Optimization of parameters for improving the performance of EEG-based BCI system

One of the main problems in Brain Computer Interface (BCI) systems is the non-stationary behavior of the electroencephalography (EEG) signals causing problems in real time applications. Another common problem in BCI systems is the situation where the labeled data are scarce. In this study, we take a semi-supervised learning perspective and propose solving both types of problems by updating the BCI system with labels obtained from the outputs of the classifier. To test the approach, data from motor imagery BCI system are used. Attributes extracted from EEG signals are classified with Linear Discriminant Analysis (LDA) and Support Vector Machines (SVM). With respect to the static classifiers, accuracy was improved approximately 4% using the proposed adaptation approach in the case of a training dataset. Even though the difference between the performance of static and adaptive classifiers decreases as the size of training data increases, the accuracy of our proposed adaptive classifier remains higher. The proposed approach has also improved the performance of a BCI system around 4% in the case of non-stationary signals as well.

Brain computer interface (BCI) is known as a good way to communicate between brain and computer or other device. There are many kinds of physiological signal can operate BCI systems. Motor imagery (MI) has been demonstrated to be a good way to operate a BCI system. In some recent studies about MI based BCI systems, low accuracy rate and time consuming are common problems. In this thesis, a novel motor imagery algorithm is proposed to improve the accuracy rate and computational efficiency at the same time. The architecture of many BCI system is quite complex and they involve time consuming processing. The electroencephalography (EEG) signal is the most commonly used inputs for BCI applications but EEG is often contaminated with noise. To overcome such drawbacks, in this paper we use the common spatial pattern (CSP) for feature extraction from EEG and the linear discriminant analysis (LDA) for motor imagery classification. In this study, CSP and LDA have been used to reduce the artifact and classify Mi-based EEG signal. We have used two-level cross validation scheme to determine the subject specific best time window and number of CSP features. We have compared the performance of our system with BCI competition results. This novel algorithm with high accuracy rate and efficiency can be applied to real time BCI system in real-life applications.

Motor imagery (MI)-based brain-computer interface (BCI) systems are being increasingly employed to provide alternative means of communication and control for people suffering from neuro-motor impairments, with a special effort to bring these systems out of the controlled lab environments. Hence, accurately classifying MI from brain signals, e.g., from electroencephalography (EEG), is essential to obtain reliable BCI systems. However, MI classification is still a challenging task, because the signals are characterized by poor SNR, high intra-subject and cross-subject variability. Deep learning approaches have started to emerge as valid alternatives to standard machine learning techniques, e.g., filter bank common spatial pattern (FBCSP), to extract subject-independent features and to increase the cross-subject classification performance of MI BCI systems. In this paper, we first present a review of the most recent studies using deep learning for MI classification, with particular attention to their cross-subject performance. Second, we propose DynamicNet, a Python-based tool for quick and flexible implementations of deep learning models based on convolutional neural networks. We show-case the potentiality of DynamicNet by implementing EEGNet, a well-established architecture for effective EEG classification. Finally, we compare its performance with FBCSP in a 4-class MI classification over public datasets. To explore its cross-subject classification ability, we applied three different cross-validation schemes. From our results, we demonstrate that DynamicNet-implemented EEGNet outperforms FBCSP by about 25%, with a statistically significant difference when cross-subject validation schemes are applied.

Event Abstract Back to Event Analysis and classification of EEG signals for brain computer interfaces Ugur Halici1*, E. Agi1, C. Özgen1 and I. Ulusoy1 1 Computer Vision and Intelligent Systems Research Laboratory, Middle East Technical University, Turkey Brain Computer Interfaces (BCIs) are systems developed in order to control devices by using only brain signals. Depending on the application, the device to be controlled might be a neuroprothesis, a wheel-chair or a computer. In BCI systems, different mental activities to be performed by the users are associated with different actions on the device to be controlled. The neurophysiologic brain signals produced as results of these mental activities are collected by a data acquisition device and converted to digital form. Electroencephalogram (EEG) is the most common technique employed for data acquisition in BCI systems today, due to its time-space resolution, its price and its easiness in use when compared to the other acquisition techniques. The digital brain signals collected are processed for feature extraction and then classified in order to decide on the associated action. A control signal produced according to the decision is used to change the state of the device to be controlled. Providing a feedback signal informing the user about the new state of the controlled device completes the BCI cycle. The early BCI systems were designed to work on pre-defined neurophysiologic features, where a feedback produced by a fixed set-up was provided to the system users. In these early BCI systems, the users should learn by themselves how to produce neurophysiologic signals in order to control the system. In several studies, it is reported that weeks, even months may be needed in order the user to learn how to produce these signals by adaptation of their brain. Recently, BCI systems in which machine learning algorithms employed for adapting the BCI system by using user specific signals are started to emerge. In several studies, it is shown that the adaptation of the BCI systems in this way reduces drastically the time required the user to able to produce appropriate signals and adaptation time in terms of days, even minutes are reported [1-4] The aim of this study is to analyze EEG signals collected for BCI purposes by using techniques Principal Component Analysis, Fisher Linear Discriminant Analysis, and Independent Component Analysis in order to reduce the dimension of the feature vectors obtained by sampling EEG signals in an representative manner and then classify the low dimension feature vectors obtained in this way by using Neural Networks and Support Vector Machines, which are able to learn through samples. The experimental studies are still going on for feature extraction and classification. The experimental results obtained will be presented in the final paper for comparison purposes.

