Tesis sobre el tema "Brain-Computer Interfaces (BCIs)"

Les interfaces cerveau-ordinateur (BCI) basées sur la géométrie riemannienne ont gagné en popularité au cours de la dernière décennie, démontrant des améliorations significatives dans divers contextes de classification des BCI. Malgré ces avancées, les systèmes BCI restent insuffisamment fiables pour des applications pratiques. L'une des principales difficultés auxquelles les BCI sont confrontées est la variabilité considérable de l'électroencéphalogramme (EEG). Cette variabilité est censée être encore plus prononcée lorsque les systèmes BCI sont utilisés sur plusieurs jours ou en dehors des environnements de laboratoire contrôlés. Cette thèse aborde la grande variabilité de l'EEG sous divers angles sur la variété riemannienne des matrices symétriques définies positives (SPD). Nos six contributions au total peuvent être divisées en trois catégories. Dans la première section, nous proposons deux approches pour atténuer la variabilité de la distribution des données intra-utilisateur sur une variété SPD. La première contribution est une méthode automatique de détection des valeurs aberrantes basée sur le regroupement spectral des matrices SPD EEG, qui permet de détecter les valeurs aberrantes plus précisément que les méthodes existantes de manière entièrement centrée sur les données. La deuxième contribution propose un modèle de classification qui prend en compte les distributions multimodales des matrices SPD sur un collecteur. Notre classificateur multimodal améliore de manière significative la précision de la classification pour un ensemble de données très variables par rapport à un classificateur unimodal standard. La deuxième section traite de la variabilité inter-utilisateur en proposant deux méthodes de sélection de paramètres personnalisées. La première méthode implique une réduction dimensionnelle pour projeter les matrices SPD dans des sous-espaces de basse dimension plus discriminants entre classes, améliorant significativement la précision de classification par rapport à l'espace dimensionnel original. La deuxième méthode est une approche de sélection de bandes de fréquences et de fenêtres temporelles discriminantes basée sur la distinctivité des classes sur une variété SPD. Notre approche de sélection a considérablement amélioré la précision de la classification par rapport à une référence sans sélection de paramètres personnalisés et à une méthode de sélection conventionnelle bien connue. Dans la section finale, nous nous concentrons sur la conception de caractéristiques de classification moins variables dérivées de mesures neurophysiologiques qui ont été sous-utilisées dans les études BCI. Nous proposons de nouvelles représentations de matrices SPD exploitant des couplages inter-fréquences comme caractéristiques de classification, améliorant significativement la précision par rapport aux représentations SPD riemanniennes conventionnelles. De plus, nous avons exploré l'efficacité de la suppression d'une composante hautement variable du signal neural basée sur la paramétrisation périodique/aperiodique des signaux EEG. Cela pourrait contribuer au développement de stratégies neuroscientifiquement interprétables pour aborder la grande variabilité des EEG/BCI. Nos résultats empiriques issus de ces six contributions ouvrent la voie au développement d'algorithmes qui traitent plus efficacement la variabilité significative de l'EEG, faisant progresser la conception d'applications BCI fiables
Riemannian geometry-based Brain-Computer Interfaces (BCIs) have gained momentum over the last decade, demonstrating significant improvements in various BCI classification contexts. Despite these advancements, BCI systems remain insufficiently reliable for practical applications. One of the obstacles facing BCIs is the considerable variability of electroencephalogram (EEG). This variability is expected to be even more pronounced when BCI systems are used over multiple days or outside controlled laboratory environments. This thesis tackled the large variability of EEG data from a variety of angles on the Riemannian manifold of symmetric positive definite (SPD) matrices. Our six contributions can be divided into three categories. In the first section, we proposed two approaches to mitigate the variability of intra-user data distribution on an SPD manifold. The first contribution is an automatic outlier detection method based on spectral clustering for EEG SPD matrices, which could detect outliers more accurately than existing methods in a fully data-driven manner. The second contribution proposed a classification model that accounts for multimodal distributions of SPD matrices on a manifold. Our classifier significantly improved accuracy for a highly variable dataset compared to a standard unimodal classifier. The second section tackled inter-user variability by proposing two personalized parameters selection methods. The first method involves dimensionality reduction to project SPD matrices into more class-discriminating low-dimensional subspaces, significantly enhancing classification accuracy from the original high-dimensional space. The second method is a discriminative frequency band and time window selection approach based on class distinctiveness on an SPD manifold. Our selection approach substantially improved classification accuracy over both a baseline without personalized parameters selection and a well-known conventional selection method. In the final section, we focused on designing less variable classification features derived from neurophysiological measurements that have been underutilized in BCI studies. We propose novel SPD matrix representations that exploit multiple cross-frequency coupling as classification features, significantly improving classification accuracy over conventional Riemannian SPD representations. Additionally, we explored the effectiveness of removing a highly variable component of neural signal based on periodic/aperiodic parameterization of EEG signals. This could contribute to the development of neuroscientifically interpretable strategies for addressing large variability in EEG/BCI. Our empirical findings from these six contributions collectively pave the way for algorithm developments that more effectively address significant EEG variability, advancing the design of reliable BCI applications

In recent years, brain computer interfaces (BCIs) have gained popularity in non-medical domains such as the gaming, entertainment, personal health, and marketing industries. A growing number of companies offer various inexpensive consumer grade BCIs and some of these companies have recently introduced the concept of BCI "App stores" in order to facilitate the expansion of BCI applications and provide software development kits (SDKs) for other developers to create new applications for their devices. The BCI applications access to users' unique brainwave signals, which consequently allows them to make inferences about users' thoughts and mental processes. Since there are no specific standards that govern the development of BCI applications, its users are at the risk of privacy breaches. In this work, we perform first comprehensive analysis of BCI App stores including software development kits (SDKs), application programming interfaces (APIs), and BCI applications w.r.t privacy issues. The goal is to understand the way brainwave signals are handled by BCI applications and what threats to the privacy of users exist. Our findings show that most applications have unrestricted access to users' brainwave signals and can easily extract private information about their users without them even noticing. We discuss potential privacy threats posed by current practices used in BCI App stores and then describe some countermeasures that could be used to mitigate the privacy threats. Also, develop a prototype which gives the BCI app users a choice to restrict their brain signal dynamically.

C’è un crescente interesse nella comunità scientifica per l’applicazione delle tecniche della bioingegneria nel campo delle interfacce fra cervello e computer. Questo interesse nasce dal fatto che in Europa ci sono almeno 300.000 persone con paralisi agli arti inferiori, con una età media piuttosto bassa (31 anni), registrandosi circa 5.000 nuovi casi ogni anno, in maggioranza dovuti ad incidenti automobilistici. Tali lesioni traumatiche spinali inducono delle disfunzioni sensoriali a causa dell’interruzione tra gli arti e i centri sopraspinali. Per far fronte a questi problemi gli scienziati si sono sempre più proiettati verso un nuovo settore: il Brain Computer Interaction, ossia un ambito della ricerca volto alla costruzione di interfacce in grado di collegare direttamente il cervello umano ad un dispositivo elettrico come un computer.

Con Brain-Computer Interface si intende un collegamento diretto tra cervello e macchina, che essa sia un computer o un qualsiasi dispositivo esterno, senza l’utilizzo di muscoli. Grazie a sensori applicati alla cute del cranio i segnali cerebrali del paziente vengono rilevati, elaborati, classificati (per mezzo di un calcolatore) e infine inviati come output a un device esterno. Grazie all'utilizzo delle BCI, persone con gravi disabilità motorie o comunicative (per esempio malati di SLA o persone colpite dalla sindrome del chiavistello) hanno la possibilità di migliorare la propria qualità di vita. L'obiettivo di questa tesi è quello di fornire una panoramica nell'ambito dell'interfaccia cervello-computer, mostrando le tipologie esistenti, cercando di farne un'analisi critica sui pro e i contro di ogni applicazione, ponendo maggior attenzione sull'uso dell’elettroencefalografia come strumento per l’acquisizione dei segnali in ingresso all'interfaccia.

Les Interfaces Cerveau-Ordinateur basées sur l’Imagerie Mentale (IM-ICO) permettent auxutilisateurs d’interagir uniquement via leur activité cérébrale, grâce à la réalisation de tâchesd’imagerie mentale. Cette thèse se veut contribuer à l’amélioration des IM-ICO dans le but deles rendre plus utilisables. Les IM-ICO sont extrêmement prometteuses dans de nombreuxdomaines allant de la rééducation post-AVC aux jeux-vidéo. Malheureusement, leurdéveloppement est freiné par le fait que 15 à 30% des utilisateurs seraient incapables de lescontrôler. Nombre de travaux se sont focalisés sur l’amélioration des algorithmes de traitementdu signal. Par contre, l’impact de l’entraînement des utilisateurs sur leur performance estsouvent négligé. Contrôler une IM-ICO nécessite l’acquisition de compétences et donc unentraînement approprié. Or, malgré le fait qu’il ait été suggéré que les protocolesd’entraînement actuels sont théoriquement inappropriés, peu d’efforts sont mis en oeuvre pourles améliorer. Notre principal objectif est de comprendre et améliorer l’apprentissage des IMICO.Ainsi, nous cherchons d’abord à acquérir une meilleure compréhension des processussous-tendant cet apprentissage avant de proposer une amélioration des protocolesd’entraînement afin qu’ils prennent en compte les facteurs cognitifs et psychologiquespertinents et qu’ils respectent les principes issus de l’ingénierie pédagogique. Nous avonsainsi défini 3 axes de recherche visant à investiguer l’impact (1) de facteurs cognitifs, (2) de lapersonnalité et (3) du feedback sur la performance. Pour chacun de ces axes, nous décrivonsd’abord les études nous ayant permis de déterminer les facteurs impactant la performance ;nous présentons ensuite le design et la validation de nouvelles approches d’entraînementavant de proposer des perspectives de travaux futurs. Enfin, nous proposons une solution quipermettrait d’étudier l’apprentissage de manière mutli-factorielle et dynamique : un systèmetutoriel intelligent
Mental-imagery based brain-computer interfaces (MI-BCIs) enable users to interact with theirenvironment using their brain-activity alone, by performing mental-imagery tasks. This thesisaims to contribute to the improvement of MI-BCIs in order to render them more usable. MIBCIsare bringing innovative prospects in many fields, ranging from stroke rehabilitation tovideo games. Unfortunately, most of the promising MI-BCI based applications are not yetavailable on the public market since an estimated 15 to 30% of users seem unable to controlthem. A lot of research has focused on the improvement of signal processing algorithms.However, the potential role of user training in MI-BCI performance seems to be mostlyneglected. Controlling an MI-BCI requires the acquisition of specific skills, and thus anappropriate training procedure. Yet, although current training protocols have been shown tobe theoretically inappropriate, very little research is done towards their improvement. Our mainobject is to understand and improve MI-BCI user-training. Thus, first we aim to acquire a betterunderstanding of the processes underlying MI-BCI user-training. Next, based on thisunderstanding, we aim at improving MI-BCI user-training so that it takes into account therelevant psychological and cognitive factors and complies with the principles of instructionaldesign. Therefore, we defined 3 research axes which consisted in investigating the impact of(1) cognitive factors, (2) personality and (3) feedback on MI-BCI performance. For each axis,we first describe the studies that enabled us to determine which factors impact MI-BCIperformance; second, we describe the design and validation of new training approaches; thethird part is dedicated to future work. Finally, we propose a solution that could enable theinvestigation of MI-BCI user-training using a multifactorial and dynamic approach: an IntelligentTutoring System

Pazienti con lesioni cerebrali o spinali possono essere affetti da gravi deficit nelle funzioni sensoriali, motorie e comunicative; sono perciò sempre più necessarie tecniche di riabilitazione avanzate, personalizzate e adattative, per limitare i deficit insorti e restituire al paziente una vita il più normale possibile. Negli ultimi decenni, numerosi gruppi di ricerca hanno sviluppato Brain-Computer Interface (BCI) basate sul segnale elettroencefalografico (EEG) con l’obbiettivo di fornire mezzi di comunicazione o riabilitazione motoria funzionale. Tuttavia, le tecnologie BCI hanno un ampio potenziale al di là della sola riabilitazione motoria. Applicazioni dei sistemi BCI in protocolli di riabilitazione cognitiva, ad esempio, hanno conseguito risultati promettenti nella prospettiva di migliorare funzioni quali l’attenzione, l'apprendimento e la memoria in pazienti con disturbi delle funzioni cognitive. In questo lavoro di Tesi si analizzano i principi di funzionamento dei sistemi BCI, a partire dall’acquisizione del segnale elettroencefalografico fino all’estrazione e alla classificazione delle feature del segnale per decodificare intenzioni motorie e processi cognitivi (memoria, attenzione) dell’utente. Viene poi presentata un’analisi della letteratura per quando riguarda gli approcci BCI in riabilitazione sia motoria che cognitiva, prestando particolare attenzione ai metodi utilizzati per l’elaborazione e traduzione del segnale EEG. Sono stati considerati con particolare attenzione studi che valutano gli effetti dell’applicazione di BCI non solo attraverso performance motorie e cognitive ma anche utilizzando tecniche di neuro-imaging avanzate, per indagare possibili cambiamenti nell’organizzazione funzionale della corteccia cerebrale sottostanti i risultati positivi ottenuti. Infine, vengono commentati i vantaggi e le limitazioni di queste tecnologie riabilitative e i problemi ancora aperti.

