Blind source separation using temporal predictability

Brain Connectivity Studies on Structure-Function Relationships: A Short Survey with an Emphasis on Machine Learning

This short survey reviews the recent literature on the relationship between the brain structure and its functional dynamics. Imaging techniques such as diffusion tensor imaging (DTI) make it possible to reconstruct axonal fiber tracks and describe the structural connectivity (SC) between brain regions. By measuring fluctuations in neuronal activity, functional magnetic resonance imaging (fMRI) provides insights into the dynamics within this structural network. One key for a better understanding of brain mechanisms is to investigate how these fast dynamics emerge on a relatively stable structural backbone. So far, computational simulations and methods from graph theory have been mainly used for modeling this relationship. Machine learning techniques have already been established in neuroimaging for identifying functionally independent brain networks and classifying pathological brain states. This survey focuses on methods from machine learning, which contribute to our understanding of functional interactions between brain regions and their relation to the underlying anatomical substrate.

Sparse and Random Sampling Techniques for High-Resolution, Full-Field, BSS-Based Structural Dynamics Identification from Video

Video-based techniques for identification of structural dynamics have the advantage that they are very inexpensive to deploy compared to conventional accelerometer or strain gauge techniques. When structural dynamics from video is accomplished using full-field, high-resolution analysis techniques utilizing algorithms on the pixel time series such as principal components analysis and solutions to blind source separation the added benefit of high-resolution, full-field modal identification is achieved. An important property of video of vibrating structures is that it is particularly sparse. Typically video of vibrating structures has a dimensionality consisting of many thousands or even millions of pixels and hundreds to thousands of frames. However the motion of the vibrating structure can be described using only a few mode shapes and their associated time series. As a result, emerging techniques for sparse and random sampling such as compressive sensing should be applicable to performing modal identification on video. This work presents how full-field, high-resolution, structural dynamics identification frameworks can be coupled with compressive sampling. The techniques described in this work are demonstrated to be able to recover mode shapes from experimental video of vibrating structures when 70% to 90% of the frames from a video captured in the conventional manner are removed.

Separation Enhancement of Power Line Noise from Human ECG Signal Based on Stone Technique

The cardiac signal is very important for the heart disease diagnosis and evaluation. The noise cancelation represent one of the most preprocessing step in ECG signal processing, usually, this signal is very sensitive and varies with time. The ECG signal is mostly contaminated by different signals like Power line noise signal, Baseline signal and muscle signal. The power line interference signal is the most effected signal on the ECG during data recording. Several papers try to cancel the noise based on different ways and to extract the useful information. In this paper a novel approach based on stone blind source extraction is used to extract the pure ECG signal from raw ECG, the main advantage of the proposed approach compared with the classical technique is to separate all the useful information without filtering or cancelling the suitable data from the recording signal. Real ECG data from MIT-BIH databases is taken and the MATLAB program is used to evaluate the experimental results. The performance of the proposed approach is measured based on SNR and MSE. The main contribution of this paper is to use Stone blind source separation technique as a first time in ECG signal analysis and prove that this method is the best technique compared with conventional ways. The obtained result proves Stone BSS technique is very efficient to remove the power line noise.

Sequential Subspace Clustering via Temporal Smoothness for Sequential Data Segmentation

This paper develops a novel sequential subspace clustering method for sequential data. Inspired by the state-of-the-art methods, ordered subspace clustering, and temporal subspace clustering, we design a novel local temporal regularization term based on the concept of temporal predictability. Through minimizing the short-term variance on historical data, it can recover the temporal smoothness relationships in sequential data. Moreover, we claim that the local temporal regularization is more important than the global structural regularization for a specific task, such as sequential subspace clustering, which leads to a concise minimization objective function. To solve the bi-convex objective function, a simple and efficient optimization algorithm based on the alternate convex search method is devised to jointly learn the coding matrix and the dictionary. Furthermore, five baseline methods are also devised for comparison with our proposed method from different aspects. Extensive experimental results and comparisons with the state-of-the-art methods on three data sets demonstrate the effectiveness of the proposed temporal smoothness sequential subspace clustering method for sequential data.