A new research methodology named Brain Computer Interface (BCI) studies novel human-computer interactions; by means of BCI electronics devices, paralyzed patients are able to interact with the environment using no muscular contractions. This technique provides an external electronics support to all persons with severe motor disabilities, by acquiring in continuous mode the electroencephalogram (EEG) signals and operating some processing to control a computer or other domotics devices. Patients are so allowed to control external devices or to communicate simple messages through the computer, just concentrating their attention either on codified movements or on a letter or icon on a digital keyboard. The use of a customized and optimized spatial filtering technique embedded in the BCI system, based on the detection of the Electroencephalographic activity, improves the accuracy of BCI system itself, thanks to the explicit separation of the signal activity of interest from artefact signals. In this chapter, after an overview of the state-of-the-art research on BCI systems, the spatial filtering problem in EEG signals acquisition will be illustrated. In particular, a spatial filtering algorithm, known as ICA (Independent Component Analysis) and its application will be discussed. Finally, the design and implementation of an embedded system for EEG signals acquisition and real-time processing for BCI applications will be presented. The system is based onto a very performing and reconfigurable hardware platform. Moreover ICA algorithm has been implemented for noise reduction and artifacts removal.

Brain Computer Interfacing (BCI) systems, which are a new communicating channel between humans and the computers are growing rapidly. One such a method is based on the Steady State Visual Evoked Potentials (SSVEP), which can be recorded during visual stimulating of the subject by a twinkling light source with a fixed frequency. An important parameter to be considered is the effect of the inter-sources distance on the accuracy of such BCI systems. In particular inter-sources (LEDs) distances of 4, 14, 24, 44 and 64 cm when the sources plane is located 60 cm away from the subject's eyes (producing inter-sources visual angles of 3.8°, 13.4°, 22.6°, 40.2° and 56° respectively) were examined. In addition, four different sweep lengths of 0.5, 1, 2 and 3 seconds are considered. In addition, due to the usage of the AR models for feature extraction from the SSVEP signals, selection of the best AR model together with the best classifier among the LDA, the SVM and the Naive Bayes are studied. It is showed that the BCI system with D=44 cm, AR order of 13 and either the LDA or the SVM classifiers could produce the best results compared to the other cases.

A brain-computer interface(BCI) system is a communication channel which transforms a subject's thought process into command signals to control various devices. These systems use electroencephalographic signals or the neuronal activity of many single neurons. The presented study deals with an efficient analysis method of neuronal signals from a BCI System using an independent component analysis(ICA) algorithm. The BCI system was implemented to generate event signals coding movement information of the subject. To apply the ICA algorithm, we obtained the perievent histograms of neuronal signals recorded from prefrontal cortex(PFC) region during target-to-goal(TG) task trials in the BCI system. The neuronal signals were then smoothed over 5ms intervals by low-pass filtering. The matrix of smoothed signals was then rearranged such that each signal was represented as a column and each bin as a row. Each column was also normalized to have a unit variance. As a result, we verified that different patterns of the neuronal signals are dependent on the target position and predefined event signals.

