Dynamic Modeling of Carbon Dioxide Transport through the Skin Using a Capnometry Wristband.

The development of a capnometry wristband is of great interest for monitoring patients at home. We consider a new architecture in which a non-dispersive infrared (NDIR) optical measurement is located close to the skin surface and is combined with an open chamber principle with a continuous circulation of air flow in the collection cell. We propose a model for the temporal dynamics of the carbon dioxide exchange between the blood and the gas channel inside the device. The transport of carbon dioxide is modeled by convection-diffusion equations. We consider four compartments: blood, skin, the measurement cell and the collection cell. We introduce the state-space equations and the associated transition matrix associated with a Markovian model. We define an augmented system by combining a first-order autoregressive model describing the supply of carbon dioxide concentration in the blood compartment and its inertial resistance to change. We propose to use a Kalman filter to estimate the carbon dioxide concentration in the blood vessels recursively over time and thus monitor arterial carbon dioxide blood pressure in real time. Four performance factors with respect to the dynamic quantification of the CO2 blood concentration are considered, and a simulation is carried out based on data from a previous clinical study. These demonstrate the feasibility of such a technological concept.

Transfer Learning for P300 Brain-Computer Interfaces by Joint Alignment of Feature Vectors.

This article presents a new transfer learning method named group learning, that jointly aligns multiple domains (many-to-many) and an extension named fast alignment that aligns any further domain to previously aligned group of domains (many-to-one). The proposed group alignment algorithm (GALIA) is evaluated on brain-computer interface (BCI) data and optimal hyper-parameter values of the algorithm are studied for classification performance and computational cost. Six publicly available P300 databases comprising 333 sessions from 177 subjects are used. As compared to the conventional subject-specific train/test pipeline, both group learning and fast alignment significantly improve the classification accuracy except for the database with clinical subjects (average improvement: 2.12±1.88%). GALIA utilizes cyclic approximate joint diagonalization (AJD) to find a set of linear transformations, one for each domain, jointly aligning the feature vectors of all domains. Group learning achieves a many-to-many transfer learning without compromising the classification performance on non-clinical BCI data. Fast alignment further extends the group learning for any unseen domains, allowing a many-to-one transfer learning with the same properties. The former method creates a single machine learning model using data from previous subjects and/or sessions, whereas the latter exploits the trained model for an unseen domain requiring no further training of the classifier.

Micro-pursuit: A class of fixational eye movements correlating with smooth, predictable, small-scale target trajectories.

Humans generate ocular pursuit movements when a moving target is tracked throughout the visual field. In this article, we show that pursuit can be generated and measured at small amplitudes, at the scale of fixational eye movements, and tag these eye movements as micro-pursuits. During micro-pursuits, gaze direction correlates with a target's smooth, predictable target trajectory. We measure similarity between gaze and target trajectories using a so-called maximally projected correlation and provide results in three experimental data sets. A first observation of micro-pursuit is provided in an implicit pursuit task, where observers were tasked to maintain their gaze fixed on a static cross at the center of screen, while reporting changes in perception of an ambiguous, moving (Necker) cube. We then provide two experimental paradigms and their corresponding data sets: a first replicating micro-pursuits in an explicit pursuit task, where observers had to follow a moving fixation cross (Cross), and a second with an unambiguous square (Square). Individual and group analyses provide evidence that micro-pursuits exist in both the Necker and Cross experiments but not in the Square experiment. The interexperiment analysis results suggest that the manipulation of stimulus target motion, task, and/or the nature of the stimulus may play a role in the generation of micro-pursuits.

Graph Lasso-Based Test for Evaluating Functional Brain Connectivity in Sickle Cell Disease.

Sickle cell disease (SCD) is a vascular disorder that is often associated with recurrent ischemia-reperfusion injury, anemia, vasculopathy, and strokes. These cerebral injuries are associated with neurological dysfunction, limiting the full developing potential of the patient. However, recent large studies of SCD have demonstrated that cognitive impairment occurs even in the absence of brain abnormalities on conventional magnetic resonance imaging (MRI). These observations support an emerging consensus that brain injury in SCD is diffuse and that conventional neuroimaging often underestimates the extent of injury. In this article, we postulated that alterations in the cerebral connectivity may constitute a sensitive biomarker of SCD severity. Using functional MRI, a connectivity study analyzing the SCD patients individually was performed. First, a robust learning scheme based on graphical lasso model and Fréchet mean was used for estimating a consistent descriptor of healthy brain connectivity. Then, we tested a statistical method that provides an individual index of similarity between this healthy connectivity model and each SCD patient's connectivity matrix. Our results demonstrated that the reference connectivity model was not appropriate to model connectivity for only 4 out of 27 patients. After controlling for the gender, two separate predictors of this individual similarity index were the anemia (p = 0.02) and white matter hyperintensities (WMH) (silent stroke) (p = 0.03), so that patients with low hemoglobin level or with WMH have the least similarity to the reference connectivity model. Further studies are required to determine whether the resting-state connectivity changes reflect pathological changes or compensatory responses to chronic anemia.

