Synchronized oscillations of ensembles of neurons in the brain underlie human cognition and behaviors. Neuronal network oscillations can be described by the physics of coupled dynamical systems. This dissertation examines the dynamic network activities in two distinct neurocognitive networks, the salience network (SN) and the ventral temporal cortex-dorsolateral prefrontal cortex (VTC-DLPFC) network, during perceptual decision-making (PDM). The key nodes of the SN include the right anterior insula (rAI), left anterior insula (lAI), and dorsal anterior cingulate cortex (dACC) in the brain. When and how a sensory signal enters and organizes within the SN before reaching the central executive network including the prefrontal cortex has been a mystery. Second, prior studies also report that perception of visual objects (face and house) involves a network of the VTC—the fusiform face area (FFA) and para-hippocampal place area (PPA)—and the DLPFC. How sensory information enters and organizes within the VTC-DLPFC network is not well understood, in milliseconds time-scale of human’s perception and decision-making. We used clear and noisy face/house image categorization tasks and scalp electroencephalography (EEG) recordings to study the dynamics of these networks. We demonstrated that beta (13–30 Hz) oscillation bound the SN, became most active around 100 ms after the stimulus onset, the rAI acted as a main outflow hub within the SN, and the SN activities were negatively correlated with the difficult tasks. We also uncovered that the VTC-DLPFC network activities were mediated by beta (13-30 Hz) and gamma (30-100 Hz) oscillations. Beta activities were enhanced in the time frame 125-250 ms after stimulus onset, the VTC acted as main outflow hub, and network activities were negatively correlated with the difficult tasks. In contrast, gamma activities were elevated in the time frame 0-125 ms, the DLPFC acted as a main outflow hub, and network activities—specifically the FFA-PPA pair—were positively correlated with the difficult tasks. These findings significantly enhance our understanding of how sensory information enters and organizes within the SN and the VTC-DLPFC network, respectively in PDM.

Η ηλεκτρομαγνητική δραστηριότητα του εγκεφάλου μελετάται με τη βοήθεια των μη παρεμβατικών μεθόδων της Ήλεκτροεγκεφαλογραφίας και της Μαγνητοεγκεφαλογραφίας. Ειδικότερα, κάθε ηλεκτροχημικά παραγόμενο ρεύμα στο εσωτερικό του εγκεφάλου δημιουργεί ένα ηλεκτρικό και ένα μαγνητικό πεδίο, στο εσωτερικό και στο εξωτερικό του εγκεφάλου αντίστοιχα. Τα πεδία αυτά καταγράφονται στην επιφάνεια και στον εξωτερικό χώρο του κρανίου και δίνουν το Ηλεκτροεγκεφαλόγραφημα (EEG) και το Μαγνητοεγκεφαλόγραφημα (MEG) αντίστοιχα, τα οποία μεταφέρουν πληροφορίες για τη λειτουργία του εγκεφάλου τη χρονική στιγμή της καταγραφής. Η παρούσα διατριβή αφορά στη μαθηματική ανάλυση ευθέων και αντίστροφων προβλημάτων που συνδέονται με τις μεθόδους αυτές με σκοπό τον εντοπισμό και το χαρακτηρισμό της πηγής που παρήγαγε τα μετρούμενα πεδία. Στο Μέρος Ι μελετάται αναλυτικά η δομή και λειτουργία του εγκεφάλου, περιγράφεται το φυσικό πρότυπο που χρησιμοποιούμε και γίνεται αναφορά τόσο στη σφαιρική όσο και στην ελλειψοειδή γεωμετρία που αποτελούν τα γεωμετρικά υπόβαθρα. Στο Μέρος ΙΙ επιλύεται το ευθύ πρόβλημα του Βιοηλεκτρισμού στην περίπτωση του σφαιρικού ομογενούς προτύπου για τον ανθρώπινο εγκέφαλο, όπου η πηγή είναι αυθαίρετη κατανομή ρεύματος. Αποδεικνύεται ό,τι, στο εξωτερικό ηλεκτρικό δυναμικό δεν εμπεριέχεται η συνεισφορά του σωληνοειδούς μέρους της εφαπτομενικής συνιστώσας του ρεύματος και συνεπώς το αντίστοιχο αντίστροφο πρόβλημα είναι μη μοναδικό. Με την απαίτηση το ρεύμα να ελαχιστοποιεί την , το αντίστροφο πρόβλημα επιλύεται μοναδικά και προσδιορίζονται οι συνιστώσες του νευρωνικού ρεύματος από γνωστές μετρήσεις του ηλεκτρικού δυναμικού. Τα κύρια χαρακτηριστικά καθώς και οι περιορισμοί που επιβάλλουν το φυσικό και το γεωμετρικό πρόβλημα αναλύονται λεπτομερώς. Στο Μέρος ΙΙΙ επιλύονται ευθέα προβλήματα