. Correia and J. Cardoso

Abstract: This work presents a free head eye-tracking solution created for use as a complementary tool in the parameterization of cochlear implants. Nowadays, the parameterization of these implants is a long and cumbersome process performed by audiologists and speech therapists that throughout many periodic evaluations where audiometric tests and electrode adjustments are performed. The eye tracking system will assist this process through detection of saccades generated when a subject hears sounds produced during the audiometric test procedure. The main purpose is to ease and improve the implant re-parameterization procedure with uncooperative subjects, like children. The developed system is composed of three digital video cameras where two of the cameras are responsible of the detection of the position of the face and eyes and the third is responsible of the gaze detection. The developed face and eye detectors are also compared in order to choose the best combination of algorithms to perform robust eye detection with unpredictable subjects. The best combination of algorithms is the Viola-Jones face detector combined with an eye detector Ring Gabor filters, that correctly detected the eye-position in 76,81% of the tested videos at 18 frames per second.

Paper Nr: 115
Title:

BRAIN SEGMENTATION IN HEAD CT IMAGES

Authors:

Ana Sofia Torres and Fernando C. Monteiro

Abstract: Brain segmentation in head computed tomography scans is essential for the development of computer-aided diagnostic methods for identifying the brain diseases. In this paper we present a hybrid framework to brain segmentation which joints region-based information based on watershed transform with clustering techniques. A pre-processing step is used to reduce the spatial resolution without losing important image information. An initial partitioning of the image into primitive regions is set by applying a rainfalling watershed algorithm on the image gradient magnitude. This initial partition is the input to a computationally efficient region segmentation process which produces the final segmentation. We have applied our approach on several head CT images and the results reveal the robustness and accuracy of this method.

Paper Nr: 116
Title:

REGULARIZED RECONSTRUCTION OF ULTRASONIC IMAGING AND THE REGULARIZATION PARAMETER CHOICE

Authors:

Leonardo G. S. Zanin, Fábio K. Schneider and Marcelo V. W. Zibetti

Abstract: Ultrasound image reconstruction based on inverse problems has attracted attention to the ultrasonic imaging research community recently. Different from standard beamforming-based methods techniques, this new imaging method tries to solve a linear system g=Hf as a form of reconstructing the ultrasound image. In order to understand the behaviour of this imaging system, it is important to analyse the forward problem. In this paper, we analyse the effect of the noise in acquisition matrix using singular value decomposition. Also, the effect of regularization parameter in dealing with the noise is investigated in regularized. This analysis provides some interesting insights in the understanding of how the inverse reconstruction can be improve some aspects higher than beamforming.

Paper Nr: 127
Title:

DECODING SSVEP RESPONSES BASED ON PARAFAC DECOMPOSITION

Authors:

Nikolay V. Manyakov, Nikolay Chumerin, Adrien Combaz, Arne Robben, Marijn van Vliet and Marc M. Van Hulle

Abstract: In this position paper, we investigate whether a parallel factor analysis (Parafac) decomposition is beneficial to the decoding of steady-state visual evoked potentials (SSVEP) present in electroencephalogram (EEG) recordings taken from the subject’s scalp. In particular, we develop an automatic algorithm aimed at detecting the stimulation frequency after Parafac decomposition. The results are validated on recordings made from 54 subjects using consumer-grade EEG hardware (Emotiv’s EPOC headset) in a real-world environment. The detection of one frequency among 12, 4 and 2 possible was considered to assess the feasibility for Brain Computer Interfacing (BCI). We determined the frequencies subsets, among all subjects, that maximize the detection rate.

Paper Nr: 129
Title:

RELATIONSHIP BETWEEN QUANTITATIVE T-WAVE ALTERNANS ESTIMATES AND PARAMETERS DESCRIBING CLINICAL STATUS OF PATIENT IN ACCUTE PHASE OF MYOCARDIAL INFARCTION AND OUTCOME RESULTS

Authors:

Algimantas Krisciukaitis, Renata Simoliuniene, Andrius Macas, Robertas Petrolis, Eimante Kamile Puodziunaite, Zita Bertasiene and Viktoras Saferis

Abstract: Relationship between parameters describing clinical status of the patient in acute phase of myocardial infarction, outcome results and quantitative estimates of T-wave alternans, known prognostic factor of severe cardiac arrhythmias or sudden cardiac death, was investigated in aim to reveal their usefulness and incremental diagnostic utility. Integrated estimate, reflecting differences between S-T,T complexes of odd and even cardiocycles, obtained by means of Principal Component Analysis showed significant correlation with Left Ventricular Ejection Fraction and rehospitalization of patient within 6 months.

