Source Separation From Single-Channel Recordings by Combining Empirical-Mode Decomposition and Independent Component Analysis

Today, it is generally accepted that current methods for biophysical antenatal surveillance do not facilitate a comprehensive and reliable assessment of foetal well-being and thus, that continuing research into alternative methods is necessary to improve antenatal monitoring procedures. Here, attention has been paid to the abdominal phonogram, a signal that is recorded by positioning an acoustic sensor on the maternal womb and contains valuable information about foetal status, but which is hidden by maternal and environmental sources. To recover such information, this work describes single-channel independent component analysis (SCICA) as an alternative signal processing approach for analyzing the abdominal phonogram. The approach, based on the method of delays, the Temporal Decorrelation Source SEParation implementation (TDSEP) of Independent Components Analysis (ICA), and an automatic grouping algorithm, has managed to successfully retrieve estimates of: (1) the foetal cardiac activity (in the form of the foetal phonocardiogram, FPCG), (2) the maternal cardiovascular activity (in the form of the maternal phonocardiogram, MPCG, and/or pulse wave), (3) the maternal respiratory activity (in the form of the maternal respirograma, MResp), and (4) noise (N). These results have been obtained from a dataset of 25 single-channel phonograms and point at the possibilities of using SCICA to address a fundamental problem faced in antenatal surveillance, i.e. the extraction of information from a non-invasive signal like the abdominal phonogram. Future work will test the possibility of using SCICA to recover information regarding the foetal breathing movements (FBM), another physiological parameter of interest in foetal surveillance.

Signal estimation from functional magnetic resonance imaging data (fMRI) is a difficult and challenging task that involves carefully chosen models that can be validated by domain experts. Blind source separation (BSS) techniques such as principal component analysis (PCA), independent component analysis (ICA), and tensor decomposition have been proposed for estimating the task-fMRI signals. However, few studies have investigated their efficacy for classifying different tasks. This paper proposes yet another novel method to validate the signal estimation using predictive performance of the methods. Three commonly used BSS techniques are used for estimating signals (spatio-temporal maps) from task-fMRI. The dataset (consisting of 4 tasks) is taken from the Human Connectome Project (HCP). The extracted temporal signals are used to train a classifier that classifies two different tasks. For a binary classification, a Constrained-PARAFAC tensor decomposition achieves 92% accuracy for differentiating gambling vs. relational tasks. In addition, results for other binary task classifications are also presented. Overall, the proposed Constrained-PARAFAC performs better in prediction performance compared to PCA and ICA.

Blind source separation plays an important role in extracting the source components from one or more mixture(s) of the sources received by a sensor or receiver. It is blind since no other information besides the observed mixture signals is available. In presence of only one observed mixture, it is known as single channel blind source separation (SCBSS). This paper proposes a method of SCBSS based on variational mode decomposition (VMD) and principal component analysis (PCA). The observed signal is decomposed into a number of modes simultaneously using VMD. Then, PCA is used to select the corresponding source components from the decomposed modes. An example illustrating the separation of a sinusoid and speech signal by our proposed method has been presented. Cross-correlation coefficient has been used to measure the performance of the VMD-PCA algorithm. Our method has been compared with the single channel independent component analysis (SCICA) and ensemble empirical mode decomposition-principal component analysis-independent component analysis (EEMD-PCA-ICA) methods in terms of quality of performance and computational complexity.

In this work we present a technique for applying Blind Source Separation (BSS) to single channel recordings of Electromagnetic (EM) brain signals. Single channel recordings of brain signals are preprocessed through the method of delays, and the delay matrix processed with the BSS technique described here called LSDIAGTD which uses temporal decorrelation to implement the now popular Independent Component Analysis (ICA) algorithm. This allows the identification and extraction of statistically independent sources underlying these single channel recordings. In particular we depict the analysis of single channel recordings from a Brain-Computer Interfacing paradigm. We show that BSS technique applied in this way extracts a series of codebook vectors representing the spectral content underlying the recorded signal. It then becomes possible to identify and extract particular rhythmic activity underlying the recordings. We show that rhythmic activity in the 8 to 12Hz band can be extracted in the case of imagined hand movements for a particular BCI paradigm.