του Βιοηλεκτρομαγνητισμού σε ελλειψοειδή γεωμετρία και αντλούμε χρήσιμα συμπεράσματα για την αντιστροφή των προβλημάτων MEG. Συγκεκριμένα υπολογίστηκε η οκταπολική προσέγγιση του μαγνητικού πεδίου που παράγεται στο εξωτερικό του πλέον ρεαλιστικού ομογενούς προτύπου για τον ανθρώπινο εγκέφαλο, που είναι το ελλειψοειδές, συναρτήσει των ελλειψοειδών αρμονικών τρίτου βαθμού. Η βελτίωση αυτή είναι σημαντική καθώς αποδεικνύεται αριθμητικά ότι η μαγνητικά «σιωπηλή» πηγή της τετραπολικής προσέγγισης συνεισφέρει στις μετρήσεις του μαγνητικού πεδίου. Ως εκ τούτου, η νέα αυτή προσέγγιση του μαγνητικού πεδίου παρέχει αρκετές πληροφορίες για την πιθανή αντιστροφή του προβλήματος. Στη συνέχεια επιλύθηκε το ευθύ πρόβλημα του Βιοηλεκτρομαγνητισμού στην περίπτωση που ο εγκεφαλικός ιστός περιλαμβάνει περιοχή υγρού πυρήνα διαφορετικής αγωγιμότητας. Ο πυρήνας αυτός πληρούται από εγκεφαλονωτιαίο υγρό ενώ η πηγή βρίσκεται στον φλοιό του εγκεφαλικού ιστού. Υπολογίζεται το ηλεκτρικό δυναμικό και το μαγνητικό πεδίο στο εξωτερικό του αγωγού και τα αποτελέσματα συγκρίνονται αναλυτικά και αριθμητικά με τα αντίστοιχα αποτελέσματα του ομογενούς προτύπου του εγκέφαλου. Από την σύγκριση αυτή προκύπτει ότι τόσο η ανομοιογένεια εντός του εγκεφαλικού ιστού όσο και η θέση της πηγής υπεισέρχονται με καθοριστικό τρόπο στο μαγνητικό πεδίο του υπό μελέτη προτύπου. / The electromagnetic activity of the human brain is studying via the non invasive methods of Electroencephalography and Magnetoencephalography. It is well known that an electrochemically generated current in the interior of the brain generates an electric and a magnetic field, both in the interior and exterior of the brain. The resulting electric and magnetic fields are measured on the surface and the exterior of the head via the EEG and MEG, respectively. In the present thesis we study direct and inverse EEG and MEG problems in order to identify and characterize the source. In the First Part we describe the morphology and the functionality of the human brain and we state the physical and geometrical models that we use. In the Second Part we solved the direct problem of EEG for the spherical homogeneous model of the brain in the case of a continuously distributed neuronal current. It turns out that the electric potential is independent of the solenoid part of the tangential component of the neuronal current. Consequently, the corresponding inverse problem is not uniquely solvable. Hence, we demand that the current has minimum and in this case we ended up with the complete expansions of the visible part of the current from the knowledge of the electric field. In the Third Part we studied direct problems of MEG in ellipsoidal geometry. In particular we evaluated the octapolic term of the magnetic induction field which it’s produced in the exterior of the ellipsoidal model of the brain-head system. This term provides the highest order terms that can be expressed in closed form. It is shown numerically that the silent source of the quadrupolic term of the magnetic induction field does contribute to the octapolic term. Therefore, the knowledge of the quadrupolic and octapolic terms provides enough data to construct an effective algorithm for inversion. Finally, the direct problem of MEG is presented, in the case where the cerebral tissue is considered as an ellipsoidal conductor and surrounds a fluid ellipsoidal core of different conductivity. The fluid core is occupied by the cerebrospinal fluid and the source lies in the cerebral shell. The electric field in every region and the exterior magnetic induction field are obtained. Furthermore, we compare analytically and numerically the results of the inhomogeneous model with the homogeneous ellipsoidal model. We observed that both the inhomogeniety inside the cerebral tissue and the location of the source appear in the magnetic induction field of the inhomogeneous model. Τhe existence of the fluid core effects the monotonicity of the components of the magnetic field as well as its magnitude.