Brain Computer Interface (BCI) systems have gained great visibility in the last years as they represent a quite innovative way of communication and a new instrument aimed at exploring brain functions. A lot of research labs are developing their own BCI system, everyone being involved in some particular aspects of them. At the "Tor Vergata" University our purpose is to develop tools for the evaluation and the optimization of the performances of BCI systems and to delineate some criteria for the analysis and implementation of different BCI systems; also we have defined file formats for BCI data in order to allow the sharing of tools among groups and to create models for the generalization and therefore the unification of the resources. All the tools and routines mentioned are part of the Body Language Framework++ 2.0.

Recently single trial classification of Event Related Potentials (ERPs) has received much attention to improve Brain Computer Interface (BCI) systems. BCI is one of the most recent fields in computer science which tries to help handicapped people; however, single trial EEG analysis is hard due to its low Signal to Noise Ratio (SNR). Many BCI systems are based on the analysis of ERPs such as P300. P300 is a positive peak that occurs approximately 300ms after a visual stimulus. But electrical potentials produced by blinks and eye movements cause serious problems in analyzing of EEG data. In this paper, a two stage algorithm is presented to detect P300 waves in single trial conditions in the presence of noise and artifact. In the first stage, the raw EEG data are denoised. For this reason, an ICA-wavelet based denoising method is proposed to automatically remove ocular artifact and it is also compared to other denoising methods. In the second stage, a detection algorithm is applied on denoised EEG data in order to discover P300 waves. In this method, a set of new features are extracted by applying ICA on each unknown incoming signal and finally they are classified using a neural network. The proposed method has been tested on more than 10 runs and an average accuracy of 71.5% in these runs is achieved in detecting P300 waves.

The performance of an EEG-based brain-computer interface (BCI) system is highly dependent on signal preprocessing. This manuscript presents a filtering method to improve the feature classification algorithms typically used in BCI. A graph Laplacian quadratic form using the Phase Locking Value (PLV) is applied to generate a new filtered signal in the preprocessing stage. The accuracy of the classification algorithms improved significantly (up to 27.18% in the BCI Competition IV dataset, and up to 42.56% with records made with an Emotiv EPOC+). In addition, the proposed filtering algorithm has similar or better results when compared with the Filter Bank Common Spatial Pattern (FBCSP), which has disadvantages in a multiclass classification. This paper shows how our PLV-based filtering between EEG channels could improve the performance of a BCI.

Mostly in natural disaster situations, for example an earthquake, we will face a situation where people get trapped and unable to be found as soon as possible due to the unpredictable terrain. Hence, robot is being seen as scouts that able to adapt and perceive, and usually snake type robot is able to move throughout rubble and small confines. Then, the idea of using modular self-reconfigurable (MSR) robots based on CEBOT in 1988 spark an interest to researchers to fit the search and rescue purpose. Besides that, the idea of controlling robots without manual control but by using “thinking” have fascinated researchers and lead to the development of interfaces between human and machines. In 1970, a Brain Computer Interface (BCI) system has been establish and the BCI system usually is based on the Electroencephalogram (EEG) which is the electrical potentials generated in which the information generated by user and independent of any muscular activities. The implementations of both technologies to each other will frontier for a new alternatives which improves self-reconfigurable modular robot in term of control strategy. As a preliminary framework, a simulation is needed to determine both systems capabilities, and it can be achieved by interfacing the robot simulator, OpenVIBE and Virtual Experimentation Platform (V-REP) with the Python programming language. This study presentation will focus on how to control the modular robot in V-REP environment by using Python language which can be used for interfacing with BCI system created in OpenVIBE software and the simulated robot configuration. Each joint handles of modular robot based on V-REP hierarchy being streamed by Python language. There are several methods for establishment of communication between OpenVIBE and other applications. The method purpose in this paper will utilizing Lab Streaming Layer (LSL) and Python script as mediator. The outcome that is expected to be obtained for this study will be the success of communication establish to control virtual modular robot based on data from OpenVIBE. The future work based on this study will involving the BCI system designed using OpenVIBE and implementation to real modular robot might be applied.

The structure and functioning of the human brain has always attracted the attention of scientists. In the past 10-20 years, one of the popular areas of study related to the human brain is the Brain Computer Interface (BCI) systems. Different brain monitoring techniques such as EEG, fNIRS, MEG can be used to observe human brain activities in BCI systems. In addition to the differences in monitoring methods in BCI systems, the focus of system operation may also change. In some BCI systems, it may be preferable to focus on Motor Imagery in another system, while the focus is on the P300 signals in EEG data. This requires a differentiation between the systems during the operational phases of the BCI system. One of the most significant points of system differentiation is the classification stage of the BCI system. In this study, a compilation of classification methods commonly used in Brain Computer Interface systems based on EEG based Motor Imagery will be carried out. This study will provide researchers with information about the methods that can be used in their own systems for classification in Brain Computer Interface systems based on EEG based Motor Imagery.