Brain-Computer Interface (BCI) offers the opportunity to paralyzed patients to control their movements without any neuromuscular activity. Signal processing of neuronal activity enables to decode movement intentions. Ability for patient to control an effector is closely linked to this decoding performance. In this study, I tackle a recent way to decode neuronal activity: Deep learning. The study is based on public data extracted by Schalk et al. for BCI Competition IV. Electrocorticogram (ECoG) data from three epileptic patients were recorded. During the experiment setup, the team asked subjects to move their fingers and recorded finger movements thanks to a data glove. An artificial neural network (ANN) was built based on a common BCI feature extraction pipeline made of successive convolutional layers. This network firstly mimics a spatial filtering with a spatial reduction of sources. Then, it realizes a time-frequency analysis and performs a log power extraction of the band-pass filtered signals. The first investigation was on the optimization of the network. Then, the same architecture was used on each subject and the decoding performances were computed for a 6-class classification. I especially investigated the spatial and temporal filtering. Finally, a preliminary study was conducted on prediction of finger movement. This study demonstrated that deep learning could be an effective way to decode brain signal. For 6-class classification, results stressed similar performances as traditional decoding algorithm. As spatial or temporal weights after training are slightly described in the literature, we especially worked on interpretation of weights after training. The spatial weight study demonstrated that the network is able to select specific ECoG channels notified in the literature as the most informative. Moreover, the network is able to converge to the same spatial solution, independently to the initialization. Finally, a preliminary study was conducted on prediction of movement position and gives encouraging results.

Human to computer interaction only using thoughts is no longer a science fiction topic and recent progress made in this field are astounding. This work shows the creation of a novel upper limb pointing movement performance quantification platform (eCAM test) and its validation on a group of healthy subjects. After that, it shows that functional electrical stimulation (FES) enhances these upper limbs movements performance. Furthermore, this work shows that anodal transcranial direct current stimulation (atDCS) of the cerebellum impacts brain rhythms as well as postural tremor on a patient. Also, the MRI data gathered during this work will allow to better understand the underlying mechanisms of tDCS. Following that, it has been shown that the frequency and complexity of a tapping task increase the postural tremor of the contralateral limb. The same effect has been witnessed for neuromuscular fatigue. All these advances allowed us to place the foundations of a multimodal brain computer interface (BCI) based on sensors fusion. A development phase is now required to create this interface and test it on healthy and sick subjects.
Communiquer avec un ordinateur par le biais de la pensée n'est plus un sujet de science-fiction et les progrès effectués dans le domaine sont ahurissants. Ce travail montre la création d'une nouvelle plateforme de mesure de la performance des mouvements de pointage verticaux (eCAM test) ainsi que sa validation sur une cohorte de sujets sains. Suite à cela, il montre que la stimulation électrique fonctionnelle (FES) améliore la performance de ces mouvements des membres supérieurs. En plus il démontre que la stimulation anodale trancranienne en courant continu (atDCS) du cervelet a un effet sur les rythmes des signaux cérébraux ainsi que sur le tremblement postural d'un patient. De plus des données IRM recueillies durant ce travail permettront de mieux cerner les mécanismes d'action de la stimulation tDCS. Suite à cela, il a été montré que la fréquence et la complexité d'une tâche de tapping augmentent le tremblement postural du membre controlatéral. Le même effet est constaté pour la fatigue musculaire. Toutes ces avancées installent les fondements à la création d'une interface cerveau-machine multimodale basée sur la fusion de senseurs. Une phase de développement est maintenant nécessaire pour établir cette interface et la tester sur des sujets sains et malades.
Doctorat en Sciences biomédicales et pharmaceutiques (Médecine)
info:eu-repo/semantics/nonPublished

Thesis advisor: Robert Signorile
It has long been known that as neurons fire within the brain they produce measurable electrical activity. Electroencephalography (EEG) is the measurement and recording of these electrical signals using sensors arrayed across the scalp. Though there is copious research in using EEG technology in the fields of neuroscience and cognitive psychology, it is only recently that the possibility of utilizing EEG measurements as inputs in the control of computers has emerged. The idea of Brain-Computer Interfaces (BCIs) which allow the control of devices using brain signals evolved from the realm of science fiction to simple devices that currently exist. BCIs naturally present themselves to many extremely useful applications including prosthetic devices, restoring or aiding in communication and hearing, military applications, video gaming and virtual reality, and robotic control, and have the possibility of significantly improving the quality of life of many disabled individuals. However, current BCIs suffer from many problems including inaccuracies, delays between thought, detection, and action, exorbitant costs, and invasive surgeries. The purpose of this research is to examine the Emotiv EPOC© System as a cost-effective gateway to non-invasive portable EEG measurements and utilize it to build a thought-based BCI to control the Parallax Scribbler® robot. This research furthers the analysis of the current pros and cons of EEG technology as it pertains to BCIs and offers a glimpse of the future potential capabilities of BCI systems
Thesis (BA) — Boston College, 2010
Submitted to: Boston College. College of Arts and Sciences
Discipline: Computer Science Honors Program
Discipline: Computer Science

This thesis describes a brain-computer interface (BCI) system that was developed to allow direct cortical control of 7 active degrees of freedom in a robotic arm. Two monkeys with chronic microelectrode implants in their motor cortices were able to use the arm to complete an oriented grasping task under brain control. This BCI system was created as a clinical prototype to exhibit (1) simultaneous decoding of cortical signals for control of the 3-D translation, 3-D rotation, and 1-D finger aperture of a robotic arm and hand, (2) methods for constructing cortical signal decoding models based on only observation of a moving robot, (3) a generalized method for training subjects to use complex BCI prosthetic robots using a novel form of operator-machine shared control, and (4) integrated kinematic and force control of a brain-controlled prosthetic robot through a novel impedance-based robot controller. This dissertation describes each of these features individually, how their integration enriched BCI control, and results from the monkeys operating the resulting system.

Syftet med uppsatsen är att ge en uppdaterad översikt av området BCI (Brain Computer Interface) och undersöka vad som hänt sedan begreppet introducerades i forskningssammanhang; vilka praktiska resultat forskningen lett till och vilka tillämpningar som tillkommit. Metoden som företrädesvis används är litteraturstudie som tecknar bakgrund och enkät. Därefter följer en diskussion där utmaningar för framtiden, potential och tillämpningar i BCI-tekniken behandlas utifrån ett neuropsykologiskt perspektiv. Kommer BCI-tekniken att implementeras på samma sätt som radio, TV och telekommunikationer i samhället och vilka etiska och tekniska problem finns idag. För att skildra allmänhetens uppfattning om BCI genomfördes en webbaserad enkätundersökning (survey) i form av pilotstudie (n=32) som syftar till att ge en indikation på attityder och hur allmänhetens opinion med avseende på tillämpningar i samtiden och jämförelser med avseende på teknisk bakgrund.

In this manuscript three related aspects of research on BCI systems were discussed. These aspects were the evaluation of a nonlinear feature extraction algorithm for BCI, the analysis of the functional connectivity between the signals acquired in different brain regions when a user performs an operant conditioning paradigm with fMRI based BCI technology, and the development of BCIs applications for disabled subjects. We have introduced a new EEG signals features extraction techniques based on nonlinear time series analysis. This signal processing approach was tested offline considering three sessions of imaginary motor tasks. The main objective is increasing the performance of BCI systems extracting a more robust feature. In order to reach this objective a fast algorithm that computes the largest Lyapunov exponent, the DivA [7], was used. This implementation results to be computationally less onerous than the conventional ones, since it is not based on the time-delay embedding concept and also no intermediate computational steps are needed to obtain the final result. For this reason the DivA is particularly suitable for real time analysis, thus for BCI applications. Our evaluations underline the capability and the potentiality of this method in respect to the classical approach. The idea for future works is to integrate the nonlinear algorithm investigated in this thesis in a BCI system, thus using it on line. The design of a BCI based on our nonlinear feature extraction method could improve the performance of the systems that use sensory motor rhythms as neurophysiologic signals. The analysis of the functional connectivity between brain regions involved in the perception of pain is the second topic that were dealt with in this thesis. Thanks to the collaboration with the central institute of mental health, Heidelberg university in Mannheim, the dataset recorded with an fMRI based BCI technology have been analysed. The results reveal the possibility for a person to modulate the brain waves, in particular the neurophysiologic signals related to the perception of pain. Control over the pain modulatory system is an important target because it could enable a unique mechanism for clinical control over pain. Here, we found that using real-time functional MRI based BCI to guide training, subjects were able to learn to control activation both in anterior cingulated cortex and in the posterior insula. The BCI techniques could have an important role for treating disease, for example for the chronic pain treatment. An aspect that can be investigated in future work is the involvement of the medial cingulate cortex in the pain perception. Indeed when the subject deliberately induced increases or decreases in ACC or pIns fMRI activation, there was a corresponding change in the connection between the MCC and the other ROIs. In particular a more strong connection between MCC and pInsR can be noticed. In future works could be interesting to analyze the role of MCC in the perception of pain. Finally, we proposed the designs of different EEG based BCI applications. We aim to provide a significant quality of life improvement to users with severe disabilities. All the applications designed have been tested for able-bodied users, the future idea is to test the applicability of such tools also for the locked-in patients.

Severe neuromuscular disorders can produce locked-in syndrome (LIS), a loss of nearly all voluntary muscle control. A brain-computer interface (BCI) using the P300 event-related potential provides communication that does not depend on neuromuscular activity and can be useful for those with LIS. Currently, there is no way of determining the effectiveness of P300-based BCIs without testing a person's performance multiple times. Additionally, P300 responses in BCI tasks may not resemble the typical P300 response. I sought to clarify the relationship between the P300 response and BCI task parameters and examine the possibility of a predictive relationship between traditional oddball tasks and BCI performance. Both waveform and component analysis have revealed several task-dependent aspects of brain activity that show significant correlation with the user's performance. These components may provide a fast and reliable metric to indicate whether the BCI system will work for a given individual.

Brain-computer interfaces (BCIs) blend methods and concepts researched by cognitive neuroscience, electrophysiology, computer science and engineering, resulting in systems of bi-directional information exchange directly between brain and computer. BCIs contribute to medical applications that restore communication and mobility for disabled patients and provide new forms of sending information to devices for enhancement and entertainment. Virtual reality (VR) introduces humans into a computer-generated world, tackling immersion and involvement. VR technology extends the classical multimedia experience, as the user is able to move within the environment, interact with other virtual participants, and manipulate objects, in order to generate the feeling of presence. This essay presents the possibilities of merging BCI with VR and the challenges to be tackled in the future. Current attempts to combine BCI and VR technology have shown that VR is a useful tool to test the functioning of BCIs, with safe, controlled and realistic experiments; there are better outcomes for VR and BCI combinations used for medical purposes compared to solely BCI training; and, enhancement systems for healthy users seem promising with VR-BCIs designed for home users. Future trends include brain-to-brain communication, sharing of several users’ brain signals within the virtual environment, and better and more efficient interfaces.

La Brain Computer Interface (BCI) è una delle nuove tecnologie che si sta diffondendo negli ultimi anni con nuove ricerche e nuove applicazioni. La ricerca è continua e sempre con l'obiettivo di rendere la BCI uno strumento a disposizione di tutti e applicabile in molti contesti sfruttando la capacità di adattamento del dispositivo. In questo lavoro è stata fatta una prima introduzione alla struttura, alle caratteristiche di acquisizione e gestione del segnale e alle diverse tipologie di BCI oggi a disposizione; successivamente si è focalizzata l'attenzione sulle più recenti applicazioni e in particolare sulla la medicina riabilitativa, sia cognitiva che motoria. Infatti, sfruttando la BCI, i pazienti affetti da importanti disturbi del sistema nervoso centrale, o che hanno subito grossi traumi cerebrali, sono sempre più in grado di riguadagnare alcune delle libertà che avevano perso. Infine, nell'ultimo capitolo del lavoro è stato abbandonato l'aspetto puramente scientifico per focalizzarsi su quello etico, indagando i principali problemi e questi etici che si sono sviluppati insieme all'avvento della BCI.

2011/2012
La comunicazione alternativa attraverso le BCIs può risultare uno strumento per agevolare le condizioni di vita di pazienti affetti dai disturbi neurologici. La comunicazione via BCI è ancora molto più lenta rispetto alla comunicazione con il linguaggio naturale. L’obiettivo di questo lavoro di tesi si inserisce in tale ambito, consistendo nello sviluppo di nuove applicazioni in grado di migliorare la velocità di comunicazione con utilizzo delle BCIs. Sono stati sviluppati due sistemi di comunicazione alternativa basati sulla componente P300. Il primo sistema chiamato ‘Multimenu’ permette una selezione veloce di messaggi e di comandi impostati in una struttura gerarchica. Il secondo sistema di tipo predittivo, denominato PolyMorph, grazie ad algoritmi appositamente sviluppati, predice i caratteri e/o le parole successivi a quelli già selezionati in precedenza.
XXV Ciclo
1980

It is widely understood that neurons within the brain produce electrical activity, and electroencephalography—a technique used to measure biopotentials with electrodes placed upon the scalp—has been used to observe it. Today, scientists and engineers work to interface these electrical neural signals with computers and machines through the field of Brain-Computer Interfacing (BCI). BCI systems have the potential to greatly improve the quality of life of physically handicapped individuals by replacing or assisting missing or debilitated motor functions. This research thus aims to further improve the efficacy of the BCI based assistive technologies used to aid physically disabled individuals. This study deals with the testing and modification of a BCI system that uses the alpha and beta bands to detect motor intention by weighing online EEG output against a calibrated threshold.