Doctor of Philosophy (PhD) / Schizophrenia is a severe mental illness that affects approximately 1% of the population worldwide, and which typically has a devastating effect on the lives of its sufferers. The characteristic symptoms of the disease include hallucinations, delusions, disorganized thought and reduced emotional expression. While many of the early theories of schizophrenia focused on its psychosocial foundations, more recent theories have focused on the neurobiological underpinnings of the disease. This thesis has four primary aims: 1) to use magnetic resonance imaging (MRI) to identify the structural brain abnormalities present in patients suffering from their first episode of schizophrenia (FES), 2) to elucidate whether these abnormalities were static or progressive over the first 2-3 years of patients’ illness, 3) to identify the relationship between these neuroanatomical abnormalities and patients’ clinical profile, and 4) to identify the normative relationship between longitudinal changes in neuroanatomy and electrophysiology in healthy participants, and to compare this to the relationship observed between these two indices in patients with FES. The aim of Chapter 2 was to use MRI to identify the neuroanatomical changes that occur over adolescence in healthy participants, and to identify the normative relationship between the neuroanatomical changes and electrophysiological changes associated with healthy periadolescent brain maturation. MRI and electroencephalographic (EEG) scans were acquired from 138 healthy participants between the ages of 10 and 30 years. The MRI scans were segmented into grey matter (GM) and white matter (WM) images, before being parcellated into the frontal, temporal, parietal and occipital lobes. Absolute EEG power was calculated for the slow-wave, alpha and beta frequency bands, for the corresponding cortical regions. The age-related changes in regional tissue volumes and regional EEG power were inferred with a regression model. The results indicated that the healthy participants experienced accelerated GM loss, EEG power loss and WM gain in the frontal and parietal lobes between the ages of 10 and 20 years, which decelerated between the ages of 20 and 30 years. A linear relationship was also observed between the maturational changes in regional GM volumes and EEG power in the frontal and parietal lobes. These results indicate that the periadolescent period is a time of great structural and electrophysiological change in the healthy human brain. The aim of Chapter 3 was to identify the GM abnormalities present in patients with FES, both at the time of their first presentation to mental health services (baseline), and over the first 2-3 years of their illness (follow-up). MRI scans were acquired from 41 patients with FES at baseline, and 47 matched healthy control subjects. Of these participants, 25 FES patients and 26 controls returned 2-3 years later for a follow-up scan. The analysis technique of voxel-based morphometry (VBM) was used in conjunction with the Statistical Parametric Mapping (SPM) software package in order to identify the regions of GM difference between the groups at baseline. The related analysis technique of tensor-based morphometry (TBM) was used to identify subjects’ longitudinal GM change over the follow-up interval. Relative to the healthy controls, the FES patients were observed to exhibit widespread GM reductions in the frontal, parietal and temporal cortices and cerebellum at baseline, as well as more circumscribed regions of GM increase, particularly in the occipital lobe. Furthermore, the FES patients lost considerably more GM over the follow-up interval than the controls, particularly in the parietal and temporal cortices. These results indicate that patients with FES exhibit significant structural brain abnormalities very early in the course of their illness, and that these abnormalities progress over the first few years of their illness. Chapter 4 employed the same methodology to investigate the white matter abnormalities exhibited by the FES subjects relative to the controls, both at baseline and over the follow-up interval. Compared to controls, the FES patients exhibited volumetric WM deficits in the frontal and temporal