Single-Channel Sparse Non-Negative Blind Source Separation Method for Automatic 3-D Delineation of Lung Tumor in PET Images

In this paper, we propose a novel method for single-channel blind separation of nonoverlapped sources and, to the best of our knowledge, apply it for the first time to automatic segmentation of lung tumors in positron emission tomography (PET) images. Our approach first converts a 3-D PET image into a pseudo-multichannel image. Afterward, regularization free sparseness constrained non-negative matrix factorization is used to separate tumor from other tissues. By using complexity based criterion, we select tumor component as the one with minimal complexity. We have compared the proposed method with threshold based on 40% and 50% maximum standardized uptake value (SUV), graph cuts (GC), random walks (RW), and affinity propagation (AP) algorithms on 18 nonsmall cell lung cancer datasets with respect to ground truth (GT) provided by two radiologists. Dice similarity coefficient averaged with respect to two GTs is: 0.78 ± 0.12 by the proposed algorithm, 0.78 ± 0.1 by GC, 0.77 ± 0.13 by AP, 0.77 ± 0.07 by RW, and 0.75 ± 0.13 by 50% maximum SUV threshold. Since the proposed method achieved performance comparable with interactive methods, considering the unique challenges of lung tumor segmentation from PET images, our findings support possibility of using our fully automated method in routine clinics. The source codes will be available at www.mipav.net/English/research/research.html.

Robustness improvement of hyperspectral image unmixing by spatial second-order regularization

The acquisition of hundreds of images of a scene, each at a different wavelength, is known as hyperspectral imaging. This high amount of data allows the extraction of much more information from hyperspectral images compared with conventional color images. The forward-looking imaging approach emerged from remote sensing, but is still not very widespread in industrial and other practical applications. Spectral unmixing, in particular, aims at the determination of the components present in a scene as well as the abundance to which each component contributes. This information is valuable, for instance, when discrimination tasks are to be performed. Involving not only spectral, but also spatial information was found to have the potential to improve the unmixing results. Several publications use spatial first-order regularization (closely related to the total variation approach) to incorporate this spatial information. Like in classical image processing, this approach favors piecewise constant pixel transitions. This is why it was proposed in the literature to use second-order regularization instead of first order to approach piecewise-linear transitions. Therefore, we introduce Hessian-based regularization to hyperspectral unmixing and propose an algorithm to calculate the regularized result. We use simulated data and images measured in our laboratory to show that both the first- and second-order approaches share many properties and produce similar results. The second-order approach, however, is more robust and thus more accurate in finding the minimum. Both methods smoothen the images in the case of supervised unmixing (i.e., the component spectra are known beforehand) and enhance unsupervised unmixing (when the spectra are not known).

Recursive approach of EEG-segment-based principal component analysis substantially reduces cryogenic pump artifacts in simultaneous EEG-fMRI data

Electroencephalography (EEG) data simultaneously acquired with functional magnetic resonance imaging (fMRI) data are preprocessed to remove gradient artifacts (GAs) and ballistocardiographic artifacts (BCAs). Nonetheless, these data, especially in the gamma frequency range, can be contaminated by residual artifacts produced by mechanical vibrations in the MRI system, in particular the cryogenic pump that compresses and transports the helium that chills the magnet (the helium-pump). However, few options are available for the removal of helium-pump artifacts. In this study, we propose a recursive approach of EEG-segment-based principal component analysis (rsPCA) that enables the removal of these helium-pump artifacts. Using the rsPCA method, feature vectors representing helium-pump artifacts were successfully extracted as eigenvectors, and the reconstructed signals of the feature vectors were subsequently removed. A test using simultaneous EEG-fMRI data acquired from left-hand (LH) and right-hand (RH) clenching tasks performed by volunteers found that the proposed rsPCA method substantially reduced helium-pump artifacts in the EEG data and significantly enhanced task-related gamma band activity levels (p=0.0038 and 0.0363 for LH and RH tasks, respectively) in EEG data that have had GAs and BCAs removed. The spatial patterns of the fMRI data were estimated using a hemodynamic response function (HRF) modeled from the estimated gamma band activity in a general linear model (GLM) framework. Active voxel clusters were identified in the post-/pre-central gyri of motor area, only from the rsPCA method (uncorrected p<0.001 for both LH/RH tasks). In addition, the superior temporal pole areas were consistently observed (uncorrected p<0.001 for the LH task and uncorrected p<0.05 for the RH task) in the spatial patterns of the HRF model for gamma band activity when the task paradigm and movement were also included in the GLM.

Nonstationary source separation using sequential and variational Bayesian learning

Independent component analysis (ICA) is a popular approach for blind source separation where the mixing process is assumed to be unchanged with a fixed set of stationary source signals. However, the mixing system and source signals are nonstationary in real-world applications, e.g., the source signals may abruptly appear or disappear, the sources may be replaced by new ones or even moving by time. This paper presents an online learning algorithm for the Gaussian process (GP) and establishes a separation procedure in the presence of nonstationary and temporally correlated mixing coefficients and source signals. In this procedure, we capture the evolved statistics from sequential signals according to online Bayesian learning. The activity of nonstationary sources is reflected by an automatic relevance determination, which is incrementally estimated at each frame and continuously propagated to the next frame. We employ the GP to characterize the temporal structures of time-varying mixing coefficients and source signals. A variational Bayesian inference is developed to approximate the true posterior for estimating the nonstationary ICA parameters and for characterizing the activity of latent sources. The differences between this ICA method and the sequential Monte Carlo ICA are illustrated. In the experiments, the proposed algorithm outperforms the other ICA methods for the separation of audio signals in the presence of different nonstationary scenarios.

