Electroencephalographic (EEG) recordings are often contaminated with muscle artifacts. This disturbing myogenic activity not only strongly affects the visual analysis of EEG, but also most surely impairs the results of EEG signal processing tools such as source localization. This article focuses on the particular context of the contamination epileptic signals (interictal spikes) by muscle artifact, as EEG is a key diagnosis tool for this pathology. In this context, our aim was to compare the ability of two stochastic approaches of blind source separation, namely independent component analysis (ICA) and canonical correlation analysis (CCA), and of two deterministic approaches namely empirical mode decomposition (EMD) and wavelet transform (WT) to remove muscle artifacts from EEG signals. To quantitatively compare the performance of these four algorithms, epileptic spike-like EEG signals were simulated from two different source configurations and artificially contaminated with different levels of real EEG-recorded myogenic activity. The efficiency of CCA, ICA, EMD, and WT to correct the muscular artifact was evaluated both by calculating the normalized mean-squared error between denoised and original signals and by comparing the results of source localization obtained from artifact-free as well as noisy signals, before and after artifact correction. Tests on real data recorded in an epileptic patient are also presented. The results obtained in the context of simulations and real data show that EMD outperformed the three other algorithms for the denoising of data highly contaminated by muscular activity. For less noisy data, and when spikes arose from a single cortical source, the myogenic artifact was best corrected with CCA and ICA. Otherwise when spikes originated from two distinct sources, either EMD or ICA offered the most reliable denoising result for highly noisy data, while WT offered the better denoising result for less noisy data. These results suggest that the performance of muscle artifact correction methods strongly depend on the level of data contamination, and of the source configuration underlying EEG signals. Eventually, some insights into the numerical complexity of these four algorithms are given.

No abstract

The Mpl receptor (Mpl-R) is a cytokine receptor belonging to the hematopoietin receptor superfamily for which a ligand has been recently characterized. To study the lineage distribution of Mpl-R in normal hematopoietic cells, we developed a monoclonal antibody (designated M1 MoAb) by immunizing mice with a soluble form of the human Mpl-R protein. With few exceptions, Mpl-R was detected by indirect immunofluorescent analysis on all human leukemic hematopoietic cell lines with pluripotential and megakaryocytic phenotypes, but not on other cell lines. By immunoprecipitation and immunoblotting, M1 MoAb recognized a band at 82 to 84 kD corresponding to the expected size of the glycosylated receptor. Among normal hematopoietic cells, M1 MoAb strongly stained megakaryocytes (MK) and Mpl-R was detected on platelets by indirect immunofluorescence staining or immunoblotting. On purified CD34+ cells, less than 2% of the population was stained, but the labeling was weak and just above the threshold of detection. However, dual-labeling with the M1 and antiplatelet glycoprotein MoAbs showed that most Mpl-R+/CD34+ cells coexpressed CD41a, CD61, or CD42a, suggesting that cell surface appearance of Mpl-R and platelet glycoproteins could be coordinated. M1-positive and M1-negative subsets were sorted from purified CD34+ cell populations. Colony assays showed that the absolute number of hematopoietic progenitors was extremely low and no primitive progenitors were present in the CD34+/Mpl-R+ fraction. However, this cell fraction was significantly enriched in low proliferative colony-forming units-MK. When the CD34+/Mpl-R+ fraction was grown in liquid culture containing human aplastic serum and a combination of growth factors, mature MK were seen as early as day 4, whereas the predominant cell population was erythroblasts on day 8. Similar data were also obtained with the CD34+/Mpl-R- fraction with, however, a delay in the time of appearance of both MK and erythroblasts. In conclusion, Mpl-R is a cytokine receptor restricted to the MK cell lineage. Its expression is low on CD34+ cells and these cells mainly correspond to late MK progenitors and transitional cells. These data indicate that the action of the Mpl-R ligand might predominate during the late stages of human MK differentiation.

A number of application areas such as biomedical engineering require solving an underdetermined linear inverse problem. In such a case, it is necessary to make assumptions on the sources to restore identifiability. This problem is encountered in brain source imaging when identifying the source signals from noisy electroencephalographic or magnetoencephalographic measurements. This inverse problem has been widely studied during the last decades, giving rise to an impressive number of methods using different priors. Nevertheless, a thorough study of the latter, including especially sparse and tensorbased approaches, is still missing. In this paper, we propose i) a taxonomy of the algorithms based on methodological considerations, ii) a discussion of identifiability and convergence properties, advantages, drawbacks, and application domains of various techniques, and iii) an illustration of the performance of selected methods on identical data sets. Directions for future research in the area of biomedical imaging are eventually provided.