Paper Nr: 130
Title:

EMOTION CLASSIFICATION BASED ON PHYSIOLOGICAL RESPONSES INDUCED BY NEGATIVE EMOTIONS - Discrimination of Negative Emotions by Machine Learning Algorithms

Authors:

Eun-Hye Jang, Byoung-Jun Park, Sang-Hyeob Kim and Jin-Hun Sohn

Abstract: The one of main topic of emotion recognition or classification research is to recognize human’s feeling or emotion using physiological signals, which is one of the core processes to implement emotional intelligence in HCI research. The aim of this study was to identify the optimal algorithm to discriminate negative emotions (sadness, anger, fear, surprise, and stress) using physiological features. Physiological signals such as EDA, ECG, PPG, and SKT were recorded and analysed. 28 features were extracted from these signals. For classification of negative emotions, five machine learning algorithms, namely, LDF, CART, SOM, Naïve Bayes and SVM were used. Result of emotion classification showed that an accuracy of emotion classification using SVM was the highest (100.0\%) and that of LDA was the lowest (41.3\%). 78.2\%, 45.8\%, and 73.3% were shown as the accuracy of emotion classification in CART, SOMs and Naïve Bayes, respectively. This can be helpful to provide the basis for the emotion recognition technique in HCI.

Paper Nr: 134
Title:

A RULE-BASED CLASSIFICATION OF LARYNGOPATHIES BASED ON SPECTRUM DISTURBANCE ANALYSIS - An Exemplary Study

Authors:

Krzysztof Pancerz, Wiesław Paja, Jarosław Szkoła, Jan Warchoł and Grażyna Olchowik

Abstract: Our research concerns data derived from the examined patient’s speech signals for a non-invasive diagnosis of selected larynx diseases. The paper is devoted to the rule-based classification of patients on the basis of a family of coefficients reflecting spectrum disturbances around basic tones and their multiples. The paper presents a proposed procedure for feature selection and classification as well as the experiments carried out on real-life data.

Paper Nr: 136
Title:

COMPLEX NETWORK PROPERTIES OF EYE-TRACKING IN THE FACE RECOGNITION PROCESS - An Initial Study

Authors:

Bolesław Jaskuła, Jarosław Szkoła and Krzysztof Pancerz

Abstract: In the paper, we propose to investigate eye-tracking sequences obtained in the face recognition process in terms of complex networks. A proper algorithm for transformation sequences coming from eye-tracking into complex networks is described. The analysis of parameters of obtained complex networks can be helpful in better understanding and classifying human mental behaviors and activities.

Paper Nr: 137
Title:

ANALYSIS OF BERG BALANCE SCALE IN HIP FRACTURE PATIENTS USING FUZZY CLUSTERING

Authors:

Aleksandar Jeremic, Natasa Radosavljevic, Dejan Nikolic and Milica Lazovic

Abstract: Hip fractures are most frequent cause of hospitalization after the fall in older population and consequently have been subject of great interest in medicine and biomedical engineering. It has been observed that the incidence of hip fractures is rising at the approximate rate of 1-3% per year, with subsequent mortality rates at approximately 33% in first year after the fracture. In this paper we propose to classify patients at the time of admission into several clusters with respect to their ability for successful recovery. To this purpose we first evaluate the efficacy of rehabilitation program based on the balance function improvement measured by Berg Balance Scale (BBS) in elderly (in the remainder of the paper defined as above 65 years of life) after hip fractures, and evaluate influence of gender, age and comorbidity on balance function improvement in these patients. Then we design clustering procedure in which the patients are clustered according to BBS improvement using statistically most significant parameters. We then evaluate the proposed clustering procedure on a data sample consisting of 203 patients that have been admitted to the Institute for Rehabilitation, Belgrade, Serbia.