Clariant Catalysts & Lummus Technology expand CATOFIN footprint with India's first PDH plant

A brain-computer interface (BCI) is a system that enables control of devices or communication with other persons, only through cerebral activity, without using muscles. The main application for BCIs is assistive technology for disabled persons. Examples for devices that can be controlled by BCIs are artificial limbs, spelling devices, or environment control systems. BCI research has seen renewed interest in recent years, and it has been convincingly shown that communication via a BCI is in principle feasible. However, present day systems still have shortcomings that prevent their widespread application. In part, these shortcomings are caused by limitations in the functionality of the pattern recognition algorithms used for discriminating brain signals in BCIs. Moreover, BCIs are often tested exclusively with able-bodied persons instead of conducting tests with the target user group, namely disabled persons. The goal of this thesis is to extend the functionality of pattern recognition algorithms for BCI systems and to move towards systems that are helpful for disabled users. We discuss extensions of linear discriminant analysis (LDA), which is a simple but efficient method for pattern recognition. In particular, a framework from Bayesian machine learning, the so-called evidence framework, is applied to LDA. An algorithm is obtained that learns classifiers quickly, robustly, and fully automatically. An extension of this algorithm allows to automatically reduce the number of sensors needed for acquisition of brain signals. More specifically, the algorithm allows to perform electrode selection. The algorithm for electrode selection is based on a concept known as automatic relevance determination (ARD) in Bayesian machine learning. The last part of the algorithmic development in this thesis concerns methods for computing accurate estimates of class probabilities in LDA-like classifiers. These probabilities are used to build a BCI that dynamically adapts the amount of acquired data, so that a preset, approximate bound on the probability of misclassifications is not exceeded. To test the algorithms described in this thesis, a BCI specifically tailored for disabled persons is introduced. The system uses electroencephalogram (EEG) signals and is based on the P300 evoked potential. Datasets recorded from five disabled and four able-bodied subjects are used to show that the Bayesian version of LDA outperforms plain LDA in terms of classification accuracy. Also, the impact of different static electrode configurations on classification accuracy is tested. In addition, experiments with the same datasets demonstrate that the algorithm for electrode selection is computationally efficient, yields physiologically plausible results, and improves classification accuracy over static electrode configurations. The classification accuracy is further improved by dynamically adapting the amount of acquired data. Besides the datasets recorded from disabled and able-bodied subjects, benchmark datasets from BCI competitions are used to show that the algorithms discussed in this thesis are competitive with state-of-the-art electroencephalogram (EEG) classification algorithms. While the experiments in this thesis are uniquely performed with P300 datasets, the presented algorithms might also be useful for other types of BCI systems based on the EEG. This is the case because functionalities such as robust and automatic computation of classifiers, electrode selection, and estimation of class probabilities are useful in many BCI systems. Seen from a more general point of view, many applications that rely on the classification of cerebral activity could possibly benefit from the methods developed in this thesis. Among the potential applications are interrogative polygraphy ("lie detection") and clinical applications, for example coma outcome prognosis and depth of anesthesia monitoring.

Brain-computer interface (BCI) systems are proposed as a means of communication for locked-in patients. One common BCI paradigm is motor imagery in which the user controls a BCI by imagining movements of different body parts. It is known that imagining different body parts results in event-related desynchronization (ERD) in various frequency bands. Existing methods such as common spatial patterns (CSP) and its refinement filterbank common spatial patterns (FB-CSP) aim at finding features that are informative for classification of the motor imagery class. Our proposed method is a temporally adaptive common spatial patterns implementation of the commonly used filter-bank common spatial patterns method using convolutional neural networks; hence it is called TA-CSPNN. With this method we aim to: (1) make the feature extraction and classification end-to-end, (2) base it on the way CSP/FBCSP extracts relevant features, and finally, (3) reduce the number of trainable parameters compared to existing deep learning methods to improve generalizability in noisy data such as EEG. More importantly, we show that this reduction in parameters does not affect performance and in fact the trained network generalizes better for data from some participants. We show our results on two datasets, one publicly available from BCI Competition IV, dataset 2a and another in-house motor imagery dataset.