Neurological degenerative diseases like stroke, Alzheimer, Amyothrophic Lateral Sclerosis (ALS), Parkinson and many others are constantly increasing their incidence in the world health statistics as far as the mean age of the global population is getting higher and higher. This leads to a general need for effective, at-home and low-cost rehabilitative and health-daily-care tools. The latter should consist either of technological devices implemented for operating in a remote way, i.e. tele-medicine is quickly spreading around the world, or very-advanced computer-based and robotic systems to realize intense and repetitive trainings. This is the challenge in which Information and Communications Technology (ICT) is asked to play a major role in order to bring medicine to reach further advancements. Indeed, no way to cope with these issues is possible outside a strong and vivid cooperation among multi-disciplinary teams of clinicians, physicians, biologists, neuropsychologists and engineers and without a resolute pushing towards a widespread interoperability between Institutes, Hospitals and Universities all over the world, as recently highlighted during the main International conferences on ICT in healthcare. The establishment of well-defined standards for gathering and sharing data will then represent a key element to enhance the efficacy of the aforementioned collaborations. Among the others, stroke is one of the most common neurological pathologies being the second or third cause of mortality in the world; moreover, it causes more than sixty percent survivors remain with severe cognitive and motor impairments that impede them in living normal lives and require a twenty-four-hours daily care. As a consequent, on one side stroke survivors experience a frustrating condition of being completely dependent on other people even to perform simple daily actions like reach and grasp an object,hold a glass of water to drink it and so on. States, by their side, have to take into account additional costs to provide stroke patients and their families with appropriate cares and supports to cope with their needs. For this reason, more and more fundings are recently made available by means of grants, European and International projects, programs to exchange different expertise among various countries with the aim to study how to accelerate and make more effective the recovery process of chronic stroke patients. The global research about this topic is conducted on several parallel aspects: as regard as the basic knowledge of brain processes, neurophysiologists, biologists and engineers are particularly interested in an in-depth understanding of the so-called neuroplastic changes that brain daily operates in order to adapt individuals to life changes, experiences and to realize more extensively their own potentialities. Neuroplasticity is indeed the corner stone for most of the trainings nowadays adopted by the standard as well as the more innovative methods in the rehabilitative programs for post-stroke recovery. Specifically speaking, motor rehabilitation usually includes longterm, repetitive and intense goal-directed exercises that promote neuroplastic mechanisms such as neural sprouting, synapto-genesis and dendritic branching. These processes are strictly related with motor improvements and their study could - one day - serve as prognostic measures of the recovery. Another aspect of this field of neuroscience research is the number of applications that it makes feasible. One of the most exciting is to connect an injured brain to a computer or a robotic device in a Brain-Computer or Brain-Machine Interface (BCI or BMI) scheme aiming at bypassing the impairments of the patient and make him/her autonomously move again or train his/her motor abilities in a more effective way. This kind of research can already count an amount of literature that provides several proofs of concept that these heterogeneous systems constituted by humans and robots can work at the purpose. A particular application of BCI for restoring or enhancing, at least, the reaching abilities of chronic stroke survivors was implemented and is still currently being improved at I.R.C.C.S. San Camillo Hospital Foundation, an Institute for the rehabilitation from neurological diseases located in Lido of Venice and partially technically supported by the Department of Information Engineering of Padua in range of an agreement signed in 2009. This specific BCI platform allows patients to train and improve their reaching movements by means of a robotic arm that provides a force that helps patients in completing the training exercise, i.e. to hit a predetermined target. This force feedback is however subject to a strict condition: during the movement, the person has to produce the expected pattern of cerebral activity. Whenever this is accomplished, a force is delivered proportionally to the entity of the latter activity, otherwise the patient is obliged to operate without any help. In this way, this platform implements the so-called operant-learning, that is one of the most effective conditioning techniques to make a subject learn or relearn a task. If, on one hand, the primary and explicit task is to improve a movement, on the other side the secondary but most important task is to deploy the perilesional part of the brain - still healthy - in becoming responsible for the control of the movement. It is a popular and widely-accepted opinion within the neuroscience community, indeed, that a healthy region of the sensorimotor area nearby the damaged one - which was previously in charge of performing the (reaching) movement - can optimally accomplish the impaired motor function substituting the original control area. Technically speaking, the main crucial feature that can ensure the effectiveness of the whole system is the precise and in real-time identification and quantification of the cerebral pattern associated with the movement, the worldwide named movement-related desynchronization (MRD). Starting from its original definition, passing through the most used techniques for its recognition, the thesis work presents a series of criticisms of the current signal processing method to detect the MRD and a complete analysis of the possible features that can better represent the movement condition and that can be more easily extracted during the on-line operations. Brain - it is well-known - learns by trials and errors and it needs a slightly-delayed (in the range of fraction of seconds) feedback of its performance to learn a task in the best way. This BCI application was born with the purpose to provide the above-mentioned feedback: however, this is only feasible if a computationally easy and contingent signal processing technique is available. This thesis work would like to cope with the lack of a well-planned real-time signal analysis in the current experimental protocol.

In this study, a P300 based Brain-Computer Interface (BCI) system design is realized by the implementation of the Spelling Paradigm. The main challenge in these systems is to improve the speed of the prediction mechanisms by the application of different signal processing and pattern classification techniques in BCI problems. The thesis study includes the design and implementation of a 10 channel Electroencephalographic (EEG) data acquisition system to be practically used in BCI applications. The electrical measurements are realized with active electrodes for continuous EEG recording. The data is transferred via USB so that the device can be operated by any computer. v Wiener filtering is applied to P300 Speller as a signal enhancement tool for the first time in the literature. With this method, the optimum temporal frequency bands for user specific P300 responses are determined. The classification of the responses is performed by using Support Vector Machines (SVM&rsquo
s) and Bayesian decision. These methods are independently applied to the row-column intensification groups of P300 speller to observe the differences in human perception to these two visual stimulation types. It is observed from the investigated datasets that the prediction accuracies in these two groups are different for each subject even for optimum classification parameters. Furthermore, in these datasets, the classification accuracy was improved when the signals are preprocessed with Wiener filtering. With this method, the test characters are predicted with 100% accuracy in 4 trial repetitions in P300 Speller dataset of BCI Competition II. Besides, only 8 trials are needed to predict the target character with the designed BCI system.

Le interfacce cervello-computer o brain-computer interface sono dei sistemi che attraverso la misurazione e l'analisi di segnali provenienti dall'attività cerebrale traducono in azione il segnale registrato, realizzando un canale di comunicazione alternativo alle normali vie neurali. L'obbiettivo di questa tesi è quello di analizzare e differenziare le varie tipologie di BCI dal punto di vista della struttura e dei metodi di acquisizione del segnale e infine classificarle a seconda della funzione che andranno a ricoprire: motoria o cognitiva. Le Interfacce cervello-computer nascono come una tecnologia per ripristinare le funzioni motorie perse a causa di patologie neurodegenerative. Negli ultimi anni però si è sviluppato un crescente interesse nell’utilizzo di queste interfacce anche per ripristinare funzioni cognitive, come ad esempio la memoria o l’attenzione. Lo scopo principale delle brain-computer interfaces utilizzate in medicina riabilitativa è dunque quello di permettere a pazienti che hanno subito gravi lesioni nel sistema nervoso centrale o periferico di essere più autonomi e di migliorare le loro capacità di comunicazione dei loro bisogni e delle loro esigenze ai caregivers.

Une Interface Cerveau-Ordinateur (ICO) est un système de communication direct entre le cerveau humain et un ordinateur. Les ICOs offrent un nouveau type d'interaction basée sur "la pensée" pour différent contexte d'utilisation comme la robotique, la domotique, le multimédia, la réalité virtuelle et les jeux vidéo. Dans ce travail, nous nous concentrons sur l'utilisation des ICOs pour interagir avec des environnements virtuels. Nous nous intéressons plus particulièrement à l'utilisation des signaux cérébraux en lien avec la charge mentale. Dans un premier temps, nous avons étudié les signaux électroencéphalographiques (EEG) et les ICOs qui peuvent être utilisés afin de déterminer en temps réel l'état mental de l'utilisateur comme sa charge mentale, son état de concentration et son état de relaxation. Nous les avons étudiés dans le contexte des ICOs actives, c. -à-d. Quand un utilisateur essaie de contrôler volontairement son activité cérébrale. Nous avons conçu et évalué une ICO active basée sur les états de relaxation et de concentration de l'utilisateur. Dans un second temps, nous avons étudié l'utilisation des ICOs passives afin d'améliorer l'interaction avec une ICO active. Lorsqu'un utilisateur utilise une ICO passive, il n'a pas besoin d'essayer de contrôler son activité mentale, il peut rester concentré sur sa tâche principale, l'activité cérébrale est mesurée de façon passive et utiliser afin d'adapter l'application. Nous avons introduit le concept d'un inhibiteur d'ICO, qui peut être défini comme une ICO passive qui surveille l'état mental de l'utilisateur afin de démarrer l'interaction avec une ICO active quand le "cerveau" de l'utilisateur est prêt. Ensuite nous avons étudié l'usage des ICOs passives basées sur la charge mentale dans le contexte de la réalité virtuelle afin d'assister l'utilisateur quand il présente une charge mentale importante. De façon à illustrer cette approche, nous avons conçu un système visuohaptique dans lequel un système de guidage haptique est automatiquement mis en route lorsque l'utilisateur présente une charge mentale élevée durant une tache de suivi de chemin. Enfin, nous avons étudié comment notre approche pourrait s'appliquer au contexte de la réalité virtuelle et des applications médicales. Nous avons proposé d'utiliser une ICO passive afin d'adapter en temps réel un simulateur médical à l'état de l'utilisateur. Son utilité. Ce type de simulateur pourrait être utilisé afin de créer une nouvelle génération de simulateur médical "intelligent" qui prend en compte l'état mental de l'utilisateur lors des procédures d'entrainements
A Brain Computer Interface (BCI) is a communication system between a user and a computer in which the carried message is the measured user's brain activity. BCIs offer a novel kind of interaction "by thought" for different application contexts such as robotics, domotics, multimedia, virtual reality, and video games. Ln this work we focus on the use of BCIs for interacting with virtual environments. We also focus on the use of brain signals that are related to the user's mental workload or closely related states. First, we have studied the electroencephalography (EEG) markers and the BCI setups that can be used to read out in real-ti me the user mental states such as mental workload, concentration and relaxation states. We have studied them in the «active» BCI context, i. E. , when the user deliberately tries to control his/her brain activity. We have designed and evaluated an active BCI system based on relaxation and concentration mental states. Then we have explored the use of "passive" BCIs to improve the use of an active BCI. Ln passive BCIs the user does not try to control his/ her brain activity, and he/she can remain mainly concerned by his/her primary task. The brain activity is analysed to read out the user mental state which is used to adapt the application. We have introduced the concept of the BCI inhibitor which can be defined as a passive BCI system that pauses the active BCI until the user's "brain" is ready. Then we have studied the use of passive BCIs based on mental workload in a virtual reality context to assist the user when a high workload is detected. To illustrate this approach we have designed a visuo-haptic virtual reality setup in which a haptic guiding system is automatically toggled based on mental workload during a path-following task. We have conducted an experiment to evaluate the operability and efficiency of the proposed system. Results suggest that the proposed passive BCI system is able to measure a mental workload index that seems well correlated with the difficulty of the task. Activation of guides based on the measured mental workload index allowed to increase task performances by significantly reducing the number of collisions. We have finally studied how our approach could be applied to virtual reality and medical applications. We have proposed to use a passive BCI to adapt in real-time a medical simulator to the user's mental states. If the user becomes too heavily engaged in the medical operation, some visual and haptic cues are automatically activated. This could pave the way to a future generation of «smart» medical simulator taking into account brain signals and mental states in training procedures

Con il termine Brain-Computer Interface (BCI) si indica un sistema hardware e software che permette, a partire da segnali di origine cerebrale, di tradurre le intenzioni dell’utente in comandi per il controllo di un dispositivo di output, come computer, sintetizzatori vocali, apparecchi di assistenza e neuroprotesi. Questa tipologia di applicazioni non richiede l’impiego di muscoli periferici, poiché sfrutta solamente specifici segnali generati dall’attività cerebrale. In particolare, il presente elaborato tratta le BCI-speller, ovvero sistemi che permettono la scrittura di un testo sfruttando le variazioni del segnale elettroencefalografico (EEG) suscitate attraverso un’interfaccia grafica (GUI). La GUI è costituita da lettere, simboli e numeri opportunamente presentati, così che, se il soggetto presta attenzione ad uno di essi, particolari potenziali cerebrali vengono elicitati nell’EEG e sfruttati per identificare e quindi selezionare tale simbolo. L’obiettivo dell’elaborato è introdurre l’emergente e promettente campo di ricerca delle BCI, facendo luce sulle caratteristiche dei componenti che le caratterizzano e sulle varie applicazioni a cui si prestano, concentrandosi sulle BCI-speller. A tal fine, sono state presentate BCI-speller basate su due particolari potenziali cerebrali (P300 e SSVEP), ponendo particolare attenzione sugli aspetti che possono portare ad un loro miglioramento. Maggiore enfasi è stata quindi posta sulle diverse GUI delle BCI-speller basate su questi potenziali, in quanto modifiche associate alla presentazione dello stimolo e alla facilità d’uso di queste interfacce possono migliorare in prima battuta il potenziale elicitato, e di conseguenza la prestazione generale della BCI. Infine, sono stati evidenziati i vantaggi e i limiti associati a queste tecnologie, nonché gli sviluppi futuri nella prospettiva di un impiego quotidiano di queste tecnologie da parte di pazienti affetti da disturbi neuromuscolari.