lobes at baseline, as well as volumetric increases at the fronto-parietal junction bilaterally. Furthermore, the FES patients lost considerably more WM over the follow-up interval than did the controls in the middle and inferior temporal cortex bilaterally. While there is substantial evidence indicating that abnormalities in the maturational processes of myelination play a significant role in the development of WM abnormalities in FES, the observed longitudinal reductions in WM were consistent with the death of a select population of temporal lobe neurons over the follow-up interval. The aim of Chapter 5 was to investigate the clinical correlates of the GM abnormalities exhibited by the FES patients at baseline. The volumes of four distinct cerebral regions where 31 patients with FES exhibited reduced GM volumes relative to 30 matched controls were calculated and correlated with patients’ scores on three primary symptom dimensions: Disorganization, Reality Distortion and Psychomotor Poverty. The results indicated that the greater the degree of atrophy exhibited by the FES patients in three of these four ‘regions-of-reduction’, the less severe their degree of Reality Distortion. These results suggest that an excessive amount of GM atrophy may in fact preclude the formation of hallucinations or highly systematized delusions in patients with FES. The aim of Chapter 6 was to identify the relationship between the longitudinal changes in brain structure and brain electrophysiology exhibited by 19 FES patients over the first 2-3 years of their illness, and to compare it to the normative relationship between the two indices reported in Chapter 2. The methodology employed for the parcellation of the MRI and EEG data was identical to Chapter 2. The results indicated that, in contrast to the healthy controls, the longitudinal reduction in GM volume exhibited by the FES patients was not associated with a corresponding reduction in EEG power in any brain lobe. In contrast, EEG power was observed to be maintained or even to increase over the follow-up interval in these patients. These results were consistent with the FES patients experiencing an abnormal elevation of neural synchrony. Such an abnormality in neural synchrony could potentially form the basis of the dysfunctional neural connectivity that has been widely proposed to underlie the functional deficits present in patients with schizophrenia. The primary aim of Chapter 7 was to assimilate the findings from the preceding empirical chapters with the theoretical framework provided in the literature, into an integrated and testable model of schizophrenia. The model emphasized dysfunctions in brain maturation, specifically in the normative processes of synaptic ‘pruning’ and axonal myelination, as playing a key role in the development of disintegrated neural activity and the subsequent onset of schizophrenic symptoms. The model concluded with the novel proposal that disintegrated neural activity arises from abnormal elevations in the synchrony of synaptic activity in patients with first-episode schizophrenia.

Προτείνουμε μια προσέγγιση για να αναλύσουμε τα δεδομένα που συλλέγουμε από το παράδειγμα του Ορθογράφου P300, χρησιμοποιώντας την τεχνική μηχανικής μάθησης, support vector machines. Στο συγκεκριμένο πλαίσιο κατηγοριοποίησης, πετύχαμε την σωστή λύση μετά από πέντε επαναλήψεις. Ενώ η κατηγοριοποίηση στους διαγωνισμούς BCI είναι για offline ανάλυση, η προσέγγιση μας μπορεί να θεωρηθεί ως μια online λύση για τον πραγματικό κόσμο. Είναι γρήγορη, απαιτεί λίγες (λιγότερες από 10) θέσεις ηλεκτροδίων, απαιτεί μόνο ένα μικρό ποσοστό προεπεξεργασίας και η επιλογή των τιμών για κρίσιμες παραμέτρους έχει αυτοματοποιηθεί. / We propose an approach to analyze data from the P300 speller paradigm using the machine-learning technique support vector machines. In a conservative classification scheme, we found the correct solution after five repetitions. While the classification within the BCI competition is designed for offline analysis, our approach is also well-suited for a real-world online solution: It is fast, requires few electrode positions (less than 10), demands only a small amount of preprocessing and selection of values for critical parameters is automated.