Blind Hyperspectral Unmixing Using an Extended Linear Mixing Model to Address Spectral Variability.

Spectral unmixing is one of the main research topics in hyperspectral imaging. It can be formulated as a source separation problem, whose goal is to recover the spectral signatures of the materials present in the observed scene (called endmembers) as well as their relative proportions (called fractional abundances), and this for every pixel in the image. A linear mixture model (LMM) is often used for its simplicity and ease of use, but it implicitly assumes that a single spectrum can be completely representative of a material. However, in many scenarios, this assumption does not hold, since many factors, such as illumination conditions and intrinsic variability of the endmembers, induce modifications on the spectral signatures of the materials. In this paper, we propose an algorithm to unmix hyperspectral data using a recently proposed extended LMM. The proposed approach allows a pixelwise spatially coherent local variation of the endmembers, leading to scaled versions of reference endmembers. We also show that the classic nonnegative least squares, as well as other approaches to tackle spectral variability can be interpreted in the framework of this model. The results of the proposed algorithm on two different synthetic datasets, including one simulating the effect of topography on the measured reflectance through physical modelling, and on two real data sets, show that the proposed technique outperforms other methods aimed at addressing spectral variability, and can provide an accurate estimation of endmember variability along the scene because of the scaling factors estimation.

FUNCTIONAL CONNECTIVITY ANALYSIS FOR THALASSEMIA DISEASE BASED ON A GRAPHICAL LASSO MODEL.

Thalassemia is a congenital disorder of hemoglobin synthesis which can lead to thromboembolic events and stroke in the brain. In this work we propose to use a functional connectivity model to discriminate between control and diseased subjects. Our connectivity measure is based on functional magnetic resonance imaging, and hence common variations of the blood oxygenation level in spatially distant areas. Analyzing this connectivity could highlight abnormal neuronal activation and provide us with a descriptor (bio-marker) of the disease. To estimate the connectivity, we propose a robust learning scheme based on the graphical lasso model, whose hyperparameter is validated within a cross-validation scheme. To analyze model fit, we transfer the mean connectivity from the control group to the thalassemic patient group. Our null hypothesis is that the model learned on control subjects is perfectly adequate (in the maximum likelihood sense) to describe the patients. The results of the permutation test suggest that the some patients with thalassemia do not have the same connectivity structure as the control.

Independent Vector Analysis for Gradient Artifact Removal in Concurrent EEG-fMRI Data.

We consider the problem of removing gradient artifact from electroencephalogram (EEG) signal, recorded concurrently with functional magnetic resonance imaging (fMRI) acquisition. We estimate the artifact by exploiting its quasi-periodicity over the epochs and its similarity over the different channels by using independent vector analysis, a recent extension of independent component analysis for multiple datasets. The method fully makes use of the spatio-temporal information by using spatial dependences across channels to estimate the artifact for a particular channel. Thus, it provides robustness with respect to uncontrollable changes such as head movement and fluctuations in the B0 field during the acquisition. Results using both simulated data with gradient artifact and EEG data collected concurrently with fMRI show the desirable performance of the new method.

Deriving a multi-subject functional-connectivity atlas to inform connectome estimation.

The estimation of functional connectivity structure from functional neuroimaging data is an important step toward understanding the mechanisms of various brain diseases and building relevant biomarkers. Yet, such inferences have to deal with the low signal-to-noise ratio and the paucity of the data. With at our disposal a steadily growing olume of publicly available neuroimaging data, it is however possible to improve the estimation procedures involved in connectome mapping. In this work, we propose a novel learning scheme for functional connectivity based on sparse Gaussian graphical models that aims at minimizing the bias induced by the regularization used in the estimation, by carefully separating the estimation of the model support from the coefficients. Moreover, our strategy makes it possible to include new data with a limited computational cost. We illustrate the physiological relevance of he learned prior, that can be identified as a functional connectivity atlas, based on an experiment on 46 subjects of the Human Connectome Dataset.

Impact of autocorrelation on functional connectivity.