Neurological degenerative diseases like stroke, Alzheimer, Amyothrophic Lateral Sclerosis (ALS), Parkinson and many others are constantly increasing their incidence in the world health statistics as far as the mean age of the global population is getting higher and higher. This leads to a general need for effective, at-home and low-cost rehabilitative and health-daily-care tools. The latter should consist either of technological devices implemented for operating in a remote way, i.e. tele-medicine is quickly spreading around the world, or very-advanced computer-based and robotic systems to realize intense and repetitive trainings. This is the challenge in which Information and Communications Technology (ICT) is asked to play a major role in order to bring medicine to reach further advancements. Indeed, no way to cope with these issues is possible outside a strong and vivid cooperation among multi-disciplinary teams of clinicians, physicians, biologists, neuro-psychologists and engineers and without a resolute pushing towards a widespread inter-operability between Institutes, Hospitals and Universities all over the world, as recently highlighted during the main International conferences on ICT in healthcare. The establishment of well-defined standards for gathering and sharing data will then represent a key element to enhance the efficacy of the aforementioned collaborations. Among the others, stroke is one of the most common neurological pathologies being the second or third cause of mortality in the world; moreover, it causes more than sixty percent survivors remain with severe cognitive and motor impairments that impede them in living normal lives and require a twenty-four-hours daily care. As a consequence, on one side stroke survivors experience a frustrating condition of being completely dependent on other people even to perform simple daily actions like reach and grasp an object, hold a glass of water to drink it and so on. States, by their side, have to take into account additional costs to provide stroke patients and their families with appropriate cares and supports to cope with their needs. For this reason, more and more fundings are recently made available by means of grants, European and International projects, programs to exchange different expertise among various countries with the aim to study how to accelerate and make more effective the recovery process of chronic stroke patients. The global research about this topic is conducted on several parallel aspects: as regard as the basic knowledge of brain processes, neurophysiologists, biologists and engineers are particularly interested in an in-depth understanding of the so-called neuroplastic changes that brain daily operates in order to adapt individuals to life changes, experiences and to realize more extensively their own potentialities. Neuroplasticity is indeed the corner stone for most of the trainings nowadays adopted by the standard as well as the more innovative methods in the rehabilitative programs for post-stroke recovery. Specifically speaking, motor rehabilitation usually includes long term, repetitive and intense goal-directed exercises that promote neuroplastic mechanisms such as neural sprouting, synapto-genesis and dendritic branching. These processes are strictly related with motor improvements and their study could - one day - serve as prognostic measures of the recovery. Another aspect of this eld of neuroscience research is the number of applications that it makes feasible. One of the most exciting is to connect an injured brain to a computer or a robotic device in a Brain-Computer or Brain-Machine Interface (BCI or BMI) scheme aiming at bypassing the impairments of the patient and make him/her autonomously move again or train his/her motor abilities in a more effective way. This kind of research can already count an amount of literature that provides several proofs of concept that these heterogeneous systems constituted by humans and robots can work at the purpose. A particular application of BCI for restoring or enhancing, at least, the reaching abilities of chronic stroke survivors was implemented and is still currently being improved at I.R.C.C.S. San Camillo Hospital Foundation, an Institute for the rehabilitation from neurological diseases located in Lido of Venice and partially technically supported by the Department of Information Engineering of Padua in range of an agreement signed in 2009. This specific BCI platform allows patients to train and improve their reaching movements by means of a robotic arm that provides a force that helps patients in completing the training exercise, i.e. to hit a predetermined target. This force feedback is however subject to a strict condition: during the movement, the person has to produce the expected pattern of cerebral activity. Whenever this is accomplished, a force is delivered proportionally to the entity of the latter activity, otherwise the patient is obliged to operate without any help. In this way, this platform implements the so-called operant-learning, that is one of the most effective conditioning techniques to make a subject learn or re-learn a task. If, on one hand, the primary and explicit task is to improve a movement, on the other side the secondary but most important task is to deploy the perilesional part of the brain - still healthy - in becoming responsible for the control of the movement. It is a popular and widely-accepted opinion within the neuroscience community, indeed, that a healthy region of the sensorimotor area nearby the damaged one - which was previously in charge of performing the (reaching) movement - can optimally accomplish the impaired motor function substituting the original control area. Technically speaking, the main crucial feature that can ensure the effectiveness of the whole system is the precise and in real-time identification and quantification of the cerebral pattern associated with the movement, the worldwide named movement-related desynchronization (MRD). Starting from its original definition, passing through the most used techniques for its recognition, the thesis work presents a series of criticisms of the current signal processing method to detect the MRD and a complete analysis of the possible features that can better represent the movement condition and that can be more easily extracted during the on-line operations. Brain - it is well-known - learns by trials and errors and it needs a slightly-delayed (in the range of fraction of seconds) feedback of its performance to learn a task in the best way. This BCI application was born with the purpose to provide the above-mentioned feedback: however, this is only feasible if a computationally easy and contingent signal processing technique is available. This thesis work would like to cope with the lack of a well-planned real-time signal analysis in the current experimental protocol.
L'identificazione e la quantificazione in tempo reale dei correlati cerebrali del movimento e' uno degli aspetti piu' critici nell'ambito delle cosiddette Brain-Computer Interface (BCI), ovvero quelle applicazioni in cui un individuo, dopo uno specifico percorso di apprendimento, impara a controllare un computer (o un altro dispositivo) tramite la modulazione volontaria della sua attivita' cerebrale, con lo scopo finale di trarre vantaggio dall'utilizzo dell'apparecchiatura cosi' controllata. La BCI e' solo uno delle molteplici tecniche di riabilitazione motoria oggigiorno disponibili. Indipendentemente dalla specifica tecnica scelta, il metodo riabilitativo mira a recuperare le funzionalita' del cosiddetto sistema sensorimotorio, quel complesso di aree corticali e strutture sotto-corticali che permette ad un individuo di ricevere impulsi somatosensoriali dal mondo esterno, di elaborare la risposta motoria piu' opportuna e realizzarla grazie agli attuatori finali del movimento rappresentati dai muscoli. Per perseguire questo obbiettivo, la maggior parte delle tecniche riabilitative, ivi compresa la BCI, pongono grande attenzione alla promozione e sfruttamento di quei processi spontanei che il cervello costantemente impiega per svolgere le sue attivita', adattarsi a nuove condizioni ambientali e anche cercare di recuperare le abilita' perse a seguito di un evento dannoso quale un ictus. Questi fenomeni vengono generalmente riassunti nel termine neuroplasticita' e sono stati paradossalmente sfruttati per anni nella pratica clinica, ma solo recentemente sono diventati materia di rigorosi e approfonditi studi scientici portati avanti da neuroscienziati provenienti da ogni tipo di formazione (neurologi, neurofisiologi, biologi, ingegneri, neuropsicologi, ...). Nel particolare contesto della riabilitazione motoria che ambisce essenzialmente a promuovere fenomeni di (ri-)apprendimento motorio da parte del paziente colpito da ictus, la letteratura ha fermamente indicato il condizionamento operante come la strategia piu' efficace per favorire i processi di plasticita' corticale e, di conseguenza, il recupero, seppur parziale, della motricita'. Il condizionamento operante si applica tramite la ripetitiva associazione di un comportamento corretto (scorretto) effettuato dal soggetto e uno stimolo gratificante (penalizzante) dato contestualmente o in un tempo appena successivo all'esecuzione del comportamento. Sfruttando il fondamentale meccanismo di apprendimento del cervello per prove ed errori, esso impara a utilizzare il comportamento corretto per svolgere il compito richiesto. Tale comportamento corretto e', in questo caso, lo sfruttamento di risorse ridondanti prima dell'ictus ma ancora sane dopo l'evento per controllare le funzioni motorie rimaste indebolite o completamente perse a seguito del danno cerebrale. Data la fondamentale importanza della contingenza tra l'attivita' cerebrale del paziente coinvolto nell'esperimento e lo stimolo di feedback, un robusto algoritmo di elaborazione del segnale cerebrale si rende fortemente necessario. In questo lavoro di tesi e' stata analizzata una particolare applicazione BCI per la riabilitazione motoria di pazienti lievemente o moderatamente aetti da emiparesi dovuta a ictus e sono stati proposti degli algoritmi con la relativa soluzione software per la realizzazione ottimale della strategia di apprendimento operante durante il recupero di un movimento di raggiungimento. Il vincolo principale da considerare per ottenere questo tipo di risultato e' la possibilita' di identificare e quanticare il correlato neurofisiologico legato al movimento, la cosiddetta desincronizzazione movimento correlata, in tempo reale durante l'esecuzione del movimento da parte del soggetto. Nell'ottica del condizionamento operante, un dispositivo haptico inserito nel sistema funge da feedback positivo che aiuta il paziente a completare il movimento nel solo caso in cui venga riconosciuta la desincronizzazione. In caso contrario, il soggetto non riceve alcun feedback o il suo movimento viene reso piu' difficile. Dopo un primo capitolo introduttivo sul sistema sensorimotorio arricchito con alcune informazioni riguardanti la particolare patologia in questione, il secondo capitolo introduce gli elementi fondamentali della piattaforma sperimentale utilizzata, ovvero l'elettroencefalogramma (EEG) e la BCI nella sua accezione generale, accennando anche ai maggiori successi in questo ambito delle ICT applicate alla medicina riabilitativa. Nel capitolo 3 viene descritta la particolare piattaforma BCI basata su EEG menzionata sopra e nei capitoli 4 e 5 vengono presentate ampiamente le analisi, le procedure e i risultati sviluppati ed ottenuti in questi anni di studio. In particolare, nella prima parte del capitolo 4 viene illustrato l'algoritmo con cui e' possibile identificare e rimuovere in tempo reale un particolare evento di disturbo che puo' verificarsi durante la registrazione EEG, il cosiddetto artefatto electrode pop dovuto al temporaneo sollevamento di un elettrodo che causa abnormi valori negativi nelle tracce EEG. Una volta rimosso questo tipo di eventp, il segnale viene filtrato e nella seconda parte del capitolo 4 viene presentata un'esaustiva analisi dell'energia delle tracce EEG acquisite durante la registrazione dell'esperimento di cui sopra in soggetti sani di controllo e in alcuni pazienti reduci da ictus. Inoltre, viene suggerita una versione modificata del piu' noto metodo di quantificazione della desincronizzazione fornito da Pfurtscheller e colleghi a partire dagli anni '70 i cui risultati promettenti sono forniti e discussi nel capitolo finale della tesi. La tesi si conclude con una breve sezione dedicata alle prospettive future di applicazione della piattaforma con l'integrazione delle soluzioni software apportate da questa tesi e alle questioni ancora aperte da risolvere al fine di ottimizzare il sistema BCI in tutti i suoi aspetti in modo da realizzare nel modo piu efficace il condizionamento operante e promuovere quei processi spontanei che sottostanno al recupero funzionale della motricita'.

Submitted by Automa??o e Estat?stica (sst@bczm.ufrn.br) on 2017-10-02T22:27:19Z No. of bitstreams: 1 GabrielAlvesVasiljevicMendes_DISSERT.pdf: 3791566 bytes, checksum: e847396390a6b6ca2128eefd4423f561 (MD5)
Approved for entry into archive by Arlan Eloi Leite Silva (eloihistoriador@yahoo.com.br) on 2017-10-06T23:26:09Z (GMT) No. of bitstreams: 1 GabrielAlvesVasiljevicMendes_DISSERT.pdf: 3791566 bytes, checksum: e847396390a6b6ca2128eefd4423f561 (MD5)
Made available in DSpace on 2017-10-06T23:26:09Z (GMT). No. of bitstreams: 1 GabrielAlvesVasiljevicMendes_DISSERT.pdf: 3791566 bytes, checksum: e847396390a6b6ca2128eefd4423f561 (MD5) Previous issue date: 2017-07-31
Conselho Nacional de Desenvolvimento Cient?fico e Tecnol?gico (CNPq)
Brain-computer interfaces (BCIs) are specialized systems that allow users to control a computer or a machine using their brain waves. BCI systems allow patients with severe physical impairments, such as those suffering from amyotrophic lateral sclerosis, cerebral palsy and locked-in syndrome, to communicate and regain physical movements with the help of specialized equipment. With the development of BCI technology in the second half of the 20th century and the advent of consumer-grade BCI devices in the late 2000s, brain-controlled systems started to find applications not only in the medical field, but in areas such as entertainment. One particular area that is gaining more evidence due to the arrival of consumer-grade devices is the field of computer games, which has become increasingly popular in BCI research as it allows for more user-friendly applications of BCI technology in both healthy and unhealthy users. However, numerous challenges are yet to be overcome in order to advance in this field, as the origins and mechanics of the brain waves and how they are affected by external stimuli are not yet fully understood. In this sense, a systematic literature review of BCI games based on consumer-grade technology was performed. Based on its results, two BCI games, one using attention and the other using meditation as control signals, were developed in order to investigate key aspects of player interaction: the influence of graphical elements on attention and control; the influence of auditory stimuli on meditation and work load; and the differences both in performance and multiplayer game experience, all in the context of neurofeedback-based BCI games.