Analysis of functional networks involved in motor execution and motor imagery using combined hierarchical clustering analysis and independent component analysis

Cognitive experiments involving motor execution (ME) and motor imagery (MI) have been intensively studied using functional magnetic resonance imaging (fMRI). However, the functional networks of a multitask paradigm which include ME and MI were not widely explored. In this article, we aimed to investigate the functional networks involved in MI and ME using a method combining the hierarchical clustering analysis (HCA) and the independent component analysis (ICA). Ten right-handed subjects were recruited to participate a multitask experiment with conditions such as visual cue, MI, ME and rest. The results showed that four activation clusters were found including parts of the visual network, ME network, the MI network and parts of the resting state network. Furthermore, the integration among these functional networks was also revealed. The findings further demonstrated that the combined HCA with ICA approach was an effective method to analyze the fMRI data of multitasks.

On blind separability based on the temporal predictability method

This letter discusses blind separability based on temporal predictability (Stone, 2001 ; Xie, He, & Fu, 2005 ). Our results show that the sources are separable using the temporal predictability method if and only if they have different temporal structures (i.e., autocorrelations). Consequently, the applicability and limitations of the temporal predictability method are clarified. In addition, instead of using generalized eigendecomposition, we suggest using joint approximate diagonalization algorithms to improve the robustness of the method. A new criterion is presented to evaluate the separation results. Numerical simulations are performed to demonstrate the validity of the theoretical results.

Independent Component Analysis Separates Spikes of Different Origin in the EEG

Independent component analysis (ICA) is a novel system that finds independent sources in recorded signals. Its usefulness in separating epileptiform activity of different origin has not been determined. The goal of this study was to demonstrate that ICA is useful for separating different spikes using samples of EEG of patients with focal epilepsy. Digital EEG samples from four patients with focal epilepsy were included. The patients had temporal (n = 2), centrotemporal (n = 1) or frontal spikes (n = 1). Twenty-six samples with two (or more) spikes from two different patients were created. The selection of the two spikes for each mixed EEG was performed randomly, trying to have all the different combinations and rejecting the mixture of two spikes from the same patient. Two different examiners studied the EEGs using ICA with JADE paradigm in Matlab platform, trying to separate and to identify the spikes. They agreed in the correct separation of the spikes in 24 of the 26 samples, classifying the spikes as frontal, temporal or centrotemporal, left or right sided. The demonstration of the possibility of detecting different artificially mixed spikes confirms that ICA may be useful in separating spikes or other elements in real EEGs.

Independent component analysis: an introduction

Independent component analysis (ICA) is a method for automatically identifying the underlying factors in a given data set. This rapidly evolving technique is currently finding applications in analysis of biomedical signals (e.g. ERP, EEG, fMRI, optical imaging), and in models of visual receptive fields and separation of speech signals. This article illustrates these applications, and provides an informal introduction to ICA.

Unified SPM–ICA for fMRI analysis

A widely used tool for functional magnetic resonance imaging (fMRI) data analysis, statistical parametric mapping (SPM), is based on the general linear model (GLM). SPM therefore requires a priori knowledge or specific assumptions about the time courses contributing to signal changes. In contradistinction, independent component analysis (ICA) is a data-driven method based on the assumption that the causes of responses are statistically independent. Here we describe a unified method, which combines ICA, temporal ICA (tICA), and SPM for analyzing fMRI data. tICA was applied to fMRI datasets to disclose independent components, whose number was determined by the Bayesian information criterion (BIC). The resulting components were used to construct the design matrix of a GLM. Parameters were estimated and regionally-specific statistical inferences were made about activations in the usual way. The sensitivity and specificity were evaluated using Monte Carlo simulations. The receiver operating characteristic (ROC) curves indicated that the unified SPM-ICA method had a better performance. Moreover, SPM-ICA was applied to fMRI datasets from twelve normal subjects performing left and right hand movements. The areas identified corresponded to motor (premotor, sensorimotor areas and SMA) areas and were consistently task related. Part of the frontal lobe, parietal cortex, and cingulate gyrus also showed transiently task-related responses. The unified method requires less supervision than the conventional SPM and enables classical inference about the expression of independent components. Our results also suggest that the method has a higher sensitivity than SPM analyses.

