Javier Garcia-Casado

Noninvasive measurement and analysis of intestinal myoelectrical activity using surface electrodes

Intestinal myoelectrical activity (IMA), which determines bowel mechanical activity, is the result of two components: a low-frequency component [slow wave (SW)] that is always present, and a high-frequency component [spike bursts (SB)] which is associated with bowel contractions. Despite of the diagnostic significance of internal recordings of IMA, clinical application of this technique is limited due to its invasiveness. Thus, surface recording of IMA which is also called electroenterogram (EEnG) could be a solution for noninvasive monitoring of intestinal motility. The aim of our work was to identify slow wave and spike burst activity on surface EEnG in order to quantify bowel motor activity. For this purpose, we conducted simultaneous recordings of IMA in bowel serosa and on abdominal surface of five Beagle dogs in fast state. Surface EEnG was studied in spectral domain and frequency bands for slow wave and spike burst energy were determined. Maximum signal-to-interference ratio (7.5dB/spl plusmn/36%) on SB frequency band was obtained when reducing upper frequency limit of signal analysis. Energy of external EEnG in reduced SB frequency band (2-7.9Hz) presented a high correlation (0.71/spl plusmn/7%) with internal intensity of contractions. Our results suggest that energy of SB can be quantified on external EEnG which could provide a noninvasive method for monitoring intestinal mechanical activity.

Active concentric ring electrode for non-invasive detection of intestinal myoelectric signals

Although the surface electroenterogram (EEnG) is a weak signal contaminated by strong physiological interference, such as ECG and respiration, abdominal surface recordings of the EEnG could provide a non-invasive method of studying intestinal activity. The goal of this work was to develop a modular, active, low-cost and easy-to-use sensor to obtain a direct estimation of the Laplacian of the EEnG on the abdominal surface in order to enhance the quality of bipolar surface monitoring of intestinal activity. The sensor is made up of a set of 3 concentric dry Ag/AgCl ring electrodes and a battery-powered signal-conditioning circuit. Each section is etched on a different printed circuit board (PCB) and the sections are joined to each other by surface mount technology connectors. This means the sensing electrodes can be treated independently for purposes of maintenance and replacement and the signal conditioning circuit can be re-used. A total of ten recording sessions were carried out on humans. The results show that the surface recordings of the EEnG obtained by the active sensor present significantly less ECG and respiration interference than those obtained by bipolar recordings. In addition, bioelectrical sources whose frequency fitted with the slow wave component of the EEnG (SW) were identified by parametric spectral analysis in the surface signals picked up by the active sensors.

Active flexible concentric ring electrode for non-invasive surface bioelectrical recordings

Bioelectrical surface recordings are usually performed by unipolar or bipolar disc electrodes even though they entail the serious disadvantage of having poor spatial resolution. Concentric ring electrodes give improved spatial resolution, although this type of electrode has so far only been implemented in rigid substrates and as they are not adapted to the curvature of the recording surface may provide discomfort to the patient. Moreover, the signals recorded by these electrodes are usually lower in amplitude than conventional disc electrodes. The aim of this work was thus to develop and test a new modular active sensor made up of concentric ring electrodes printed on a flexible substrate by thick-film technology together with a reusable battery-powered signal-conditioning circuit. Simultaneous ECG recording with both flexible and rigid concentric ring electrodes was carried out on ten healthy volunteers at rest and in motion. The results show that flexible concentric ring electrodes not only present lower skin–electrode contact impedance and lower baseline wander than rigid electrodes but are also less sensitive to interference and motion artefacts. We believe these electrodes, which allow bioelectric signals to be acquired non-invasively with better spatial resolution than conventional disc electrodes, to be a step forward in the development of new monitoring systems based on Laplacian potential recordings.

Comparison of non-invasive electrohysterographic recording techniques for monitoring uterine dynamics

Non-invasive recording of uterine myoelectric activity (electrohysterogram, EHG) could provide an alternative to monitoring uterine dynamics by systems based on tocodynamometers (TOCO). Laplacian recording of bioelectric signals has been shown to give better spatial resolution and less interference than mono- and bipolar surface recordings. The aim of this work was to study the signal quality obtained from monopolar, bipolar and Laplacian techniques in EHG recordings, as well as to assess their ability to detect uterine contractions. Twenty-two recording sessions were carried out on singleton pregnant women during the active phase of labour. In each session the following simultaneous recordings were obtained: internal uterine pressure (IUP), external tension of abdominal wall (TOCO) and EHG signals (5 monopolar and 4 bipolar recordings, 1 discrete approximation to the Laplacian of the potential and 2 estimates of the Laplacian from two active annular electrodes). The results obtained show that EHG is able to detect a higher number of uterine contractions than TOCO. Laplacian recordings give improved signal quality over monopolar and bipolar techniques, reduce maternal cardiac interference and improve the signal-to-noise ratio. The optimal position for recording EHG was found to be the uterine median axis and the lower centre-right umbilical zone.