In the last two decades, Brain-Computer Interfaces (BCIs) have been investigated mainly for the purpose of implementing assistive technologies able to provide new channels for communication and control for people with severe disabilities. Nevertheless, more recently, thanks to technical and scientific advances in the different research fields involved, BCIs are gaining greater attention also for their adoption by healthy users, as new interaction devices. This thesis is dedicated to to the latter goal and in particular will deal with BCIs based on the Steady State Visual Evoked Potential (SSVEP), which in previous works demonstrated to be one of the most flexible and reliable approaches. SSVEP based BCIs could find applications in different contexts, but one which is particularly interesting for healthy users, is their adoption as new interaction devices for Virtual Reality (VR) environments and Computer Games. Although being investigated since several years, BCIs still poses several limitations in terms of speed, reliability and usability with respect to ordinary interaction devices. Despite of this, they may provide additional, more direct and intuitive, explicit interaction modalities, as well as implicit interaction modalities otherwise impossible with ordinary devices. This thesis, after a comprehensive review of the different research fields being the basis of a BCI exploiting the SSVEP modality, present a state-of-the-art open source implementation using a mix of pre-existing and custom software tools. The proposed implementation, mainly aimed to the interaction with VR environments and Computer Games, has then been used to perform several experiments which are hereby described as well. Initially performed experiments aim to stress the validity of the provided implementation, as well as to show its usability with a commodity bio-signal acquisition device, orders of magnitude less expensive than commonly used ones, representing a step forward in the direction of practical BCIs for end users applications. The proposed implementation, thanks to its flexibility, is used also to perform novel experiments aimed to investigate the exploitation of stereoscopic displays to overcome a known limitation of ordinary displays in the context of SSVEP based BCIs. Eventually, novel experiments are presented investigating the use of the SSVEP modality to provide also implicit interaction. In this context, a first proof of concept Passive BCI based on the SSVEP response is presented and demonstrated to provide information exploitable for prospective applications.

Cette thèse a pour but le développement d’une Interface cerveau-machine (ICM) à partir de la mesure EEG,permettant à l’utilisateur de communiquer avec un dispositif externe directement par l’intermédiaire de son activité cérébrale. Ces travaux ont été menés avec comme ligne directrice le développement d'un système d'ICM utilisable dans un contexte de vie courante, le but étant de réaliser une ICM simple d'utilisation, robuste et ergonomique, permettant le contrôle d'un effecteur avec un temps de calibration minimal.Un brain-switch ou interrupteur cérébral a été réalisé et permet à l'utilisateur d'envoyer une commande binaire. La réalisation d'une telle ICM implique le développement d'algorithmes robustes et leurs mises en œuvre expérimentales. Les travaux réalisés comportent deux volets, l'un concerne le développement de nouveaux algorithmes, l'autre concerne la réalisation de campagne de tests
This thesis presents the development of a Brain computer Interface (BCI) based on EEG signal, allowing its user to communicates with an external device solely by the mean of brain activity. This work as been conduct with the goal of designing a robust, ergonomic and easy to use BCI system for real life applications.In this context, a brain-switch has been developed, allowing it's user to send a binary command to a homeautomation system. This goal can only be achieved by developing new methodologies and algorithms, while testing them on real life experiments. Therefore, this works is two part, the first one is focus on the design of new algorithms, the secondon the design of experimental paradigm

Les interfaces cerveau machines ont quitté le domaine de la science-fiction dans les années 70 avec la réflexion portée par Jacques Vidal sur la faisabilité d’utiliser des signaux électro-encéphalogrammes comme moyen de communication entre le cerveau et des dispositifs extérieurs. Pendant le développement de ce domaine de recherche, différentes pistes ont été explorées afin de créer une véritable interface entre le cerveau et une machine. Les méthodes d’acquisition ont pris différentes formes et avec l'amélioration des capacités informatiques et l'avènement de l'apprentissage automatique, les méthodes de classification des données cérébrales sont devenues plus fines, capables de capturer des informations complexes sur la donnée cérébrale. Une piste prometteuse pour aider les patients concerne le processus de rééducation suivant un accident vasculaire cérébrale (AVC). En effet, en effectuant des mouvements ou en informant les patients de leur activité cérébrale, il est possible d’aider leur cerveau à s’adapter au déficit neuronal et ainsi de les aider à surmonter leur lésion. Dans ce contexte, les ICO jouent le rôle d'une béquille sur lequel s’appuyer le temps de la rééducation. Dans tout cela, comment devrions-nous contribuer aux systèmes BCI ? Un défi particulier pour les sujets est de créer des patterns cérébraux différenciables au niveau de l'EEG. Pour créer ces patterns, nous nous appuyons sur des tâches cognitives qui modifient le profil d'activité cérébrale. Nous pouvons évoquer l'imagerie motrice des mouvements des membres, une tâche qui consiste à imaginer des mouvements sans les exécuter. Cette tâche particulière est peu familière pour beaucoup et donc en devient complexe à exécuter. Une manière d'aider les sujets à l'effectuer est d'utiliser des retours évocateurs dans le contexte d’une ICO, tels qu’un bras robotique. Le fait de nourrir le sentiment de contrôle sur le bras ainsi que sur les mouvements produits par celui-ci contribue à obtenir des patterns cérébraux plus différenciables. Cependant, en raison des limitations actuelles en ce qui concerne le degré de contrôle permis par les systèmes d’ICO, un contrôle complet d'un bras robotique n'est pas possible. Une solution est de coupler l’ICO avec une autre technologie pour augmenter le degré de contrôle et renforcer le sentiment d’agentivité des sujets sur le bras. Parmi les technologies qui offrent un aperçu de l'intention des sujets, l’oculomètre semble être une solution élégante. L'intégration des deux composants crée un système BCI hybride capable de contrôler le bras de manière intuitive. Cette hybridation a été développée précédemment en tant que preuve de concept, mais l'impact de l'intégration de ces modalités sur le cerveau reste à étudier. En effet, une meilleure compréhension de la manière dont nous devrions façonner l'interaction entre les éléments permettrait de savoir pourquoi nous obtenons de bonnes performances et comment susciter ces patterns discriminants. Le travail de cette thèse consistait en la création d'une plateforme expérimentale entrelaçant les différentes modalités pour établir un contrôle robuste sur le bras. En réalisant un protocole expérimental, nous avons évalué comment définir l'interaction par une analyse approfondie couvrant la performance pure du système, les réponses physiologiques et neurophysiologiques des sujets. Nous avons constaté que la constance était essentielle dans l'interaction et nous avons démontré l'importance du mouvement pour susciter des réponses cérébrales dans ce contexte particulier. Notre travail ouvre la voie à une meilleure compréhension de la dynamique du cerveau dans son contrôle sur les dispositifs externes dans une configuration multimodale. Cette thèse est structurée en cinq chapitres différents couvrant le contexte des ICOs, le développement de la plateforme expérimentale, les résultats du protocole expérimental associés à leur discussion, et enfin une conclusion générale
Brain-machine interfaces (BMIs) left the realm of science fiction in the 1970s with Jacques Vidal's reflection on the feasibility of using electroencephalogram signals as a means of communication between the brain and external devices. During the development of this research field, various approaches have been explored to create a true interface between the brain and a machine. The methods of data acquisition have taken different forms, and with the improvement of computer capabilities and the advent of machine learning, methods to classify brain data have become more refined, capable of capturing complex information from brain data. A promising avenue to assist patients lies in the rehabilitation process following a stroke. By performing movements or providing feedback to patients about their brain activity, it is possible to help their brains adapt to neuronal deficits and aid them in overcoming their impairments. In this context, BMIs act as a support, like crutches, during the rehabilitation period. However, in this context, one challenge is to create differentiable brain patterns at the EEG level. To create these patterns, cognitive tasks that modify brain activity profiles are relied upon. One such task is motor imagery of limb movements, where subjects imagine movements without executing them. This particular task is unfamiliar to many and thus becomes complex to execute. One way to assist subjects in performing it is to provide evocative feedback in the BMI context, such as a robotic arm. Nurturing a sense of control over the arm and its movements helps elicit more differentiable brain patterns. Nevertheless, due to current limitations in the degree of control permitted by BMI systems, full control of a robotic arm is not yet possible. A solution is to couple the BMI with another technology to increase the degree of control and reinforce subjects' sense of agency over the arm. Among the technologies offering insights into subjects' intentions without requiring movement, the eye tracker appears to be an elegant solution. The integration of these two components creates a hybrid BCI system capable of intuitively controlling the arm. This hybridization has been demonstrated as a proof of concept in the BMI field, but the impact of integrating these modalities on the brain remains to be studied. A better understanding of how to shape the interaction between the eye tracker, BMI, and robotic arm would shed light on why good performances are obtained and how to elicit these discriminant brain patterns. The work in this thesis focused on creating an experimental platform that intertwines these different modalities to establish robust control over the arm. Through an experimental protocol, we assessed how to define the interaction through in-depth analysis, covering the system's pure performance and the physiological and neurophysiological responses of subjects. We found that consistency is crucial in the interaction, and we demonstrated the importance of movement in eliciting brain responses in this particular context. Our work paves the way for a better understanding of the brain's dynamics in controlling external devices in a multimodal setup. Additionally, we propose a new framework for controlling a robotic arm using a hybrid BCI. The thesis is structured into five chapters, covering the overall context of BMIs, the development of the experimental platform, the results from the experimental protocol and their discussion, and finally, a general conclusion

Cette thèse s'inscrit dans le cadre du projet Européen intégré VERE (Virtual Embodiement and Robotics re-Embodiement). Il s'agit de proposer une architecture logicielle intégrant un ensemble de stratégies de contrôle et de retours informationnels basés sur la "fonction tâche" pour incorporer (embodiment) un opérateur humain dans un humanoïde robotique ou un avatar notamment par la pensée. Les problèmes sous-jacents peuvent se révéler par le démonstrateur suivant (auquel on souhaite aboutir à l'issue de cette thèse). Imaginons un opérateur doté d'une interface cerveau-ordinateur ; le but est d'arriver à extraire de ces signaux la pensée de l'opérateur humain, de la traduire en commandes robotique et de faire un retour sensoriel afin que l'opérateur s'approprie le "corps" robotique ou virtuel de son "avatar". Une illustration cinématographique de cet objectif est le film récent "Avatar" ou encore "Surrogates". Dans cette thèse, on s'intéressera tout d'abord à certains problèmes que l'on a rencontré en travaillant sur l'utilisation des interfaces cerveau-ordinateur pour le contrôle de robots ou d'avatars, par exemple, la nécessité de multiplier les comportements ou les particularités liées aux retours sensoriels du robot. Dans un second temps, nous aborderons le cœur de notre contribution en introduisant le concept d'interface cerveau-ordinateur orienté objet pour le contrôle de robots humanoïdes. Nous présenterons ensuite les résultats d'une étude concernant le rôle du son dans le processus d'embodiment. Enfin, nous montrerons les premières expériences concernant le contrôle d'un robot humanoïde en interface cerveau-ordinateur utilisant l'électrocorticographie, une technologie d'acquisition des signaux cérébraux implantée dans la boîte crânienne
This thesis is part of the European project VERE (Virtual Embodiment and Robotics re-Embodiment). The goal is to propose a software framework integrating a set of control strategies and information feedback based on the "task function" in order to embody a human operator within a humanoid robot or a virtual avatar using his thoughts. The underlying problems can be shown by considering the following demonstrator. Let us imagine an operator equipped with a brain-computer interface; the goal is to extract the though of the human operator from these signals, then translate it into robotic commands and finally to give an appropriate sensory feedback to the operator so that he can appropriate the "body", robotic or virtual, of his avatar. A cinematographic illustration of this objective can be seen in recent movies such as "Avatar" or "Surrogates". In this thesis, we start by discussing specific problems that we encountered while using a brain-computer interface for the control of robots or avatars, e.g. the arising need for multiple behaviours or the specific problems induced by the sensory feedback provided by the robot. We will then introduce our main contribution which is the concept of object-oriented brain-computer interface for the control of humanoid robot. We will then present the results of a study regarding the role of sound in the embodiment process. Finally, we show some preliminary experiments where we used electrocorticography (ECoG)~--~a technology used to acquire signals from the brain that is implanted within the cranium~--~to control a humanoid robot

There are hundreds of thousands of people who could benefit from a Brain-Computer Interface. However, not all are willing to undergo surgery, so an EEG is the prime candidate for use as a BCI. The features of Event-Related Desynchronization and Synchronization could be used for a switch and have been in the past. A new method of feature selection was proposed to optimize classification of active motor movement vs a non-active idle state. The previous method had pre-selected which frequency and electrode to use as electrode C3 at the 20Hz bin. The new method used SPSS statistical software to determine the most significant frequency and electrode combination. This improved method found increased accuracy in classifying cases as either active or idle states. Future directions could be using multiple features for classification and BCI control, or exploiting the difference between ERD and ERS, though for either of these a more advanced algorithm would be required.