Electroencephalographic (EEG) signals recorded from the scalp surface are generally highly correlated. Each channel is a linear mixture of concurrently active brain and non-brain electrical sources whose activities are volume conducted to the scalp electrodes with broadly overlapping patterns (Nunez et al., 1997). This property is particularly relevant to connectivity analyses, which seek to detect and characterize active interactions between brain regions. Therefore, meaningful connectivity patterns can be derived only from measures of cortical source activities and not directly from EEG channel activities (Michel and Murray, 2012). However, this poses a serious problem, since estimating the nature, number, brain (or non-brain) locations, and time courses of the active sources contributing to the scalp EEG is not straightforward (Baillet et al., 2001). Several methods have been proposed to estimate source activities from multi-channel EEG recordings, thereby removing the confounding effects of volume conduction. These methods can be grouped into three categories: (a) simple spatial filters that seek to reduce correlations between scalp channels based on idealized assumptions; (b) more complex spatial filters that seek to estimate net activities within ROIs based on detailed neurophysiological head models; and (c) blind spatial source separation methods that seek to separately identify source signals by exploiting source signal information differences. Whereas simple spatial filters such as bipolar derivations and Laplacian filters can reduce, to some extent, correlations among scalp-recorded channels (Fisch, 2012), more sophisticated spatial filtering methods use inverse imaging methods to estimate the time courses of cortical sources in given or estimated region of interest (ROI) (Baillet et al., 2001). Blind source separation techniques, in particular independent component analysis, by contrast, learn spatial filters from the EEG time courses that separate the data into constituent independent source activities. Their corresponding brain (or non-brain) locations can then be estimated using neurophysiological inverse imaging methods (Makeig et al., 1996; Jung et al., 2001; Delorme et al., 2012). Vector autoregressive (VAR) models are versatile tools for analyzing multivariate time series, including multi-channel EEG or multivariate source activities. VAR models predict current values of time series from their recent past (Lutkepohl, 2005). Importantly, they can be used to derive various electrophysiological connectivity measures (Schlogl and Supp, 2006). Volume conduction in biological tissue can be modeled as instantaneous propagation of activity from sources to recording channels. The resulting zero-phase connectivity may be treated as noise added to lagged connectivity patterns of interest. Although some measures, including the imaginary part of the coherency (Nolte et al., 2004), are insensitive to zero-phase connectivity, measures derived from VAR model coefficients do not include such zero-phase terms. Thus, volume conduction effects are not accounted for by the model and affect the correlation structure of the model residuals, which are normally assumed to be uncorrelated. Popular connectivity measures derived from VAR models include the Directed Transfer Function (DTF) (Kaminski and Blinowska, 1991) and the Partial Directed Coherence (PDC) (Baccala and Sameshima, 2001). Whereas the PDC is defined in terms of the system matrix (a frequency domain representation of the VAR model), the DTF is based on the inverse of the system matrix. Viewing lagged dependencies between source signals as information flow, the DTF may be said to be normalized by the inflow of information to some sink, while the PDC is normalized by the outflow of information from some source. As we will demonstrate below, both the DTF and the PDC are indeed adversely affected by volume conduction from multiple sources to the scalp electrodes, in contrast to the claim of Kaminski and Blinowska in their recent opinion article (Kaminski and Blinowska, 2014). Thus, in general direct application of connectivity measures to scalp EEG signals produces less than accurate results and also does not allow their clear interpretation in terms of underlying source dynamics.

In the recent past, blind source separation (BSS) algorithms using multivariate statistical data analysis technique have been successfully used for source identification and separation in the field of biomedical and statistical signal processing. Recently numbers of different BSS techniques have been developed. With BSS methods being the feasible method for source separation and decomposition of biosignals, it is important to compare the different techniques and determine the most suitable method for the applications. This paper presents the performance of five BSS algorithms (SOBI, TDSEP, FastICA, JADE and Infomax) for decomposition of sEMG to identify subtle finger movements. It is observed that BSS algorithms based on second-order statistics (SOBI and TDSEP) gives better performance compared to algorithms based on higher-order statistics (FastICA, JADE and infomax).