SummaryObjective: Non-rapid eye movement (NREM) sleep is known to be a brain state associated with an activation of interictal epileptic activity. The goal of this work was to quantify topographic changes occurring during NREM sleep in comparison with wakefulness. Method: We studied intracerebral recordings of 20 patients who underwent stereo-electroencephalography (SEEG) during presurgical evaluation for pharmacoresistant focal epilepsy. We measured the number of interictal spikes (IS) and quantified the co-occurrence of IS between brain regions during 1 hour of NREM sleep and 1 hour of wakefulness. Co-occurrence is a method to estimate IS networks based on a temporal concordance between IS of different brain regions.Each studied region was labeled as "seizure-onset zone" (SOZ), "propagation zone" (PZ), or "not involved region" (NIR). Results: During NREM sleep, the number of interictal spikes significantly increased in all regions (mean of 68%). This increase was higher in medial temporal regions than in other regions, whether involved in the SOZ. Spike co-occurrence increased significantly in all regions during NREM sleep in comparison with wakefulness but was greater in neocortical regions. Spike co-occurrence in medial temporal regions was not higher than in other regions, suggesting that the increase of the number of spikes in this region was in great part a local effect. Significance: This study demonstrated that medial temporal regions show a greater propensity to spike production or propagation during NREM sleep compared to other brain regions, even when the medial temporal lobe is not involved in the SOZ. K E Y W O R D Sco-occurrence, focal epilepsy, interictal spikes, NREM sleep, stereo-electroencephalography

These findings may contribute to the development of EEG network-based test that could strengthen results obtained from currently-used neurophysiological tests in neurodegenerative diseases.

The brain is a complex, plastic, electrical network whose dysfunctions result in neurological disorders. Multichannel transcranial electrical stimulation (tCS) is a non-invasive neuromodulatory technique with the potential for network-oriented therapy. Challenges to realizing this vision include the proper identification of involved networks in a patient-specific context, a deeper understanding of the effects of stimulation on interconnected neuronal populations-both immediate and plastic-and, based on these, developing strategies to personalize brain stimulation interventions. For this reason, personalized hybrid biophysical and physiological models of brain networks are poised to play a key role in the evolution of networkoriented transcranial stimulation. We review some of the recent work in this emerging area of research and provide an outlook for future modeling and experimental work, as well as for developing its clinical applications in fields such as epilepsy. Research Projects Activity (IARPA), via 2014-13121700007. The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of the ODNI, IARPA, or the U.S. Government. ES is supported by the Beth Israel Deaconess Medical Center (BIDMC) via the Chief Academic Officer (CAO) Award 2017, and the Defense Advanced Research Projects Agency (DARPA) via HR001117S0030. The content of this paper is solely the responsibility of the authors and does not necessarily represent the official views of Harvard University, and its affiliated academic health care centers. DECLARATION OF INTEREST Giulio Ruffini is a shareholder and works for Neuroelectrics, a company designing medical devices for brain stimulation. Roser Sanchez-Todo is a researcher at Neuroelectrics. REFERENCES

Abstract. Independent Component Analysis (ICA) plays an important role in biomedical engineering. Indeed, the complexity of processes involved in biomedicine and the lack of reference signals make this blind approach a powerful tool to extract sources of interest. However, in practice, only few ICA algorithms such as SOBI, (extended) InfoMax and FastICA are used nowadays to process biomedical signals. In this paper we raise the question whether other ICA methods could be better suited in terms of performance and computational complexity. We focus on ElectroEncephaloGraphy (EEG) data denoising, and more particularly on removal of muscle artifacts from interictal epileptiform activity. Assumptions required by ICA are discussed in such a context. Then fifteen ICA algorithms, namely JADE, CoM2, SOBI, SOBI rob , (extended) InfoMax, PICA, two different implementations of FastICA, ERICA, SIMBEC, FOBIUMJAD, TFBSS, ICAR3, FOOBI1 and 4-CANDHAPc are briefly described. Next they are studied in terms of performance and numerical complexity. Quantitative results are obtained on simulated epileptic data generated with a physiologically-plausible model. These results are also illustrated on real epileptic recordings.