Les Interfaces Cerveau Ordinateur (ICO) permettent l’interaction à partir de l’activité cérébrale. La Réalité Augmentée (RA) elle, permet d’intégrer des éléments virtuels dans un environnement réel. Dans cette thèse, nous avons cherché à concevoir des systèmes interactifs exploitant des ICO dans des environnements RA, afin de proposer de nouveaux moyens d’interagir avec des éléments réels et virtuels. Dans la première partie de cette thèse, nous avons étudié la possibilité d’extraire différents signaux cérébraux dans un contexte de RA. Nous avons ainsi montré qu’il était possible d’exploiter les Potentiels Evoqués Visuels Stationnaires (SSVEP) en RA. Puis, nous avons montré la possibilité d’extraire des Potentiels d’Erreur des signaux cérébraux, lorsqu’un utilisateur est soumis à des types d’erreurs fréquents en RA. Dans la seconde partie, nous avons approfondi nos recherches sur l’utilisation des SSVEP pour l’interaction en RA. Nous avons notamment proposé HCCA, un nouvel algorithme permettant la reconnaissance asynchrone de réponses SSVEP. Nous avons ensuite étudié la conception d’interfaces de RA, pour des systèmes interactifs, intuitifs performants. Enfin nous avons illustré nos résultats à travers le développement d’un système de domotique utilisant les SSVEP et la RA, qui s’intègre à une plateforme de maison intelligente industrielle
Brain-Computer Interfaces (BCI) enable interaction directly from brain activity. Augmented Reality (AR) on the other hand, enables the integration of virtual elements in the real world. In this thesis, we aimed at designing interactive systems associating BCIs and AR, to offer new means of hands-free interaction with real and virtual elements. In the first part, we have studied the possibility to extract different BCI paradigms in AR. We have shown that it was possible to use Steady-State Visual Evoked Potentials (SSVEP) in AR. Then, we have studied the possibility to extract Error-Related Potentials (ErrPs) in AR, showing that ErrPs were elicited in users facing errors, often occurring in AR. In the second part, we have deepened our research in the use of SSVEP for direct interaction in AR. We have proposed HCCA, a new algorithm for self-paced detection of SSVEP responses. Then, we have studied the design of AR interfaces, for the development of intuitive and efficient interactive systems. Lastly, we have illustrated our results, through the development of a smart-home system combining SSVEP and AR, which integrates in a commercially available smart-home system

Cette thèse présente trois contributions pour améliorer la reconnaissance d’imaginations motrices utilisées par de nombreuses interfaces cerveau-ordinateur (BCI) comme moyen d'interaction. Tout d'abord, nous proposons d'estimer la qualité des images motrices en détectant des valeurs aberrantes et de les supprimer avant apprentissage. Ensuite, nous étudions la sélection des caractéristiques pour sept imaginations de mouvements. Enfin, nous présentons une architecture d'apprentissage profond reprenant les principes du réseaux EEGnet applicable directement sur des signaux électro-encéphalographiques simplement filtrés et adapté au nombre d’électrodes. Nous montrons en particulier ses bénéfices pour l'amélioration de la détection des réveils peropératoires et d'autres applications
This thesis presents three contributions to improve the recognition of motor imaginary movements used by numerous brain-computer interfaces (BCI) as types of interaction. First of all, we propose to estimate the quality of motor images by detecting outliers and removing them before training. Next, we study the feature selection for seven different motor imaginary movements. Finally, we present a deep learning architecture based on the principles of EEGNet network applied directly on raw electroencephalographic signals and adapted to the number of electrodes. We show in particular its benefits for improving the detection of intraoperative awareness and other applications

Restorer la faculté de parler chez des personnes paralysées et aphasiques pourrait être envisagée via l’utilisation d’une interface cerveau-machine permettant de contrôler un synthétiseur de parole en temps réel. L’objectif de cette thèse était de développer trois aspects nécessaires à la mise au point d’une telle preuve de concept.Premièrement, un synthétiseur permettant de produire en temps-réel de la parole intelligible et controlé par un nombre raisonable de paramètres est nécessaire. Nous avons choisi de synthétiser de la parole à partir des mouvements des articulateurs du conduit vocal. En effet, des études récentes ont suggéré que l’activité neuronale du cortex moteur de la parole pourrait contenir suffisamment d’information pour décoder la parole, et particulièrement ses propriété articulatoire (ex. l’ouverture des lèvres). Nous avons donc développé un synthétiseur produisant de la parole intelligible à partir de données articulatoires. Dans un premier temps, nous avons enregistré un large corpus de données articulatoire et acoustiques synchrones chez un locuteur. Ensuite, nous avons utilisé des techniques d’apprentissage automatique, en particulier des réseaux de neurones profonds, pour construire un modèle permettant de convertir des données articulatoires en parole. Ce synthétisuer a été construit pour fonctionner en temps réel. Enfin, comme première étape vers un contrôle neuronal de ce synthétiseur, nous avons testé qu’il pouvait être contrôlé en temps réel par plusieurs locuteurs, pour produire de la parole inetlligible à partir de leurs mouvements articulatoires dans un paradigme de boucle fermée.Deuxièmement, nous avons étudié le décodage de la parole et de ses propriétés articulatoires à partir d’activités neuronales essentiellement enregistrées dans le cortex moteur de la parole. Nous avons construit un outil permettant de localiser les aires corticales actives, en ligne pendant des chirurgies éveillées à l’hôpital de Grenoble, et nous avons testé ce système chez deux patients atteints d’un cancer du cerveau. Les résultats ont montré que le cortex moteur exhibe une activité spécifique pendant la production de parole dans les bandes beta et gamma du signal, y compris lors de l’imagination de la parole. Les données enregistrées ont ensuite pu être analysées pour décoder l’intention de parler du sujet (réelle ou imaginée), ainsi que la vibration des cordes vocales et les trajectoires des articulateurs principaux du conduit vocal significativement au dessus du niveau de la chance.Enfin, nous nous sommes intéressés aux questions éthiques qui accompagnent le développement et l’usage des interfaces cerveau-machine. Nous avons en particulier considéré trois niveaux de réflexion éthique concernant respectivement l’animal, l’humain et l’humanité
Restoring natural speech in paralyzed and aphasic people could be achieved using a brain-computer interface controlling a speech synthesizer in real-time. The aim of this thesis was thus to develop three main steps toward such proof of concept.First, a prerequisite was to develop a speech synthesizer producing intelligible speech in real-time with a reasonable number of control parameters. Here we chose to synthesize speech from movements of the speech articulators since recent studies suggested that neural activity from the speech motor cortex contains relevant information to decode speech, and especially articulatory features of speech. We thus developed a speech synthesizer that produced intelligible speech from articulatory data. This was achieved by first recording a large dataset of synchronous articulatory and acoustic data in a single speaker. Then, we used machine learning techniques, especially deep neural networks, to build a model able to convert articulatory data into speech. This synthesizer was built to run in real time. Finally, as a first step toward future brain control of this synthesizer, we tested that it could be controlled in real-time by several speakers to produce intelligible speech from articulatory movements in a closed-loop paradigm.Second, we investigated the feasibility of decoding speech and articulatory features from neural activity essentially recorded in the speech motor cortex. We built a tool that allowed to localize active cortical speech areas online during awake brain surgery at the Grenoble Hospital and tested this system in two patients with brain cancer. Results show that the motor cortex exhibits specific activity during speech production in the beta and gamma bands, which are also present during speech imagination. The recorded data could be successfully analyzed to decode speech intention, voicing activity and the trajectories of the main articulators of the vocal tract above chance.Finally, we addressed ethical issues that arise with the development and use of brain-computer interfaces. We considered three levels of ethical questionings, dealing respectively with the animal, the human being, and the human species

Thesis (MEng)--Stellenbosch University, 2014.
ENGLISH ABSTRACT: Muscular dystrophy, spinal cord injury, or amyotrophic lateral sclerosis (ALS) are injuries and disorders that disrupts the neuromuscular channels of the human body thus prohibiting the brain from controlling the body. Brain computer interface (BCI) allows individuals to bypass the neuromuscular channels and interact with the environment using the brain. The system relies on the user manipulating his neural activity in order to control an external device. Electroencephalography (EEG) is a cheap, non-invasive, real time acquisition device used in BCI applications to record neural activity. However, noise, known as artifacts, can contaminate the recording, thus distorting the true neural activity. Eye blinks are a common source of artifacts present in EEG recordings. Due to its large amplitude it greatly distorts the EEG data making it difficult to interpret data for BCI applications. This study proposes a new combination of techniques to detect and correct eye blink artifacts to improve the quality of EEG for BCI applications. Independent component analysis (ICA) is used to separate the EEG signals into independent source components. The source component containing eye blink artifacts are corrected by detecting each eye blink within the source component and using a trained wavelet neural network (WNN) to correct only a segment of the source component containing the eye blink artifact. Afterwards, the EEG is reconstructed without distorting or removing the source component. The results show a 91.1% detection rate and a 97.9% correction rate for all detected eye blinks. Furthermore for channels located over the frontal lobe, eye blink artifacts are corrected preserving the neural activity. The novel combination overall reduces EEG information lost, when compared to existing literature, and is a step towards improving EEG pre-processing in order to provide cleaner EEG data for BCI applications.
AFRIKAANSE OPSOMMING: Spierdistrofie, ’n rugmurgbesering, of amiotrofiese laterale sklerose (ALS) is beserings en steurnisse wat die neuromuskulêre kanale van die menslike liggaam ontwrig en dus verhoed dat die brein die liggaam beheer. ’n Breinrekenaarkoppelvlak laat toe dat die neuromuskulêre kanale omlei word en op die omgewing reageer deur die brein. Die BCI-stelsel vertrou op die gebruiker wat sy eie senuwee-aktiwiteit manipuleer om sodoende ’n eksterne toestel te beheer. Elektro-enkefalografie (EEG) is ’n goedkoop, nie-indringende, intydse dataverkrygingstoestel wat gebruik word in BCI toepassings. Nie net senuwee aktiwiteit nie, maar ook geraas , bekend as artefakte word opgeneem, wat dus die ware senuwee aktiwiteit versteur. Oogknip artefakte is een van die algemene artefakte wat teenwoordig is in EEG opnames. Die groot omvang van hierdie artefakte verwring die EEG data wat dit moeilik maak om die data te ontleed vir BCI toepassings. Die studie stel ’n nuwe kombinasie tegnieke voor wat oogknip artefakte waarneem en regstel om sodoende die kwaliteit van ’n EEG vir BCI toepassings te verbeter. Onafhanklike onderdeel analise (Independent component analysis (ICA)) word gebruik om die EEG seine te skei na onafhanklike bron-komponente. Die bronkomponent wat oogknip artefakte bevat word reggestel binne die komponent en gebruik ’n ervare/geoefende golfsenuwee-netwerk om slegs ’n deel van die komponent wat die oogknip artefak bevat reg te stel. Daarna word die EEG hervorm sonder verwringing of om die bron-komponent te verwyder. Die resultate toon ’n 91.1% opsporingskoers en ’n 97.9% regstellingskoers vir alle waarneembare oogknippe. Oogknip artefakte in kanale op die voorste lob word reggestel en behou die senuwee aktiwiteit wat die oorhoofse EEG kwaliteit vir BCI toepassings verhoog.

