Ahmad Diab

Effect of filtering on the classification rate of nonlinear analysis methods applied to uterine EMG signals

Nonlinear time series analysis can provide useful information regarding nonlinear features of biological signals. The effect of filtering on the performance of nonlinear methods is not well-understood. In this work, we investigate the effects of signal filtering on the sensitivity of four nonlinear methods: Time reversibility, Sample Entropy, Lyapunov Exponents and Delay Vector Variance. These methods were applied to uterine EMG signals with the aim of using them to discriminate between pregnancy and labor contractions. The signals were filtered using three different band-pass filters before the application of the methods. Results showed that the sensitivity of some methods such as sample entropy was significantly improved with filtering. On the other hand, filtering had little effect on some other methods such as time reversibility. This study concludes that while filtering increases computation time, it may be necessary for some nonlinear methods particularly those with low sensitivity.

Classification of pregnancy and labor contractions using a graph theory based analysis.

In this paper, we propose a new framework to characterize the electrohysterographic (EHG) signals recorded during pregnancy and labor. The approach is based on the analysis of the propagation of the uterine electrical activity. The processing pipeline includes i) the estimation of the statistical dependencies between the different recorded EHG signals, ii) the characterization of the obtained connectivity matrices using network measures and iii) the use of these measures in clinical application: the classification between pregnancy and labor. Due to its robustness to volume conductor, we used the imaginary part of coherence in order to produce the connectivity matrix which is then transformed into a graph. We evaluate the performance of several graph measures. We also compare the results with the parameter mostly used in the literature: the peak frequency combined with the propagation velocity (PV +PF). Our results show that the use of the network measures is a promising tool to classify labor and pregnancy contractions with a small superiority of the graph strength over PV+PF.

Towards a usable and an efficient elder fall detection system

In-house monitoring of elders and automatic fall detection using intelligent sensors is a very desirable service that has the potential of increasing autonomy and independence while minimizing the risks of living alone. The efforts of building such systems have been spanning for decades, but there still is a lot of room for improvement. This paper proposes a novel approach to make a successful monitoring and assistive services for elderly. Moreover, we present our current progress of data collection, parameters extraction and parameters selection that are essential phases of our project. Our results on the data demonstrate that the proposed system methods are efficient and accurate and can be easily used in a real fall detection system.

Non-linear analysis of human stability during static posture

The goal of this paper is to analyze the human stability in the upright position. This is achieved by analyzing the visual input, feet placement, and age affects on the human body equilibrium using stabilometric signals. Twenty eight healthy subjects (14 young and 14 elderly subjects) participated to this study to collect their stabilometric signals under four conditions. Three non linear methods are used: time reversibility, sample Entropy and Lyapunov exponents methods. Lyapunov exponents show the best results especially between feet apart and feet together condition and between young and elderly subjects with significant differences. ROC analysis and Anova test were used and the results show significant differences between groups with p-values less than 0.001.

Pregnancy monitoring using graph theory based analysis

Monitoring pregnancy using noninvasive recordings of uterine contractions is still an unsolved issue. Here, we propose a new way to tackle this problem using the electrohysterographic (EHG) signals recorded during pregnancy and labor. The new approach is based on the analysis of the propagation of the uterine electrical activity. The proposed pipeline includes i) the computation of the statistical dependencies between the multichannel (4 × 4 matrix) EHG signals, ii) the characterization of the connectivity matrices using network measures (graph-theory based analysis) and iii) the use of these measures in pregnancy monitoring. Due to its robustness to volume conduction, we used the imaginary part of coherence function to create the connectivity matrices transformed then into graphs. The method is evaluated on a dataset of EHG signals to track the correlation between uterine signals with weeks before labor. The results show a difference in the graph densities from pregnancy to labor. We speculate that the network based analysis is a very promising tool for pregnancy monitoring.

Evaluation of HD- sEMG grid misalignment with muscle fibers using nonlinear correlation

In a recent past, several techniques have been developed to analyze the effect of electrode grid alignment with the muscle fibers. But, they focused on pattern matching of Motor Unit (MU) electrical signatures and not on the interference signal. In this study, we use a method which is widely applied in connectivity and directionality analysis of stochastic and complex signals, namely, the nonlinear correlation coefficient (h2) on HD-sEMG signal matrices. The approach is first applied on simulated data from recent generation model through an (8×8) simulated electrode matrix, then on real signals using the same grid specifications on the Biceps Brachii. Both simulated and real data were evaluated with three angles of grid alignment with respect to muscle fibers. For this purpose, five parameters were extracted from obtained h2 correlation matrices and tested. According to the obtained results, a relationship between h2 values and the electrode matrix alignment seems to exist. However, further efforts are needed to design parameters more sensitive to grid misalignment with respect to muscle fibers.

Variation-based sparse source imaging in localizing uterine activity

Electrohysterogram source imaging, i.e., moving from the electrode/sensor space to the source space using EHG signals, provide an estimate of spatial distributions of uterine activity at millisecond scale. This paper aims to study the ability of different distributed source localization methods to recover uterine electrical activity sources. Performance was quantified using a detection accuracy index. Our result suggests that the variation based method is able to reconstruct extended uterus sources with the overall high accuracy, where the increasing of the electrodes numbers and the decreasing of the fat thickness induce a better accuracy in localization.

Separation and localization of EHG sources using tensor models

The use of the Electrohysterogram (EHG) for imaging the sources of the uterine electrical activity is a new and powerful diagnosis technique. However, its performance is limited as the uterus often demonstrates several simultaneously active regions and as EHGs present low signal-to-noise ratios. To overcome these problems, tensor-based preprocessing can be applied, which consists in constructing a space-time-frequency (STF) or space-time-wave-vector (STWV) tensor and decomposing it by using the Canonical Polyadic (CP) decomposition. In this paper, we present an algorithm for the accurate localization of extended sources based on the results of the tensor decomposition. Furthermore, we analyse its performance on realistic simulated data in comparison to conventional source localization algorithms.

Precise assessment of HD-sEMG grid misalignment using nonlinear correlation

The aim of this paper is to perform a precise assessment of the misalignment of high-density surface electromyogram (HD-sEMG) grid according to its rotation with the muscle fibers. For this purpose, a generic and quantitative method was used in this study, namely the nonlinear correlation coefficient (h2) that is widely applied in connectivity and directionality of stochastic and complex signals. This approach is applied on simulated data through an (8×8) simulated electrode matrix placed on the simulated Biceps Brachii using a multilayered cylindrical model (muscle, fat tissue and skin tissue). Simulations with 5 different anatomies and 5 different motor unit recruitment patterns were computed at three constant contractions levels 20%, 40%, and 60% of the maximal voluntary contraction (MVC) for grid rotation from −10° to 10°, with a step of one degree according to muscle fibers. The obtained results show an important sensitivity of the nonlinear correlation parameter, namely the mean of the image matrix value, to the grid rotation angle. Furthermore, this parameter seems to be also sensitive to the contraction level and muscle anatomy.

Ambient assistive living system using RGB-D camera

One of the main problems that are being addressed intensively in modern societies is the ageing of population. Todays challenge is to allow elderly people to remain autonomous at their home as much as possible. Currently, one of the active research fields is the development of an assistive living system (ALS) that aims to support people at home. This can help elderly people to stay at home as long as possible, which consequently satisfies elders and reduces the health care cost. In this paper, we propose a simple and efficient way to track and recognize basic elderly activities using a single RGB-D camera. The main benefit of this method is the low cost and the effortless deployment and installation, as well as an overall recall of 97%.