Arjon Turnip

Improvement of BCI Performance Through Nonlinear Independent Component Analysis Extraction

Electroencephalogram (EEG) recordings provide an important means of brain-computer communication, but their classification accuracy and transfer rate are limited by unexpected signal variations due to artifacts and noises. In this paper, a nonlinear independent component analysis (NICA) extraction method for brain signal based EEG-P300 are proposed. The performance of the proposed method is investigated through a comparison of well known extraction methods (i.e., AAR, JADE, and SOBI algorithms). Finally, the promising results reported here reflect the considerable potential of EEG for the continuous classification of mental states.

Removal artifacts from EEG signal using independent component analysis and principal component analysis

In recording the EEG signals are often contamination signal called artifacts. There are different kinds of artifacts such as power line noise, electromyogram (EMG), electrocardiogram (ECG) and electrooculogram (EOG). This research will compare two methods for removing artifacts, i.e. ICA and PCA methods. EEG signals are recorded on three conditions, which is normal conditions, closed eyes, and blinked eyes. After processing, the dominant frequency of the EEG signal is obtained in the range of 12-14 Hz (alpha-beta) either on normal conditions, closed eyes, and blinked eyes. From processing with ICA and PCA methods found that ICA method are better than PCA in terms of the separation of the EEG signals from mixed signals.

Automatic artifacts removal of EEG signals using robust principal component analysis

Analysis of EEG activity usually raises the problem of differentiating between genuine EEG activity and that which is introduced through a variety of external influence. These artifacts may affect the outcome of the EEG recording. In this paper, wavelet denoising and band pass filter for preprocessing and a robustprincipal component analysis algorithm for extraction are proposed to remove the artifacts. The algorithm is designed to adaptively derive a relatively small number of decorrelated linear combinations of a set of random zero-mean variables while retaining as much of the information from the original variables as possible. The proposed method was tested in real EEG records acquired from eight subjects. The experimental result show that the proposed method can effectively remove the artifacts from all subjects.

EEG-SSVEP signals extraction with nonlinear adaptive filter for brain-controlled wheelchair

In this paper, an application of nonlinear adaptive filter on EEG-SSVEP extraction for brain-controlled wheelchair is proposed. In the experiment, four subjects with age about 25±1 years were tested. The experimental results in this work show that the implementation of the proposed method achieves a very significant statistical improvement in extracting peak amplitude features. After a short time of practice, most participants could learn to extract the EEG-SSVEP wave with greater than 95% accuracy.

Brain-controlled wheelchair based EEG-SSVEP signals classified by nonlinear adaptive filter

In this paper, an extraction for brain-controlled wheelchair by applying nonlinear adaptive filter on EEG-SSVEP is proposed. A four-choice signal paradigm with differents frequencies (i.e., from 6 to 9 Hz for left, right, bottom, and top, respectively) is used to stimulate the four subjects (about 25±1 years old) in the experiment. The experimental results show that the application of the extraction method achieves a very significant statistical improvement in extracting peak amplitude features.

Artifacts Removal of EEG Signals Using Nonlinear Adaptive Autoregressive

Analysis of EEG activity usually raises the problem of differentiating between genuine EEG activity and that which is introduced through a variety of external influence. These artifacts may affect the outcome of the EEG recording. In this paper, the Nonlinear Autoregressive (NAR) algorithm for artifacts removal of EEG signals in connection with the choice of the model structure (order) and computation of the system coefficients is proposed. The proposed method was tested in real EEG records acquired from eight subjects. The experimental result show that the proposed method can effectively remove the artifacts from all subjects.

An experiment of lie detection based EEG-P300 classified by SVM algorithm

ERP method is chosen to identify whether a person is lying or not. It comprises of three steps and utilizes signal P300 as marker. For the sake of simplicity, Matlab based program is constructed to take over the processes. Eleven males whose age is between 20 and 27 were subject to the experiment. The gathered data were then divided into training and test data to produce several models. They were then narrowed down using SVM method based on accuracy and computation time. Despite being relatively low in accuracy, the resulting model that is used in the program proved to be able to discern all of the subjects.

Removal of EOG artifacts: Comparison of ICA algorithm from recording EEG

Electroencephalogram (EEG) is the activity of bioelectrical signals that recorded from electrodes on the scalp. In EEG recording, the signal obtained is not entirely derived from the brain, but may be contaminated by other signals such as Electrooculogram (EOG), Electrocardiogram (ECG) and Electromiogram (EMG). EEG signals that recorded, especially by electrodes located near the eyes, will be affected by EOG. So that necessary action is needed to eliminate or reduce these EEG signals artifacts. This paper proposed a method using ICA for EOG artifact removal and compared which ICA algorithm (JADE and SOBI) is more effective and has better results in the removal of EOG artifacts in EEG recording.

P300 detection using nonlinear independent component analysis

In this paper, a nonlinear independent component analysis (NICA) extraction method for brain signal based EEG-P300 are proposed. The performance of the proposed method is investigated through a comparison of well-known extraction methods (i.e., AAR, JADE, and SOBI algorithms). Finally, the promising results reported here reflect the considerable potential of EEG for the continuous classification of mental states.

Real time classification of SSVEP brain activity with adaptive feedforward neural networks

In this study, real time classification of steady state-visual evoked potensials using the adaptive feedforward Neural Networks algorithm is proposed. The classification results is directly used to make a user able of controlling the directions (stop, forward, right, and left with stimuli frequencies of 7,5 10, 15, and 20 Hz, respectively) of a wheelchair based brain computer interface. The data was collected during a session in which eleven subjects with age about 24±2 years were tested. The proposed method showed a significant performance in the classification accuracy level and it gave an accuracy level of around 85%.