Το φαινόμενο του ύπνου αποτελεί ένα μυστήριο της ανθρώπινης ζωής. Η έρευνα του φαινομένου του ύπνου γίνεται κυρίως μέσω καταγραφών Ηλεκτροεγκεφαλογραφήματος. Η επεξεργασία και η ανάλυση του ηλεκτροεγκεφαλογραφήματος ύπνου με την χρήση υπολογιστικών αλγορίθμων έχει δώσει μεγάλη ώθηση στην διερεύνηση της εγκεφαλικής δραστηριότητας. Κατά την διάρκεια του ύπνου παρουσιάζονται χαρακτηριστικοί ρυθμοί και κύματα όπως είναι οι άτρακτοι (spindles) και τα συμπλέγματα Κ (K-Complex). Η ακριβής σημασία και συμβολή των συγκεκριμένων ρυθμών δεν έχει πλήρως διερευνηθεί. Πιστεύεται όμως ότι έχουν σπουδαίο υπναγωγικό ρόλο και συμβάλλουν στις διαδικασίες σταθεροποίησης της μνήμης κατά τον ύπνο. Στην παρούσα εργασία γίνεται μια προσπάθεια εύρεσης της χρονικής συσχέτισης μεταξύ των εμφανιζόμενων ρυθμών και κυμάτων κατά την διάρκεια του ύπνου. Για την ανάπτυξη του προγράμματος χρησιμοποιήθηκε το εργαλείο που έχει αναπτυχθεί και χρησιμοποιείται για επεξεργασία σήματος στο Εργαστήριο Νευροφυσιολογίας του Πανεπιστημίου Πατρών (Flying Circus). Το συγκεκριμένο πρόγραμμα επεκτάθηκε με τις κατάλληλες ρουτίνες έτσι ώστε να παράγονται γραφικές απεικονίσεις των χρονικών στιγμών εμφάνισης των ΗΕΓ χαρακτηριστικών εκ των οποίων να συνάγεται οπτικά αν υπάρχει συσχέτιση μεταξύ τους, το ψηφιδόπλεγμα (raster). Ο χρήστης μπορεί να επιλέξει είτε την συσχέτιση των ρυθμών στον χρόνο (time raster) είτε την συσχέτιση των ρυθμών στο πεδίο των συχνοτήτων (spectrum raster). Στην εργασία παρουσιάζονται οι γραφικές απεικονίσεις και τα αποτελέσματα από την χρήση του προγράμματος σε καταγραφές ύπνου που έχουν γίνει στο Εργαστήριο Νευροφυσιολογίας. / Sleep is a mysterious phenomenon of human life. Electroencephalogram (EEG) is the main tool that researchers use to explore the physiology of sleep. Advanced analysis of sleep recordings enormously contributed to the research and understanding of brain activity. During sleep we observe some characteristic rhythms and waves such as spindles and K-Complexes. The exact meaning and role of these rhythms has not yet fully discovered. However, it is believed that they play a significant sleep promoting role and partake in cognitive processes of sleep. The present work attempts to find the temporal correlation between those rhythms and waves that appear during the second stage of non-REM sleep i.e. whether any changes in the rate of occurrence of spindles can predict or can be predicted by the occurrence of K-Complexes. A graphic interface tool was developed by adding new processing functions to the tool that is used at the Neurophysiology Unit of the University of Patras’ Medical School (Flying Circus). This program is a raster used to support the EEG studies with the appropriate routines in order to produce graphical representations of the time development of various EEG characteristics and to demonstrate visually any possible correlation between them. The new environment includes methods that can correlate the appearance of rhythms and waves in the time domain or in the frequency spectrum domain using the raster graphical representation. The result of these methods is either a time raster or a spectrum raster depending on the user choices. In this paper we present the results of using these methods in real sleep recordings that have been conducted in the Neurophysiology Unit.