In the last years Independent Component Analysis (ICA) has been applied with success in signal processing and many algorithms have been developed in order to perform ICA. In this paper we review some algorithms, like INFOMAX (Bell and Sejnowski 1995), extended-INFOMAX (Lee, Girolami and Sejniowski 1997), FastICA (OjA, and Hyvarinen 1999), that solve the ICA problem under the assumption of the linear mixture model. We also show an overview of the nonlinear ICA algorithms and we discuss the MISEP (Almeida 2003). In order to test the performances of the reviewed algorithms, we present some applications of ICA in biomedical signal processing. In particular the application of ICA to the electroencephalographic (EEG) and surface electromyographic (sEMG) recordings are shown.

This paper presents a single sensor based blind source separation approach, namely, the wavelet-assisted stationary subspace analysis (WSSA), for gearbox fault diagnosis in a wind turbine. Continuous wavelet transform (CWT) is used as a preprocessing tool to decompose a single sensor measurement data into a set of wavelet coefficients to meet the multidimensional requirement of the stationary subspace analysis (SSA). The SSA is a blind source separation technique that can separate the multidimensional signals into stationary and nonstationary source components without the need for independency and prior information of the source signals. After that, the separated nonstationary source component with the maximum kurtosis value is analyzed by the enveloping spectral analysis to identify potential fault-related characteristic frequencies. Case studies performed on a wind turbine gearbox test system verify the effectiveness of the WSSA approach and indicate that it outperforms independent component analysis (ICA) and empirical mode decomposition (EMD), as well as the spectral-kurtosis-based enveloping, for wind turbine gearbox fault diagnosis.

We address the problem of Blind Source Separation (BSS) of superimposed signals in situations where one signal has constant or slowly varying intensities at some consecutive locations and at the corresponding locations the other signal has highly varying intensities. Independent Component Analysis (ICA) is a major technique for Blind Source Separation and the existing ICA algorithms fail to estimate the original intensities in the stated situation. We combine the advantages of existing sparse methods and Kernel ICA in our technique, by proposing wavelet packet based sparse decomposition of signals prior to the application of Kernel ICA. Simulations and experimental results illustrate the effectiveness and accuracy of the proposed approach. The approach is general in the way that it can be tailored and applied to a wide range of BSS problems concerning one-dimensional signals and images (two-dimensional signals).

Today, it is generally accepted that current methods for biophysical antenatal surveillance do not facilitate a comprehensive and reliable assessment of foetal well-being and that continuing research into alternative methods is necessary to improve antenatal monitoring procedures. In our research, attention has been paid to the abdominal phonogram, a signal that is recorded by positioning an acoustic sensor on the maternal womb and contains valuable information about foetal status, but which is hidden by maternal and environmental sources. To recover such information, previous work has used single-channel independent component analysis (SCICA) on the abdominal phonogram and successfully retrieved estimates of the foetal phonocardiogram, the maternal phonocardiogram, the maternal respirogram and noise. The availability of these estimates made it possible for the current study to focus on their evaluation as sources for antenatal surveillance purposes. To this end, the foetal heart rate (FHR), the foetal heart sounds morphology, the maternal heart rate (MHR) and the maternal breathing rate (MBR) were collected from the estimates retrieved from a dataset of 25 abdominal phonograms. Next, these parameters were compared with reference values to quantify the significance of the physiological information extracted from the estimates. As a result, it has been seen that the instantaneous FHR, the instantaneous MHR and the MBR collected from the estimates consistently followed the trends given by the reference signals, which is a promising outcome for this preliminary study. Thus, as far as this study has gone, it can be said that the independent traces retrieved by SCICA from the abdominal phonogram are likely to become valuable sources of information for well-being surveillance, both foetal and maternal.

Massive multiple input multiple output (MIMO) is the key technology in the 5G communication network. Blind source separation (BSS) plays an important role in MIMO communication when information about the transmitted signals or channel is not available. Independent component analysis (ICA) is one of the well-known BSS technique applied in MIMO wireless communication. Existing BSS techniques like ICA assumes number of receiving antennas to be greater than or equal to the number of transmit antennas and hence cannot be applied to multiple input single output (MISO) communication scenario, that happens during downlink massive MIMO communication. In single channel blind source separation (SCBSS) the source signals are retrieved from a single observed mixture (single channel or single receiving antenna) of the source signals. In this paper, we investigate the application of SCBSS to MISO communication system. Using ensemble empirical mode decomposition (EEMD), the MISO communication system is converted to a pseudo MIMO system followed by the application of principal component analysis (PCA) and ICA algorithms to separate the transmitted signals. The algorithm is successfully applied to 2 X 1 MISO system and with a slight modification to n X 1 MISO system. Simulation results are provided to validate the performance of the algorithm.