Although the impact of serial correlation (autocorrelation) in residuals of general linear models for fMRI time-series has been studied extensively, the effect of autocorrelation on functional connectivity studies has been largely neglected until recently. Some recent studies based on results from economics have questioned the conventional estimation of functional connectivity and argue that not correcting for autocorrelation in fMRI time-series results in "spurious" correlation coefficients. In this paper, first we assess the effect of autocorrelation on Pearson correlation coefficient through theoretical approximation and simulation. Then we present this effect on real fMRI data. To our knowledge this is the first work comprehensively investigating the effect of autocorrelation on functional connectivity estimates. Our results show that although FC values are altered, even following correction for autocorrelation, results of hypothesis testing on FC values remain very similar to those before correction. In real data we show this is true for main effects and also for group difference testing between healthy controls and schizophrenia patients. We further discuss model order selection in the context of autoregressive processes, effects of frequency filtering and propose a preprocessing pipeline for connectivity studies.

Dynamic changes of spatial functional network connectivity in healthy individuals and schizophrenia patients using independent vector analysis.

Recent work on both task-induced and resting-state functional magnetic resonance imaging (fMRI) data suggests that functional connectivity may fluctuate, rather than being stationary during an entire scan. Most dynamic studies are based on second-order statistics between fMRI time series or time courses derived from blind source separation, e.g., independent component analysis (ICA), to investigate changes of temporal interactions among brain regions. However, fluctuations related to spatial components over time are of interest as well. In this paper, we examine higher-order statistical dependence between pairs of spatial components, which we define as spatial functional network connectivity (sFNC), and changes of sFNC across a resting-state scan. We extract time-varying components from healthy controls and patients with schizophrenia to represent brain networks using independent vector analysis (IVA), which is an extension of ICA to multiple data sets and enables one to capture spatial variations. Based on mutual information among IVA components, we perform statistical analysis and Markov modeling to quantify the changes in spatial connectivity. Our experimental results suggest significantly more fluctuations in patient group and show that patients with schizophrenia have more variable patterns of spatial concordance primarily between the frontoparietal, cerebellar and temporal lobe regions. This study extends upon earlier studies showing temporal connectivity differences in similar areas on average by providing evidence that the dynamic spatial interplay between these regions is also impacted by schizophrenia.

Independent component analysis for brain FMRI does indeed select for maximal independence.

A recent paper by Daubechies et al. claims that two independent component analysis (ICA) algorithms, Infomax and FastICA, which are widely used for functional magnetic resonance imaging (fMRI) analysis, select for sparsity rather than independence. The argument was supported by a series of experiments on synthetic data. We show that these experiments fall short of proving this claim and that the ICA algorithms are indeed doing what they are designed to do: identify maximally independent sources.

A non-orthogonal SVD-based decomposition for phase invariant error-related potential estimation.

The estimation of the Error Related Potential from a set of trials is a challenging problem. Indeed, the Error Related Potential is of low amplitude compared to the ongoing electroencephalographic activity. In addition, simple summing over the different trials is prone to errors, since the waveform does not appear at an exact latency with respect to the trigger. In this work, we propose a method to cope with the discrepancy of these latencies of the Error Related Potential waveform and offer a framework in which the estimation of the Error Related Potential waveform reduces to a simple Singular Value Decomposition of an analytic waveform representation of the observed signal. The followed approach is promising, since we are able to explain a higher portion of the variance of the observed signal with fewer components in the expansion.

EEG sensor selection by sparse spatial filtering in P300 speller brain-computer interface.

A Brain-Computer Interface (BCI) is a specific type of human-machine interface that enables communication between a subject/patient and a computer by direct control from decoding of brain activity. This paper deals with the P300-speller application that enables to write a text based on the oddball paradigm. To improve the ergonomics and minimize the cost of such a BCI, reducing the number of electrodes is mandatory. We propose a new algorithm to select a relevant subset of electrodes by estimating sparse spatial filters. A l(1)-norm penalization term, as an approximation of the l(0)-norm, is introduced in the xDAWN algorithm, which maximizes the signal to signal-plus-noise ratio. Experimental results on 20 subjects show that the proposed method is efficient to select the most relevant sensors: from 32 down to 10 sensors, the loss in classification accuracy is less than 5%.

An extension of the Canonical Correlation Analysis to the case of multiple observations of two groups of variables.