Adaptive filtering of ECG interference on surface EEnGs based on signal averaging

An external electroenterogram (EEnG) is the recording of the small bowel myoelectrical signal using contact electrodes placed on the abdominal surface. It is a weak signal affected by possible movements and by the interferences of respiration and, principally, of the cardiac signal. In this paper an adaptive filtering technique was proposed to identify and subsequently cancel ECG interference on canine surface EEnGs by means of a signal averaging process time-locked with the R-wave. Twelve recording sessions were carried out on six conscious dogs in the fasting state. The adaptive filtering technique used increases the signal-to-interference ratio of the raw surface EEnG from 16.7 +/- 6.5 dB up to 31.9 +/- 4.0 dB. In addition to removing ECG interference, this technique has been proven to respect intestinal SB activity, i.e. the EEnG component associated with bowel contractions, despite the fact that they overlap in the frequency domain. In this way, more robust non-invasive intestinal motility indicators can be obtained with correlation coefficients of 0.68 +/- 0.09 with internal intestinal activity. The method proposed here may also be applied to other biological recordings affected by cardiac interference and could be a very helpful tool for future applications of non-invasive recordings of gastrointestinal signals.

Automatic Identification of Motion Artifacts in EHG Recording for Robust Analysis of Uterine Contractions

Electrohysterography (EHG) is a noninvasive technique for monitoring uterine electrical activity. However, the presence of artifacts in the EHG signal may give rise to erroneous interpretations and make it difficult to extract useful information from these recordings. The aim of this work was to develop an automatic system of segmenting EHG recordings that distinguishes between uterine contractions and artifacts. Firstly, the segmentation is performed using an algorithm that generates the TOCO-like signal derived from the EHG and detects windows with significant changes in amplitude. After that, these segments are classified in two groups: artifacted and nonartifacted signals. To develop a classifier, a total of eleven spectral, temporal, and nonlinear features were calculated from EHG signal windows from 12 women in the first stage of labor that had previously been classified by experts. The combination of characteristics that led to the highest degree of accuracy in detecting artifacts was then determined. The results showed that it is possible to obtain automatic detection of motion artifacts in segmented EHG recordings with a precision of 92.2% using only seven features. The proposed algorithm and classifier together compose a useful tool for analyzing EHG signals and would help to promote clinical applications of this technique.

Identification of the slow wave component of the electroenterogram from Laplacian abdominal surface recordings in humans

The electroenterogram (EEnG) is a surface recording of the myoelectrical activity of the smooth muscle layer of the small intestine. It is made up of two signals: a low-frequency component, known as the slow wave (SW), and high-frequency signals, known as spike bursts (SB). Most methods of studying bowel motility are invasive due to the difficult anatomic access of the intestinal tract. Abdominal surface EEnG recordings could be a noninvasive solution for monitoring human intestinal motility. However, surface EEnG recordings in humans present certain problems, such as the low amplitude of the signals and the influence of physiological interference such as the electrocardiogram (ECG) and respiration. In this study, a discrete estimation of the abdominal surface Laplacian potential was obtained using Hjorths method. The objective was to analyze the enhancement given by Laplacian EEnG estimation compared to bipolar recordings. Eight recording sessions were carried out on eight healthy human volunteers in a state of fasting. First, the ECG interference content present in the bipolar signals and in the Laplacian estimation were quantified and compared. Secondly, to identify the SW component of the EEnG, respiration interference was removed by using an adaptive filter, and spectral estimation techniques were applied. The following parameters were obtained: the dominant frequency (DF) of the signals, stability of the rhythm (RS) of the DF detected and the percentage of DFs within the typical frequency range for the SW (TFSW). Results show the better ability of the Laplacian estimation to attenuate ECG interference, as compared to bipolar recordings. As regards the identification of the SW component of the EEnG, after removing respiration interference, the mean value of the DF in all abdominal surface recording channels and in their Laplacian estimation ranged from 0.12 to 0.14 Hz (7.3 to 8.4 cycles min(-1) (cpm)). Furthermore in 80% of the cases, the detected DFs were inside the typical human SW frequency range, and the ratio of frequency change in the surface bipolar and Laplacian estimation signals, in 90% of the cases, was within the frequency change accepted for human SW. Significant statistical differences were also found between the DF of all surface signals (bipolar and Laplacian estimation) and the DF of respiration. In conclusion, it was demonstrated that the discrete Laplacian potential estimation attenuated the physiological interference present in bipolar surface recordings, especially ECG. Furthermore, a slow frequency component, whose frequency, rhythm stability and amplitude fitted with the SW patterns in humans, was identified in bipolar and Laplacian estimation signals. This could be a useful non-invasive tool for monitoring intestinal activity by abdominal surface recordings.