Recently, independent component analysis (ICA) is the most popular and promising statistical technique for blind audio source separation. This paper proposes the minimum beta-divergence based ICA as an adaptive robust audio source separation algorithm. This algorithm explores local structures of audio source signals in which the observed signals follow a mixture of several ICA models. The performance of this algorithm is equivalent to the standard ICA algorithms if observed signals are not corrupted by outliers and there exist only one structure of audio source signals in the entire data space, while it keeps better performance otherwise. It is able to extract all local audio source structures sequentially in presence of huge outliers. Our experimental results also support the above statements.

During acquisition or transmission of the Electrocardiogram, noises generated from the surrounded electrical equipment, the patient's motion, movement of the electrodes, or contraction of the muscle around the heart usually interfere with the obtained signal. The interference of these noises in the frequency domain may mask the desired signal and obstruct the diagnosis process. Blind Source Separation techniques and Model-based filtering methods have shown promising results in ECG signal processing. This work pointed to assess the performance of Independent Component Analysis and Extended Kalman Filter in removing the most common ECG noise, such as muscle contraction, baseline shift, and electrode motion artifact. Testing has been executed on a formed signal set by adding noises from the MIT noise stress test database to signals from the MIT-BIH arrhythmia database at a different signal to noise ratio. Performance comparison demonstrates that both techniques show satisfying results in muscle artifact filtering, while ICA based filtration is more accurate than EKF in reducing baseline wander and electrode movement artifacts.

We present a three stage algorithm for adaptive subtraction of free surface multiples, a processing step made necessary in the free-surface multiple elimination method, for example, by the absence of any of its deterministic prerequisites (knowledge of source wavelet, deghosted data, etc.). First, we construct multiple orders from the free surface multiple prediction formula. Each order contains unique information about the data. Second, we use the full recording duration of any given data trace to construct filters that attempt to match the data and the multiple predictions. This filter produces good phase results, but because of the order by order nature of the free surface algorithm, results that are still insufficient for straightforward subtraction. Instead, third, we construct, trace-by-trace, a mixing model where the mixtures are the data trace and its orders of multiple predictions. Corresponding to the mixtures, there are sources and a mixing process, both of which we find through a blind source separation technique, in particular by employing independent component analysis. One of the recovered sources is the desired signal. That is, it is the data without free surface multiples. This side-steps the subtraction inherent to most adaptive subtraction methods, and instead separates the desired signal from the free surface multiples.

Surface electromyography (EMG) can be used in the rehabilitation of musculoskeletal disorders, for evaluating the coordination of muscles that stabilize a joint, or as an objective tool in assessing muscle fatigue. This latter may be achieved by evaluating the low-frequency content of the EMG. However, the electrocardiogram (ECG) interference is also recorded when EMG is acquired close to the heart. It may mask or even modify the information of interest in the EMG. Different signal processing techniques have been proposed to eliminate ECG artifacts from the EMG signals, the high-pass filter being the most used one. Nevertheless, this classic filtering approach would also attenuate EMG activities below 30 Hz, which contain information from low-intensity muscle contractions. This work addresses the automatic ECG attenuation when multichannel EMG is collected on the left pectoralis major muscle during a muscle fatigue test. The proposed ECG artifact mitigation extends a successful automatic template subtraction (TS) approach, which does not require an extra reference channel and is now applied to independent EMG source signal components extracted using a blind source separation technique (ICA, Independent Component Analysis). The automatic detection of ECG components was performed through two alternative measures: entropy and Kullback-Leibler divergence. While the ECG interference seems to hamper the detection of muscle fatigue in the low-contraction regime, the association ICA+TS preserved better the EMG low-frequency content, and the entropy-based automatic detection was found to be more suitable, avoiding possible errors that might arise from manual detection procedures.