Les Interfaces Cerveaux-Ordinateur (ICO) sont des systèmes qui permettent de manipuler une machine avec sa seule activité cérébrale. Elles sont utilisées pour accomplir des objectifs variés, par exemple afin qu’un amputé puisse manipuler un bras robotique, pour une réhabilitation neuronale en cas d’accident vasculaire cérébral, dans un cadre ludique pour jouer à des jeux vidéo, etc. Une ICO comprend l’acquisition du signal cérébral (le plus souvent par électroencéphalographie, EEG), le décodage et l’interprétation de ce signal, et enfin la production d’un retour sensoriel à l’utilisateur. Ce retour guidera l’utilisateur pour réguler son activité cérébral et apprendre à manipuler la machine. La morphologie du cerveau diffère cependant entre utilisateurs, et les pensées d’un même individu varient au cours du temps. Ces fluctuations rendent les ICO moins performantes, qui sont alors difficiles à utiliser hors des conditions du laboratoire. Nous avons donc besoin d’une machine dynamique, qui puisse s’adapter au cours du temps à son utilisateur. Dans la littérature les approches proposées afin de remédier à ce problème décrivent des machines qui décodent de manière adaptative les signaux EEG, mais ces systèmes ne sont pas assez robustes et ne permettent toujours pas aux ICO d’être utilisées dans la vie quotidienne. L’objectif de cette thèse est d’améliorer les performances et l’utilisabilité des ICO basées sur de l’EEG, en les adaptant de façon innovante aux états et compétences des utilisateurs. Pour ce faire, nous avons premièrement mis en évidence tous les facteurs changeants dans une ICO en définissant trois séquences : 1. Les états psychologiques fluctuants de l’utilisateur qui modifient la signature du signal EEG ; 2. Ce signal qui varie et qui amène la machine à ajuster son décodage ; 3. La tâche qui est présentée à l’utilisateur via le retour sensoriel de la machine, et qui influence à son tour les états psychologiques de l’utilisateur. Nous avons ainsi mis en évidence la possibilité d’adopter un nouvel angle de recherche, en utilisant la tâche adaptative pour diriger les états psychologiques de l’utilisateur et aider ce dernier à manipuler une ICO. Au lieu de seulement adapter le décodage aux signaux cérébraux, nous avons donc considéré l’adaptation de l’interface (via le retour sensoriel produit par la machine) afin d’influencer les signaux et d’en faciliter le décodage. En utilisant des connaissances issues de la psychologie comportementale et des sciences de l’éducation, il est en effet possible de créer des taches et des interfaces qui incitent les utilisateurs à réussir et même à prendre plaisir à utiliser une ICO. Ces différents facteurs, liés à la motivation, participent à produire des signaux plus prédictibles et plus facilement décodables par la machine, augmentant d’autant la performance du système. Nous avons donc formulé une taxonomie des ICO adaptatives en définissant la tâche adaptative comme un nouveau moyen d’améliorer les performances des ICO. Une fois que la taxonomie des ICO adaptatives a été mis en place, nous avons cherché à identifier chez l’utilisateur quel était l’état psychologique optimal qui puisse servire de critère d’optimisation de la tâche. La littérature en psychologie indique que cet état est l’état de flow, un état d’immersion, de contrôle et de plaisir optimal qui incite les gens à se surpasser, quel que soit la tâche, le sexe, la culture ou bien encore l’âge. [...]
Brain-Computer Interfaces (BCIs) are systems that enable a person to manipulate an external device with only brain activity, often using ElectroEncephaloGraphgy (EEG). Although there is great medical potential (communication and mobility assistance, as well as neuro-rehabilitation of those who lost motor functions), BCIs are rarely used outside of laboratories. This is mostly due to users’ variability from their brain morphologies to their changeable psychological states, making it impossible to create one system that works with high success for all. The success of a BCI depends tremendously on the user’s ability to focus to give mental commands, and the machine’s ability to decode such mental commands. Most approaches consist in either designing more intuitive and immersive interfaces to assist the users to focus, or enhancing the machine decoding properties. The latest advances in machine decoding are enabling adaptive machines that try to adjust to the changeable EEG during the BCI task. This thesis is unifying the adaptive machine decoding approaches and the interface design through the creation of adaptive and optimal BCI tasks according to user states and traits. Its purpose is to improve the performance and usability of BCIs and enable their use outside of laboratories. To such end, we first created a taxonomy for adaptive BCIs to account for the various changeable factors of the system. Then, we showed that by adapting the task difficulty we can influence a state of flow, i.e., an optimal state of immersion, control and pleasure. which in turn correlates with BCI performance. Furthermore, we have identified the user traits that can benefit from particular types of task difficulties. This way we have prior knowledge that can guide the task adaptation process, specific to each user trait. As we wish to create a generic adaptation rule that works for all users, we use a probabilistic Bayesian model, called Active Inference used in neuroscience to computationally model brain behavior. When we provide such probabilistic model to the machine, it becomes adaptive in such a way that it mimics brain behavior. That way, we can achieve an automatic co-adaptive BCI and potentially get a step closer into using BCIs in our daily lives

This thesis study focuses on the realization of an online cue based Motor Imagery (MI) Brain Computer Interface (BCI). For this purpose, some signal processing and classification methods are investigated. Specifically, several time-spatial-frequency methods, namely the Short Time Fourier Transform (STFT), Common Spatial Frequency Patterns (CSFP) and the Morlet Transform (MT) are implemented on a 2-class MI BCI system. Distinction Sensitive Learning Vector Quantization (DSLVQ) method is used as a feature selection method. The performance of these methodologies is evaluated with the linear and nonlinear Support Vector Machines (SVM), Multilayer Perceptron (MLP) and Naive Bayesian (NB) classifiers. The methodologies are tested on BCI Competition IV dataset IIb and an average kappa value of 0.45 is obtained on the dataset. According to the classification results, the algorithms presented here obtain the 4th level in the competition as compared to the other algorithms in the competition. Offline experiments are performed in METU Brain Research Laboratories and Hacettepe Biophysics Department on two subjects with the original cue-based MI BCI paradigm. Average prediction accuracy of the methods on a 2-class BCI is evaluated to be 76.26% in these datasets. Furthermore, two online BCI applications are developed: the ping-pong game and the electrical wheelchair control. For these applications, average classification accuracy is found to be 70%. During the offline experiments, the performance of the developed system is observed to be highly dependent on the subject training and experience. According to the results, the EEG channels P3 and P4, which are considered to be irrelevant with the motor imagination, provided the best classification performance on the offline experiments. Regarding the observations on the experiments, this process is related to the stimulation mechanism in the cue based applications and consequent visual evoking effects on the subjects.

A Real-Time Brain-Computer Interface is a technical system classifying increased or decreased brain activity in Real-Time between different body movements, actions performed by a person. Focus in this thesis will be on testing algorithms and settings, finding the initial time interval and how increased activity in the brain can be distinguished and satisfyingly classified. The objective is letting the system give an output somewhere within 250ms of a thought of an action, which will be faster than a persons reaction time.

Algorithms in the preprocessing were Blind Signal Separation and the Fast Fourier Transform. With different frequency and time interval settings the algorithms were tested on an offline Electroencephalographic data file based on the "Ten Twenty" Electrode Application System, classified using an Artificial Neural Network.

A satisfying time interval could be found between 125-250ms, but more research is needed to investigate that specific interval. A reduction in frequency resulted in a lack of samples in the sample window preventing the algorithms from working properly. A high frequency is therefore proposed to help keeping the sample window small in the time domain. Blind Signal Separation together with the Fast Fourier Transform had problems finding appropriate correlation using the Ten-Twenty Electrode Application System. Electrodes should be placed more selectively at the parietal lobe, in case of requiring motor responses.

Une Interface Cerveau-Ordinateur (ICO) est un système de communication qui permet à ses utilisateurs d'envoyer des commandes à un ordinateur via leur activité cérébrale, cette activité étant mesurée, généralement par ÉlectroEncéphaloGraphie (EEG), et traitée par le système. Dans la première partie de cette thèse, dédiée au traitement et à la classification des signaux EEG, nous avons cherché à concevoir des ICOs interprétables et plus efficaces. Pour ce faire, nous avons tout d'abord proposé FuRIA, un algorithme d'extraction de caractéris- tiques utilisant les solutions inverses. Nous avons également proposé et étudié l'utilisation des Systèmes d'Inférences Flous (SIF) pour la classification. Nos évaluations ont montré que FuRIA et les SIF pouvaient obtenir de très bonnes performances de classification. De plus, nous avons proposé une méthode utilisant ces deux algorithmes afin de concevoir une ICO complétement interprétable. Enfin, nous avons proposé de considérer la conception d'ICOs asynchrones comme un problème de rejet de motifs. Notre étude a introduit de nouvelles techniques et a permis d'identifier les classifieurs et les techniques de rejet les plus appropriés pour ce problème. Dans la deuxième partie de cette thèse, nous avons cherché à concevoir des applications de Réalité Virtuelle (RV) controlées par une ICO. Nous avons tout d'abord étudié les performances et les préférences de participants qui interagissaient avec une application ludique de RV à l'aide d'une ICO asynchrone. Nos résultats ont mis en évidence le besoin d'utiliser des ICO adaptées à l'utilisateur ainsi que l'importance du retour visuel. Enfin, nous avons développé une application de RV permettant à un utilisateur d'explorer un musée virtuel par la pensée. Dans ce but, nous avons conçu une ICO asynchrone et proposé une nouvelle technique d'interaction permettant à l'utilisateur d'envoyer des commandes de haut niveau. Une première évaluation semble montrer que l'utilisateur peut explorer le musée plus rapidement avec cette technique qu'avec les techniques actuelles.

L'électroencéphalographie (EEG) mesure de manière non invasive l'activité électrique du cerveau par le biais de champs électromagnétiques générés par l'activité synchronisée de millions de neurones. Cela permet de collecter des données temporelles multivariées qui constituent une trace de l'activité électrique du cerveau mesurée au niveau du cuir chevelu. À tout instant, les mesures enregistrées par ces capteurs sont des combinaisons linéaires des activités électriques provenant d'un ensemble de sources sous-jacentes situées dans le cortex cérébral. Ces sources interagissent entre elles selon un modèle biophysique complexe qui reste mal compris. Dans certaines applications, telles que la planification chirurgicale, il est crucial de reconstruire avec précision ces sources électriques corticales, une tâche connue sous le nom de résolution du problème inverse de reconstruction de sources. Bien qu'intellectuellement satisfaisante et potentiellement plus précise, cette approche nécessite le développement et l'application d'un modèle spécifique au sujet, ce qui est à la fois coûteux et techniquement difficile à réaliser. Il est cependant souvent possible d'utiliser directement les mesures EEG au niveau des capteurs et d'en extraire des informations sur l'activité cérébrale. Cela réduit considérablement la complexité de l'analyse des données par rapport aux approches au niveau des sources. Ces mesures peuvent être utilisées pour une variété d'applications comme par exemple la surveillance des états cognitifs, le diagnostic des conditions neurologiques ou le développement d'interfaces cerveau-ordinateur (BCI). De fait, même sans avoir une compréhension complète des signaux cérébraux, il est possible de créer une communication directe entre le cerveau et un appareil externe à l'aide de la technologie BCI. Le travail décrit dans ce document est centré sur les interfaces cerveau-ordinateur basées sur l'EEG, qui ont plusieurs applications dans divers domaines médicaux, comme la réadaptation et la communication pour les personnes handicapées, ou dans des domaines non médicaux, notamment les jeux et la réalité virtuelle. La première contribution de cette thèse va dans ce sens, avec la proposition d'une méthode basée sur une matrice de covariance augmentée (ACM). Sur cette base, la méthode de covariance augmentée Block-Toeplitz (BT-ACM) représente une évolution notable, améliorant l'efficacité de calcul tout en conservant son efficacité et sa versatilité. Enfin, ce travail se poursuit avec la proposition d'un réseau de neurones artificiel Phase-SPDNet qui permet l'intégration de ces méthodologies dans une approche de Deep Learning et qui est particulièrement efficace même avec un nombre limité d'électrodes. Nous avons en outre proposé le cadre pseudo-on-line pour mieux caractériser l'efficacité des méthodes BCI et la plus grande étude de reproductibilité BCI basée sur l'EEG en utilisant le benchmark MOABB (Mother of all BCI Benchmarks). Cette recherche vise à promouvoir une plus grande reproductibilité et fiabilité des études BCI. En conclusion, nous relevons dans cette thèse deux défis majeurs dans le domaine des interfaces cerveau-ordinateur (BCI) basées sur l'EEG : l'amélioration des performances par le développement d'algorithmes avancés au niveau des capteurs et l'amélioration de la reproductibilité au sein de la communauté BCI
Electroencephalography (EEG) non-invasively measures the brain's electrical activity through electromagnetic fields generated by synchronized neuronal activity. This allows for the collection of multivariate time series data, capturing a trace of the brain electrical activity at the level of the scalp. At any given time instant, the measurements recorded by these sensors are linear combinations of the electrical activities from a set of underlying sources located in the cerebral cortex. These sources interact with one another according to a complex biophysical model, which remains poorly understood. In certain applications, such as surgical planning, it is crucial to accurately reconstruct these cortical electrical sources, a task known as solving the inverse problem of source reconstruction. While intellectually satisfying and potentially more precise, this approach requires the development and application of a subject-specific model, which is both expensive and technically demanding to achieve.However, it is often possible to directly use the EEG measurements at the level of the sensors and extract information about the brain activity. This significantly reduces the data analysis complexity compared to source-level approaches. These measurements can be used for a variety of applications, including monitoring cognitive states, diagnosing neurological conditions, and developing brain-computer interfaces (BCI). Actually, even though we do not have a complete understanding of brain signals, it is possible to generate direct communication between the brain and an external device using the BCI technology. This work is centered on EEG-based BCIs, which have several applications in various medical fields, like rehabilitation and communication for disabled individuals or in non-medical areas, including gaming and virtual reality.Despite its vast potential, BCI technology has not yet seen widespread use outside of laboratories. The primary objective of this PhD research is to try to address some of the current limitations of the BCI-EEG technology. Autoregressive models, even though they are not completely justified by biology, offer a versatile framework to effectively analyze EEG measurements. By leveraging these models, it is possible to create algorithms that combine nonlinear systems theory with the Riemannian-based approach to classify brain activity. The first contribution of this thesis is in this direction, with the creation of the Augmented Covariance Method (ACM). Building upon this foundation, the Block-Toeplitz Augmented Covariance Method (BT-ACM) represents a notable evolution, enhancing computational efficiency while maintaining its efficacy and versatility. Finally, the Phase-SPDNet work enables the integration of such methodologies into a Deep Learning approach that is particularly effective with a limited number of electrodes.Additionally, we proposed the creation of a pseudo online framework to better characterize the efficacy of BCI methods and the largest EEG-based BCI reproducibility study using the Mother of all BCI Benchmarks (MOABB) framework. This research seeks to promote greater reproducibility and trustworthiness in BCI studies.In conclusion, we address two critical challenges in the field of EEG-based brain-computer interfaces (BCIs): enhancing performance through advanced algorithmic development at the sensor level and improving reproducibility within the BCI community