Adaptive cancellation of the ECG interference in external electroenterogram

The small bowel myoelectrical activity (electroenterogram) can be recorded placing electrodes on the abdominal skin. This surface signal contains noise and interferences. The most significant interferences in the surface electroenterogram are the respiratory artifact and the electrocardiogram signal (ECG). Conventional filtering cannot be applied to eliminate these interferences because the signal and interferences have overlap spectra. The goal of this paper is to cancel the ECG interference from surface register through adaptive filtering. Internal and external recordings were acquired simultaneously from four Beagle dogs. In this study three different configurations of adaptive filters (based on LMS algorithm) were tested for ECG cancellation from external electroenterogram. Spectral parameters were calculated from surface recording and filtered surface recordings: energy (EF), energy over two Hz (EF2), mean frequency (FM) and frequency standard deviation (FSD). Correlation between surface recording parameters and internal motility index (IMI) is analysed. Results show that spectral parameter EF2 of filtered external electroenterogram presents high correlation coefficient value. The adaptive interference cancellation is a very efficient method to solve the problem when signal and interference have overlap spectra. Small bowel motility can be studied through filtered external register of the electroenterogram without need of surgical intervention.

Identifying physical activity type in manual wheelchair users with spinal cord injury by means of accelerometers.

Study design:This was a cross-sectional study.Objectives:The main objective of this study was to develop and test classification algorithms based on machine learning using accelerometers to identify the activity type performed by manual wheelchair users with spinal cord injury (SCI).Setting:The study was conducted in the Physical Therapy department and the Physical Education and Sports department of the University of Valencia.Methods:A total of 20 volunteers were asked to perform 10 physical activities, lying down, body transfers, moving items, mopping, working on a computer, watching TV, arm-ergometer exercises, passive propulsion, slow propulsion and fast propulsion, while fitted with four accelerometers placed on both wrists, chest and waist. The activities were grouped into five categories: sedentary, locomotion, housework, body transfers and moderate physical activity. Different machine learning algorithms were used to develop individual and group activity classifiers from the acceleration data for different combinations of number and position of the accelerometers.Results:We found that although the accuracy of the classifiers for individual activities was moderate (55–72%), with higher values for a greater number of accelerometers, grouped activities were correctly classified in a high percentage of cases (83.2–93.6%).Conclusions:With only two accelerometers and the quadratic discriminant analysis algorithm we achieved a reasonably accurate group activity recognition system (>90%). Such a system with the minimum of intervention would be a valuable tool for studying physical activity in individuals with SCI.

Enhancement of the non-invasive electroenterogram to identify intestinal pacemaker activity

Surface recording of electroenterogram (EEnG) is a non-invasive method for monitoring intestinal myoelectrical activity. However, surface EEnG is seriously affected by a variety of interferences: cardiac activity, respiration, very low frequency components and movement artefacts. The aim of this study is to eliminate respiratory interference and very low frequency components from external EEnG recording by means of empirical mode decomposition (EMD), so as to obtain more robust indicators of intestinal pacemaker activity from the external EEnG signal. For this purpose, 11 recording sessions were performed in an animal model under fasting conditions and in each individual session the myoelectrical signal was recorded simultaneously in the intestinal serosa and the external abdominal surface in physiological states. Various parameters have been proposed for evaluating the efficacy of the method in reducing interferences: the signal-to-interference ratio (S/I ratio), attenuation of the target and interference signals, the normal slow wave percentage and the stability of the dominant frequency (DF) of the signal. The results show that the S/I ratio of the processed signals is significantly greater than the original values (9.66 +/- 4.44 dB versus 1.23 +/- 5.13 dB), while the target signal was barely attenuated (-0.63 +/- 1.02 dB). The application of the EMD method also increased the percentage of the normal slow wave to 100% in each individual session and enabled the stability of the DF of the external signal to be increased considerably. Furthermore, the variation coefficient of the DF derived from the external processed signals is comparable to the coefficient obtained using internal recordings. Therefore, the EMD method could be a very useful tool to improve the quality of external EEnG recording in the low frequency range and therefore to obtain more robust indicators of the intestinal pacemaker activity from non-invasive EEnG recordings.