L'électroencéphalographie (EEG) est une technique d'imagerie cérébrale non invasive importante, capable d'enregistrer l'activité neuronale avec une grande résolution temporelle (ms), mais avec une résolution spatiale faible. Le problème inverse en EEG est un problème difficile, fortement sous-déterminé : des contraintes ou des a priori sont nécessaires pour aboutir à une solution unique. Récemment, l'intégration de signaux EEG et d'imagerie par résonance magnétique fonctionnelle (fMRI) a été largement considérée. Les données EEG et fMRI relatives à une tâche donnée, reflètent les activités neuronales des mêmes régions. Nous pouvons donc supposer qu'il existe des cartes spatiales communes entre données EEG et fMRI. En conséquence, résoudre le problème inverse en EEG afin de trouver les cartes spatiales des sources EEG congruentes avec celles obtenues par l'analyse de signaux fMRI semble être une démarche réaliste. Le grand défi reste la relation entre l'activité neuronale électrique (EEG) et l'activité hémodynamique (fMRI), qui n'est pas parfaitement connue à ce jour. La plupart des études actuelles reposent sur un modèle neurovasculaire simpliste par rapport à la réalité. Dans ce travail, nous utilisons des a priori et des faits simples et généraux, qui ne dépendent pas des données ou de l'expérience et sont toujours valides, comme contraintes pour résoudre le problème inverse en EEG. Ainsi, nous résolvons le problème inverse en EEG en estimant les sources spatiales parcimonieuses, qui présentent la plus forte corrélation avec les cartes spatiales obtenues par fMRI sur la même tâche. Pour trouver la représentation parcimonieuse du signal EEG, relative à une tâche donnée, on utilise une méthode (semi-aveugle) de séparation de sources avec référence (RSS), qui extrait les sources dont la puissance est la plus corrélée à la tâche. Cette méthode a été validée sur des simulations réalistes et sur des données réelles d'EEG intracrânienne (iEEG) de patients épileptiques. Cette représentation du signal EEG dans l'espace des sources liées à la tâche est parcimonieuse. En recherchant les fonctions d'activation de fMRI similaires à ces sources, on déduit les cartes spatiales de fMRI très précises de la tâche. Ces cartes fournissent une matrice de poids, qui impose que les voxels activés en fMRI doivent être plus importants que les autres voxels dans la résolution du problème inverse en EEG. Nous avons d'abord validé cette méthode sur des données simulées, puis sur des données réelles relatives à une expérience de reconnaissance de visages. Les résultats montrent en particulier que cette méthode est très robuste par rapport au bruit et à la variabilité inter-sujets. / Electroencephalography (EEG) is an important non-invasive imaging technique as it records the neural activity with high temporal resolution (ms), but it lacks high spatial resolution. The inverse problem of EEG is underdetermined and a constraint or prior information is needed to find a unique solution. Recently, EEG-fMRI integration is widely considered. These methods can be categoraized in three groups. First group uses the EEG temporal sources as the regressors in the generalized linear method (GLM) which is used to analyze the fMRI data. The second group analyzes EEG and fMRI simultaneously which is known as fusion technique. The last one, which we are interested in, uses the fMRI results as prior information in the EEG inverse problem. The EEG and fMRI data of a specific task, eventually reflect the neurological events of the same activation regions. Therefore, we expect that there exist common spatial patterns in the EEG and the fMRI data. Therefore, solving the EEG inverse problem to find the spatial pattern of the EEG sources which is congruent with the fMRI result seems to be close to the reality. The great challenge is the relationship between neural activity (EEG) and hemodynamic changes (fMRI), which is not discovered by now. Most of the previous studies have used simple neurovascular model because using the realistic model is very complicated. Here, we use general and simple facts as constraints to solve the EEG inverse problem which do not rely on the experiment or data and are true for all cases. Therefore, we solve the EEG inverse problem to estimate sparse connected spatial sources with the highest correlation with the fMRI spatial map of the same task. For this purpose, we have used sparse decomposition method. For finding sparse representation of the EEG signal, we have projected the data on the uncorrelated temporal sources of the activity. We have proposed a semi-blind source separation method which is called reference-based source separation (R-SS) and extracts discriminative sources between the activity and the background. R-SS method has been verified on a realistic simulation data and the intracranial EEG (iEEG) signal of five epileptic patients. We show that the representation of EEG signal in its task related source space is sparse and then a weighted sparse decomposition method is proposed and used to find the spatial map of the activity. In the weighted sparse decomposition method we put fMRI spatial map in the weighting matrix, such that the activated voxels in fMRI are considered more important than the other voxels in the EEG inverse problem. We validated the proposed method on the simulation data and also we applied the method on the real data of the face perception experiment. The results show that the proposed method is stable against the noise and subject variability.