In this contribution we present a method that extends the Canonical Correlation Analysis for two groups of variables to the case of multiple conditions. Contrary to the extensions in literature based on augmenting the number of variable groups, the addition of conditions allows for a more robust estimate of the canonical correlation structure inherently present in the data. Algorithms to solve the estimation problem are based on joint approximate diagonalization algorithms for matrix sets. Simulations show the performance of the proposed method under two different scenarios: the calculation of a latent canonical structure and the estimation of a bilinear mixture model.

A contrast function for independent component analysis without permutation ambiguity.

This brief deals with the problem of blind source separation (BSS) via independent component analysis (ICA). We prove that a linear combination of the separator output fourth-order marginal cumulants (kurtoses) is a valid contrast function for ICA under prewhitening if the weights have the same sign as the source kurtoses. If, in addition, the source kurtoses are different and so are the linear combination weights, the contrast eliminates the permutation ambiguity typical to ICA, as the estimated sources are sorted at the separator output according to their kurtosis values in the same order as the weights. If the weights equal the source kurtoses, the contrast is a cumulant matching criterion based on the maximum-likelihood principle. The contrast can be maximized by means of a cost-efficient Jacobi-type pairwise iteration. In the real-valued two-signal case, the asymptotic variance of the resulting Givens angle estimator is determined in closed form, leading to the contrast weights with optimal finite-sample performance. A fully blind solution can be implemented by computing the optimum weights from the initial source estimates obtained by a classical ICA stage. An experimental study validates the features of the proposed technique and shows its superior performance compared to related previous methods.

Atrial activity estimation from atrial fibrillation ECGs by blind source extraction based on a conditional maximum likelihood approach.

This work presents a spatial filtering method for the estimation of atrial fibrillation activity in the cutaneous electrocardiogram. A linear extraction filter is obtained by maximising the extractor output power on the significant spectral support of the signal of interest. An iterative procedure based on a quasi-maximum likelihood estimator is proposed to jointly estimate the significant spectral support and the extraction filter. Compared with a previously proposed spatio-temporal blind source separation method, our approach yields an improved atrial activity signal estimate as quantified by a higher spectral concentration of the extractor output. The proposed methodology can readily be adapted to signal extraction problems in other application domains.

Eigenvector analysis for separation of a spectrally concentrated source from a mixture.

An objective function is presented to recover a spectrally narrow band signal from multichannel measurements, as in electrocardiogram recordings of atrial fibrillation. The criterion can be efficiently maximized through the eigenvalue decomposition of some spectral correlation matrices of the whitened observations across appropriately chosen frequency bands. It is conjectured that the global optimum so attained recovers the source of interest when its spectral concentration around its modal frequency is maximal. Numerical experiments on synthetic data seem to support the validity of this hypothesis. Moreover, the components extracted from a patient data set with known atrial fibrillation show the characteristics of the associated f-wave as described in medical literature.

The exploitation of spatial topographies for atrial signal extraction in atrial fibrillation ECGs.

The accuracy in the extraction of the atrial activity (AA) from electrocardiogram (ECG) signals recorded during atrial fibrillation (AF) episodes plays an important role in the analysis and characterization of atrial arrhythmias. The present contribution puts forward a method for AA signal extraction based on a blind source separation (BSS) formulation. The latter exploits spatial information on the different components in the ECG related or not to AF. The source directions or spatial topographies of the components not related to AF are used to determine the nullspace of the AA, so that the topographies related to AA become more suitable to describe AF sources. The comparative performance of the method is evaluated on real data recorded from patients with noticeable AF. The AA extraction quality of the proposed technique is comparable to that of previous algorithms.

Cumulant matching for independent source extraction.

In this work we show how one can make use of priors on signal statistics under the form of cumulant guesses to extract an independent source from an observed mixture. The advantage of using statistical priors on the signal lies in the fact that no specific knowledge is needed about its temporal behavior, neither about its spatial distribution. We show that these statistics can be obtained either by reasoning on the theoretical values of a supposed waveform, either by using a subset of the observations from which we know that their statistics are merely hindered by interferences. Results on an electro-cardiographic recording confirm the above assumptions.

Extraction of the atrial activity from the ECG based on independent component analysis with prior knowledge of the source kurtosis signs.

In this work it will be shown that a contrast for independent component analysis based on prior knowledge of the source kurtosis signs (ica-sks) is able to extract atrial activity from the electrocardiogram when a constrained updating is introduced. A spectral concentration measure is used, only allowing signal pair updates when spectral concentration augments. This strategy proves to be valid for independent source extraction with priors on the spectral concentration. Moreover, the method is computationally attractive with a very low complexity compared to the recently proposed methods based on spatiotemporal extraction of the atrial fibrillation signal.