Die funktionelle Magnetresonanztomographie (fMRT) ist ein nicht-invasives Bildgebungsverfahren, mit dem Veränderungen der neuronalen Aktivität im Gehirn mit hoher räumlicher Auflösung erfasst werden können. Mit der fMRT-Bildgebung bei neurowissenschaftlichen Experimenten wurden in den letzten beiden Jahrzehnten bedeutende Erkenntnisse für die Hirnforschung und Medizin gewonnen. Mit Hilfe dieser Methode werden neuronale Aktivitätsunterschiede bei der Durchführung einer bestimmten Aufgabe, z. B. dem Betrachten von Bildern mit emotionalen Inhalten, erfasst und die Daten unabhängig von der Messung zu einem späteren Zeitpunkt statistisch ausgewertet. Mit Hilfe des technischen Fortschritts im letzten Jahrzehnt ist es darüber hinaus möglich geworden, fMRT-Daten direkt zur Aufnahmezeit zu verarbeiten und auszuwerten. Dies wird als Echtzeit-fMRT bezeichnet, wenn die Datenverarbeitung schneller erfolgt als die Aufnahme eines Hirnvolumens (aktuell ca. zwei Sekunden). Die Ergebnisse der Echtzeitdatenverarbeitung können dann wiederum als Steuerbefehle für verschiedene Anwendungen verwendet werden. Daher wird dies auch als Hirn-Computer-Schnittstelle (Brain Computer Interface, BCI) mittels fMRT bezeichnet. Die Echtzeitverarbeitung von fMRT-Daten ermöglicht mehrere neue Anwendungen. Dazu gehören unter anderem die Qualitätskontrolle zur Laufzeit von fMRT-Experimenten, die schnelle funktionelle Lokalisierung von Hirnarealen entweder für neurowissenschaftliche Experimente oder intraoperativ, die Kontrolle des Experimentes mittels des Probandenverhaltens und insbesondere die Möglichkeit, sogenannte fMRT-Neurofeedbackexperimente durchzuführen. Bei diesen Experimenten lernen Probanden, die Aktivität von definierten Hirnarealen willentlich zu beeinflussen. Das Ziel dabei ist, Veränderungen in ihrem Verhalten zu generieren. Die Umsetzung eines BCIs mittels Echtzeit-fMRT ist eine interdisziplinäre Aufgabenstellung von MR-Physik, Informatik und Neurowissenschaften um das Verständnis des menschlichen Gehirns zu verbessern und neue Therapieansätze zu gestalten. Für diese hard- und softwaretechnisch anspruchsvolle Aufgabe gibt es einen enormen Bedarf an neuen Algorithmen zur Datenverarbeitung und Ansätzen zur verbesserten Datenakquise. In diesem Zusammenhang präsentiert die vorliegende Arbeit ein neues Softwareframework, das einerseits eine integrierte Gesamtlösung für die Echtzeit-fMRT darstellt und in seinen Teilmodulen eine abstrakte Basis für eine universelle Methodenentwicklung anbietet. Diese Arbeit beschreibt die grundlegenden abstrakten Konzepte und die Implementierung in ein neues Softwarepaket namens ’Brain Analysis in Real-Time’ (BART). Der Fokus der Arbeit liegt auf zwei Kernmodulen, die für universelle Gestaltung von sogenannten adaptiven Paradigmen und die Einbindung von Echtzeit-fMRT-Datenverarbeitungsalgorithmen konzipiert sind. Bei adaptiven Paradigmen werden zur Laufzeit eines Experiments physiologische Parameter (z. B. Herzrate) oder Verhaltensdaten (z. B. Augenbewegungen) simultan zu den fMRT-Daten erfasst und analysiert, um die Stimulation eines Probanden entsprechend zu adaptieren. Damit kann die Zuverlässigkeit der Daten, die zur Auswertung zur Verfügung stehen, optimiert werden. Die vorliegende Arbeit präsentiert das dazu notwendige abstrakte Grundkonzept des neuen Softwareframeworks und die ersten konkreten Implementierungen für die Datenverarbeitung und adaptive Paradigmen. Das Framework kann zukünftig mit neuen methodischen Ideen erweitert werden. Dazu gehören die Einbindung neuer Datenverarbeitungsalgorithmen, wie z. B. Konnektivitätsanalysen und die Adaption von Paradigmen durch weitere physiologische Parameter. Dabei ist insbesondere die Kombination mit EEG-Signalen von großem Interesse. Außerdem bietet das System eine universelle Grundlage für die zukünftige Arbeit an Neurofeedbackexperimenten. Das in dieser Arbeit entwickelte Framework bietet im Vergleich zu bisher vorgestellten Lösungsansätzen ein Ein-Computer-Setup mit einem erweiterbaren Methodenspektrum. Damit wird die Komplexität des notwendigen technischen Setups reduziert und ist nicht auf einzelne Anwendungsfälle beschränkt. Es können flexibel neue Datenverarbeitungsalgorithmen für ein fMRT-BCI eingebunden und vielgestaltige Anwendungsfälle von adaptiven Paradigmen konzipiert werden. Eine Abstraktion der Stimulation und die Kombination mit der Echtzeitauswertung ist bisher einzigartig für neurowissenschaftliche Experimente. Zusätzlich zu den theoretischen und technischen Erläuterungen werden im empirischen Teil der vorliegenden Arbeit neurowissenschaftliche Experimente, die mit dem Softwarepaket BART durchgeführt wurden, vorgestellt und deren Ergebnisse erläutert. Dabei wird die Zuverlässigkeit und Funktionsweise der Implementierung in allen Teilschritten der Datenerfassung und -verarbeitung validiert. Die Ergebnisse verifizieren die Implementierung einer parallelisierten fMRT-Analyse.Weiterhin wird eine erste konkrete Umsetzung für ein adaptives Paradigma vorgestellt, bei dem zur Laufzeit die Blickrichtung der Probanden berücksichtigt wird. Die Ergebnisse zeigen die signifikanten Verbesserungen der Reliabilität der fMRT-Ergebnisse aufgrund der optimierten Datenqualität durch die Adaption des Paradigmas. Zusammengefasst umfasst die vorliegende Arbeit eine interdisziplinäre Aufgabe, die sich aus der Verarbeitung von MR-Daten in Echtzeit, einem neuen abstraktes Softwarekonzept für Entwicklung neuer methodischer Ansätze und der Durchführung von neurowissenschaftlichen Experimenten zusammensetzt.

To date, only one study has situationally bolstered attentional resources in an effort to improve P300-BCI performance. The current study implements a 4-week Mindfulness Meditation Training Program (MMTP) as a nonmedicinal means to increase concentrative attention and to reduce lapses of attention; MMTP is expected to improve P300-BCI performance by enhancing attentional resources and reducing distractibility. A second aim is to test the efficacy of the checkerboard paradigm (CBP) against the standard row-column paradigm (RCP). Online results show that MMTP had greater accuracies than CTRL and that CBP outperformed the RCP. MMTP participants provided greater amplitude positive target responses, but these differences were not statistically significant. CBP had greater positive amplitude peaks and negative peaks than RCP. The discussion focuses on potential benefits of MMTP for P300-based BCIs, provides further support for the construct validity of mindfulness, and addresses future directions of the translational applicability of MMTP to in-home settings.

Orientador: Prof. Dr. Diogo Coutinho Soriano
Dissertação (mestrado) - Universidade Federal do ABC. Programa de Pós-Graduação em Engenharia Biomédica, 2017.
Interfaces cérebro-computador (BCIs) definem canais de comunicação capazes de mapear sinais cerebrais em sinais de controle para dispositivos externos, sem utilização dos eferentes biológicos, utilizando comumente estratégias não invasivas para tanto, tal como obtido pela eletroencefalografia de superfície. Dentre os principais paradigmas BCI têm-se os potenciais visualmente evocados em regime estacionário (SSVEP - steady state visually evoked potential), o qual se baseia no sincronismo da atividade elétrica do córtex visual com estímulos visuais externos, permitindo assim a identificação dos eletrodos e frequências estimulatórias mais eficientes para a discriminação dos estímulos escolhidos pelo usuário via modulação da sua atenção. Tal paradigma de sistema BCI tem sido utilizado como uma importante estratégia no âmbito do desenvolvimento de tecnologias assistivas, as quais visam aumentar a qualidade de vida de pacientes com severas limitações motoras e de comunicação. Neste contexto, o presente trabalho apresenta contribuições à implementação de sistemas BCI-SSVEP operando de modo off-line e em tempo de execução (on-line). Para tanto, analisa-se aqui um conjunto de estruturas de processamento de sinais que levam ao melhor desempenho na tarefa de reconhecimento de padrões considerando técnicas clássicas de estimação de espectro e análise de correlação canônica (CCA - Canonical Correlation Analysis), um método comumente referenciado por seus bons resultados. Comparativos envolvendo variantes de pré-processamento baseados na filtragem espacial e na seleção de atributos também são apresentados. Dois conjuntos de dados foram analisados em ambiente off-line e um em tempo de execução. O primeiro conjunto de dados off-line foi analisado a partir da coleta de dados em cooperação científica no contexto do projeto DesTiNe, enquanto o segundo conjunto envolveu coleta de dados off-line e em tempo de execução no próprio laboratório de Métodos Computacionais para a Bioengenharia da UFABC. Como contribuições centrais podem-se mencionar: 1) comparativo de desempenho utilizando variantes de técnicas de filtragem espacial, extração e seleção de características em ambiente off-line; 2) implementação de um setup completo experimental para realização de experimentos BCI-SSVEP com neuro-feedback visual e auditivo; 3) Disponibilização de uma base de dados BCI-SSVEP contendo aquisições de 15 sujeitos com 12 sessões de 6 segundos para cada uma das 4 frequências (10, 11, 12 e 13 Hz), totalizando 48 sessões por sujeito, i.e. um total de 720 sessões de 6 s ou 4.320 s de dados disponibilizados para a comunidade científica; 4) Comparação de 3 métodos de extração de características em âmbito off-line (FFT, Welch e CCA); 5) Comparação de 2 métodos de extração de características em âmbito on-line, FFT e CCA; 6) Análise de desempenho on-line versus off-line.
Brain-computer interfaces (BCIs) define communication channels capable of mapping brain signals in control signals to external devices, without the use of biological efferents, using commonly non-invasive strategies for both, as obtained by surface electroencephalography. Among the main BCI paradigms are the steady state visually evoked potentials (SSVEP), which is based on the synchronization of the electrical activity of the visual cortex with external visual stimuli, thus allowing the identification of the electrodes and frequencies stimulus for discriminating the stimuli chosen by the user by modulating its attention. This BCI system paradigm has been widely used in the development of assistive technologies, which aim to increase the quality of life of patients with severe motor and communication limitations. In this context, this work presents contributions to the implementation of BCI-SSVEP systems operating offline and at run-time. To do so, we analyze here a set of signal processing structures that lead to the best pattern recognition performance considering classical techniques as spectrum estimation and Canonical Correlation Analysis (CCA), a commonly cited method for its good results. Comparisons involving preprocessing variants based on spatial filtering and attribute selection are also presented. Two sets of data were analyzed in an offline environment and one at run time. The first set of off-line data was analyzed from data collection in scientific cooperation in the context of DesTiNe project, while the second set involved off-line and run time data analysis in the Laboratory of Computational Methods for Bioengineering at UFABC. As central contributions may be mentioned: 1) comparative performance using variants of techniques of spatial filtering, feature extraction and feature selection in an offline environment; 2) implementation of a complete experimental setup to perform BCI-SSVEP experiments with visual and auditory neuro-feedback; 3) Availability of a BCI-SSVEP database containing acquisitions of 15 subjects with 12 sessions of 6 seconds for each of the 4 frequencies (10, 11, 12 and 13 Hz), totaling 48 sessions per subject, ie a total of 720 sessions of 6 s or 4,320 s of data made available to the scientific community; 4) Comparison of 3 methods of feature extraction in off-line environment (FFT, Welch and CCA); 5) Comparison of 2 methods of feature extraction in online scope, FFT and CCA; 6) Analysis of performance online versus offline.

This research set out to develop a novel technique to decompose Electroencephalograph (EEG) signal into sets of constituent peaks in order to better describe the underlying nature of these signals. It began with the question; can a localised, single stimulation of sensory nervous tissue in the body be detected in the brain? Flash Visual Evoked Potential (VEP) tests were carried out on 3 participants by presenting a flash and recording the response in the occipital region of the cortex. By focussing on analysis techniques that retain a perspective across different domains � temporal (time), spectral (frequency/scale) and epoch (multiple events) � useful information was detected across multiple domains, which is not possible in single domain transform techniques. A comprehensive set of algorithms to decompose evoked potential data into sets of peaks was developed and tested using wavelet transform singularity detection methods. The set of extracted peaks then forms the basis for a subsequent clustering analysis which identifies sets of localised peaks that contribute the most towards the standard evoked response. The technique is quite novel as no closely similar work in research has been identified. New and valuable insights into the nature of an evoked potential signal have been identified. Although the number of stimuli required to calculate an Evoked Potential response has not been reduced, the amount of data contributing to this response has been effectively reduced by 75%. Therefore better examination of a small subset of the evoked potential data is possible. Furthermore, the response has been meaningfully decomposed into a small number (circa 20) of constituent peaksets that are defined in terms of the peak shape (time location, peak width and peak height) and number of peaks within the peak set. The question of why some evoked potential components appear more strongly than others is probed by this technique. Delineation between individual peak sizes and how often they occur is for the first time possible and this representation helps to provide an understanding of how particular evoked potentials components are made up. A major advantage of this techniques is the there are no pre-conditions, constraints or limitations. These techniques are highly relevant to all evoked potential modalities and other brain signal response applications � such as in brain-computer interface applications. Overall, a novel evoked potential technique has been described and tested. The results provide new insights into the nature of evoked potential peaks with potential application across various evoked potential modalities.