Cette thèse explore l'analyse de l'activité électrique du cerveau. Un défi important de ces signaux est leur grande variabilité à travers différents essais et/ou différents sujets. Nous proposons une nouvelle méthode appelée "adaptive waveform learning" (AWL). Cette méthode est suffisamment générale pour permettre la prise en compte de la variabilité empiriquement rencontrée dans les signaux neuroélectriques, mais peut être spécialisée afin de prévenir l'overfitting du bruit. La première partie de ce travail donne une introduction sur l'électrophysiologie du cerveau, présente les modalités d'enregistrement fréquemment utilisées et décrit l'état de l'art du traitement de signal neuroélectrique. La principale contribution de cette thèse consiste en 3 chapitres introduisant et évaluant la méthode AWL. Nous proposons d'abord un modèle de décomposition de signal général qui inclut explicitement différentes formes de variabilité entre les composantes de signal. Ce modèle est ensuite spécialisé pour deux applications concrètes: le traitement d'une série d'essais expérimentaux segmentés et l'apprentissage de structures répétées dans un seul signal. Deux algorithmes sont développés pour résoudre ces problèmes de décomposition. Leur implémentation efficace basée sur des techniques de minimisation alternée et de codage parcimonieux permet le traitement de grands jeux de données.Les algorithmes proposés sont évalués sur des données synthétiques et réelles contenant des pointes épileptiformes. Leurs performances sont comparées à celles de la PCA, l'ICA, et du template-matching pour la détection de pointe. / This thesis investigates the analysis of brain electrical activity. An important challenge is the presence of large variability in neuroelectrical recordings, both across different subjects and within a single subject, for example, across experimental trials. We propose a new method called adaptive waveform learning (AWL). It is general enough to include all types of relevant variability empirically found in neuroelectric recordings, but can be specialized for different concrete settings to prevent from overfitting irrelevant structures in the data. The first part of this work gives an introduction into the electrophysiology of the brain, presents frequently used recording modalities, and describes state-of-the-art methods for neuroelectrical signal processing. The main contribution of this thesis consists in three chapters introducing and evaluating the AWL method. We first provide a general signal decomposition model that explicitly includes different forms of variability across signal components. This model is then specialized for two concrete applications: processing a set of segmented experimental trials and learning repeating structures across a single recorded signal. Two algorithms are developed to solve these models. Their efficient implementation based on alternate minimization and sparse coding techniques allows the processing of large datasets. The proposed algorithms are evaluated on both synthetic data and real data containing epileptiform spikes. Their performances are compared to those of PCA, ICA, and template matching for spike detection.

This paper addresses the problem of detecting K-complexes in sleep EEG. The study of sleep has become very essential to diagnose the brain disorders and analysis of brain activities. Since Kcomplex can have a wide variety of shapes it is very difficult to detect the K-complexes manually. In this paper, I present an automatic method for K-complexes detection based wavelet transform,TKEO and method for classification using feedforward multilayer neural network designed in Matlab. Detection performance reached the value approx. from 52,9 to 83,6 %.