A Note on Stone's Conjecture of Blind Signal Separation

Stone's method is one of the novel approaches to the blind source separation (BSS) problem and is based on Stone's conjecture. However, this conjecture has not been proved. We present a simple simulation to demonstrate that Stone's conjecture is incorrect. We then modify Stone's conjecture and prove this modified conjecture as a theorem, which can be used a basis for BSS algorithms.

Independent component analysis removing artifacts in ictal recordings

PURPOSE

Independent component analysis (ICA) is a novel algorithm able to separate independent components from complex signals. Studies in interictal EEG demonstrate its usefulness to eliminate eye, muscle, 50-Hz, electrocardiogram (ECG), and electrode artifacts. The goal of this study was to evaluate the usefulness of ICA in removing artifacts in ictal recordings with a known EEG onset.

METHODS

We studied 20 seizures of nine patients with focal epilepsy monitored in our video-EEG monitoring unit. ICA was applied to remove obvious artifacts in segments at the beginning of the seizure. The final EEGs were exported to the original format and were compared with the original EEG by two blinded examiners. We compared original recordings and the samples cleaned by digital filters (DFs), ICA and ICA plus digital filters (ICA + DFs), evaluating the possibility of finding an ictal pattern, the localization of the onset in area and time, and the global quality of the sample.

RESULTS

All the recordings except one (95%) improved after the use of ICA for the elimination of blinking and other artifacts. Three seizures were found in which in the original recordings did not permit us to detect an ictal pattern, and after ICA + DFs, an ictal onset was evident; in two of them, ICA alone was able to show this pattern. The best results in all the scores were obtained with ICA + DF. ICA was better than DFs. The agreement between the two reviewers was highly significant.

CONCLUSIONS

ICA is useful to remove artifacts from ictal recordings. When applied to ictal recordings, it increases the quality of the recording. In some cases, ICA may be useful to show ictal onsets obscured by artifacts. ICA + DFs obtained the best results regarding removal of the artifacts.

Independent Component Analysis as a Tool to Eliminate Artifacts in EEG: A Quantitative Study

Independent component analysis (ICA) is a novel technique that calculates independent components from mixed signals. A hypothetical clinical application is to remove artifacts in EEG. The goal of this study was to apply ICA to standard EEG recordings to eliminate well-known artifacts, thus quantifying its efficacy in an objective way. Eighty samples of recordings with spikes and evident artifacts of electrocardiogram (EKG), eye movements, 50-Hz interference, muscle, or electrode artifact were studied. ICA components were calculated using the Joint Approximate Diagonalization of Eigen-matrices (JADE) algorithm. The signal was reconstructed excluding those components related to the artifacts. A normalized correlation coefficient was used as a measure of the changes caused by the suppression of these components. ICA produced an evident clearing-up of signals in all the samples. The morphology and the topography of the spike were very similar before and after the removal of the artifacts. The correlation coefficient showed that the rest of the signal did not change significantly. Two examiners independently looked at the samples to identify the changes in the morphology and location of the discharge and the artifacts. In conclusion, ICA proved to be a useful tool to clean artifacts in short EEG samples, without having the disadvantages associated with the digital filters. The distortion of the interictal activity measured by correlation analysis was minimal.

Exploratory fMRI analysis by autocorrelation maximization

A novel and computationally efficient method for exploratory analysis of functional MRI data is presented. The basic idea is to reveal underlying components in the fMRI data that have maximum autocorrelation. The tool for accomplishing this task is Canonical Correlation Analysis. The relation to Principal Component Analysis and Independent Component Analysis is discussed and the performance of the methods is compared using both simulated and real data.

Spatiotemporal independent component analysis of event-related fMRI data using skewed probability density functions

We introduce two independent component analysis (ICA) methods, spatiotemporal ICA (stICA) and skew-ICA, and demonstrate the utility of these methods in analyzing synthetic and event-related fMRI data. First, stICA simultaneously maximizes statistical independence over both time and space. This contrasts with conventional ICA methods, which maximize independence either over time only or over space only; these methods often yield physically improbable solutions. Second, skew-ICA is based on the assumption that images have skewed probability density functions (pdfs), an assumption consistent with spatially localized regions of activity. In contrast, conventional ICA is based on the physiologically unrealistic assumption that images have symmetric pdfs. We combine stICA and skew-ICA, to form skew-stICA, and use it to analyze synthetic data and data from an event-related, left-right visual hemifield fMRI experiment. Results obtained with skew-stICA are superior to those of principal component analysis, spatial ICA (sICA), temporal ICA, stICA, and skew-sICA. We argue that skew-stICA works because it is based on physically realistic assumptions and that the potential of ICA can only be realized if such prior knowledge is incorporated into ICA methods.