Background: The visual system may process spatial frequency information in a low-to-high, coarse-to-fine sequence. In particular, low and high spatial frequency information may be processed via different pathways during object recognition, with LSF information projected rapidly to frontal areas and HSF processed later in visual ventral areas. In an electroencephalographic study, we examined the time course of information processing for images filtered to contain different ranges of spatial frequencies. Participants viewed either high spatial frequency (HSF), low spatial frequency (LSF), or unfiltered, broadband (BB) images of objects or nonobject textures, classifying them as showing either man-made or natural objects, or nonobjects. Event-related potentials (ERPs) and evoked and total gamma band activity (eGBA and tGBA) recorded using the electroencephalogram were compared for object and nonobject images across the different spatial frequency ranges. Results: The visual P1 showed independent modulations by object and spatial frequency, while for the N1 these factors interacted. The P1 showed more positive amplitudes for objects than nonobjects, and more positive amplitudes for BB than for HSF images, which in turn evoked more positive amplitudes than LSF images. The peak-to-peak N1 showed that the N1 was much reduced for BB non-objects relative to all other images, while HSF and LSF nonobjects still elicited as negative an N1 as objects. In contrast, eGBA was influenced by spatial frequency and not objecthood, while tGBA showed a stronger response to objects than nonobjects. Conclusions: Different pathways are involved in the processing of low and high spatial frequencies during object recognition, as reflected in interactions between objecthood and spatial frequency in the visual N1 component. Total gamma band seems to be related to a late, probably highlevel representational process.

This project added to the sparse body of research on the neural underpinnings of turn-taking with an electroencephalography (EEG) investigation of spontaneous conversation. Eighteen participants (3 male, 15 female, mean age 29.79), recruited and participating in pairs, underwent EEG hyperscanning as they conversed on a freely chosen topic for 45 minutes. In line with previous research, it was predicted that a time-frequency analysis of the EEG might reveal either increased power at around 10 Hz (the location of one of two components of the mu rhythm, an oscillation possibly involved in motor preparation for speech), or reduced alpha (8-12 Hz) power (reflecting non-motor aspects of turn preparation) prior to taking one’s turn. Increased power between 8-12 Hz was observed around 1.5 and 1 second preceding turn-taking, but similar power increases also occurred prior to turn-yielding and the conversation partner continuing after a pause, and a reduction in alpha power was found in turn-taking relative to listening to the other speaker continue after a pause. It is unclear whether this activity reflected motor or non-motor aspects of turn preparation, but the spontaneous conversation paradigm proved feasible for investigating brain activity coupled to turn-taking despite the methodological obstacles. / Detta forskningsprojekt bedrar till ett ämne där relativt få studier har genomförts med en elektroencefalografi- (EEG-) undersökning av hjärnaktivitet som är kopplad till turtagning i spontant samtal. Arton deltagare (3 män, 15 kvinnor, medelålder 29,79) som rekryterades och deltog i par, genomgick EEG-hyperscanning medan de pratade om ett fritt valt ämne i 45 minuter. Det förutsades att en tidsfrekvensanalys av EEG kan avslöja antingen ökad effekt vid cirka 10 Hz (vilket motsvarar en av två komponenter i mu-rytmen, en oscillation som eventuellt är involverad i motoriska förberedelser för tal) eller reducerad alfaeffekt (8 -12 Hz) (vilket möjligen återspeglar icke-motoriska aspekter av turtagningsförberedelser) innan man tar sin tur. Ökad effekt mellan 8-12 Hz observerades ungefär 1,5 och 1 sekund före turtagning, men liknande ökningar inträffade också innan samtalspartnern tog sin tur eller fortsatte efter en paus, och en minskning av alfaeffekt observerades när turtagning jämfördes till kontexter där försökspersonerna lyssnade när den andra talaren fortsatte efter en paus. Det är oklart om denna aktivitet återspeglade motoriska eller icke-motoriska aspekter av turtagningsförberedelser, men det visar sig vara möjligt att undersöka hjärnaktivitet kopplad till spontant samtal på ett rimligt sätt trots paradigmens metodologiska svårigheter. / Hidden events in turn-taking