Cookies on this website
We use cookies to ensure that we give you the best experience on our website. If you click 'Continue' we'll assume that you are happy to receive all cookies and you won't see this message again. Click 'Find out more' for information on how to change your cookie settings.
  • The impact of PSO based dimension reduction on EEG classification

    2 July 2018

    The high dimensional nature of EEG data due to large electrode numbers and long task periods is one of the main challenges of studying EEG. Evolutionary alternatives to conventional dimension reduction methods exhibit the advantage of not requiring the entire recording sessions for operation. Particle Swarm Optimization (PSO) is an Evolutionary method that achieves performance through evaluation of several generations of possible solutions. This study investigates the feasibility of a 2 layer PSO structure for synchronous reduction of both electrode and task period dimensions using 4 motor imagery EEG data. The results indicate the potential of the proposed PSO paradigm for dimension reduction with insignificant losses in classification and the practical uses in subject transfer applications. © 2012 Springer-Verlag.

  • Multiplying the mileage of your dataset with subwindowing

    2 July 2018

    This study is focused on improving the classification performance of EEG data through the use of some data restructuring methods. In this study, the impact of having more training instances/samples vs. using shorter window sizes is investigated. The BCI2003 IVa dataset is used to examine the results. The results not surprisingly indicate that, up to a certain point, having higher numbers of training instances significantly improves the classification performance while the use of shorter window sizes tends to worsen performance in a way that usually cannot fully be compensated for by the additional instances, but tends to provide useful gain in overall performance for small divisors into two or three subepochs. We have moreover determined that use of an incomplete set of overlapping windows can have little effect, and is inapplicable for the smallest divisors, but that use of overlapping subepochs from three specific non-overlapping areas (start, middle and end) of a superepoch tends to contribute significant additional information. Examination of a division into five equal non-overlapping areas indicates that for some subjects the first or last fifth contributes significantly less information than the middle three fifths. © 2011 Springer-Verlag.

  • Evaluation of a minimum-norm based beamforming technique, sLORETA, for reducing tonic muscle contamination of EEG at sensor level.

    2 July 2018

    BACKGROUND: Cranial and cervical muscle activity (electromyogram, EMG) contaminates the surface electroencephalogram (EEG) from frequencies below 20 through to frequencies above 100Hz. It is not possible to have a reliable measure of cognitive tasks expressed in EEG at gamma-band frequencies until the muscle contamination is removed. NEW METHOD: In the present work, we introduce a new approach of using a minimum-norm based beamforming technique (sLORETA) to reduce tonic muscle contamination at sensor level. Using a generic volume conduction model of the head, which includes three layers (brain, skull, and scalp), and sLORETA, we estimated time-series of sources distributed within the brain and scalp. The sources within the scalp were considered to be muscle and discarded in forward modelling. RESULT: (1) The method reduced EMG contamination, more strongly at peripheral channels; (2) task-induced cortical activity was retained or revealed after removing putative muscle activity. COMPARISON WITH EXISTING METHODS: This approach can decrease tonic muscle contamination in scalp measurements without relying on time-consuming processing of expensive MRI data. In addition, it is competitive to ICA in muscle reduction and can be reliably applied on any length of recorded data that captures the dynamics of the signals of interest. CONCLUSION: This study suggests that sLORETA can be used as a method to quantitate cranial muscle activity and reduce its contamination at sensor level.

  • Investigating a gaze-tracking brain computer interface concept using steady state visually evoked potentials

    2 July 2018

    This project investigated the possibility of a user's gaze being tracked within the area of a computer monitor bounded by multiple light sources, each stimulating an SSVEP. If realised, such a system would allow a spatial arrangement of interactive elements around a screen, rather than the discrete list of commands accessible through existing SSVEP based BCIs. Research-level EEG equipment would make the proposed BCI prohibitively expensive for home users. Thus, investigation was made into the utility of inexpensive consumer-grade EEG equipment, as is available for computer-gaming. SSVEPs were elicited using initially a traditional strobe light source and then a set of individual LEDs, as necessary for the simultaneous stimulation of multiple SSVEPs. Clear responses were recorded using the research EEG system for both the strobe and LED sources; however the consumer system lacked sufficient sensitivity to reliably detect the SSVEPs. Tests with two stimulating LEDs showed that two SSVEPs of differing frequencies be resolved simultaneously, and that the amplitude of the response decreases as the user's gaze is directed further from the stimulating light source. Further work will aim to derive the user's gaze location within an area bounded by multiple stimulating LEDs, using the relative amplitudes of the elicited SSVEPs. © 2012 IEEE.

  • Reducing training requirements through evolutionary based dimension reduction and subject transfer

    2 July 2018

    © 2016 Elsevier B.V. Training Brain Computer Interface (BCI) systems to understand the intention of a subject through Electroencephalogram (EEG) data currently requires multiple training sessions with a subject in order to develop the necessary expertise to distinguish signals for different tasks. Conventionally the task of training the subject is done by introducing a training and calibration stage during which some feedback is presented to the subject. This training session can take several hours which is not appropriate for on-line EEG-based BCI systems. An alternative approach is to use previous recording sessions of the same person or some other subjects that performed the same tasks (subject transfer) for training the classifiers. The main aim of this study is to generate a methodology that allows the use of data from other subjects while reducing the dimensions of the data. The study investigates several possibilities for reducing the necessary training and calibration period in subjects and the classifiers and addresses the impact of i) evolutionary subject transfer and ii) adapting previously trained methods (retraining) using other subjects data. Our results suggest reduction to 40% of target subject data is sufficient for training the classifier. Our results also indicate the superiority of the approaches that incorporated evolutionary subject transfer and highlights the feasibility of adapting a system trained on other subjects.

  • Adapting subject-independent task-specific EEG feature masks using PSO

    2 July 2018

    Dimension reduction is an important step toward asynchronous EEG based BCI systems, with EA based Feature/ Electrode Reduction (FR/ER) methods showing significant potential for this purpose. A PSO based approach can reduce 99% of the EEG data in this manner while demonstrating generalizability through the use of 3 new subsets of features/electrodes that are selected based on the best performing subset on the validation set, the best performing subset on the testing set, and the most commonly used features/electrodes in the swarm. This study is focused on applying the subsets generated from 4 subjects on a 5th one. Two schemes for this are implemented based on i) extracting separate subsets of feature/electrodes for each subject (out of 4 subjects) and combining the final products together for use with the 5th subject, and ii) concatenating the preprocessed EEG data of 4 subjects together and extracting the desired subset with PSO for use with the 5th subject. The results indicate the feasibility of generating subsets of feature/electrode indexes that are task specific and can be used on new subjects. © 2012 IEEE.

  • Evolutionary feature selection and electrode reduction for EEG classification

    2 July 2018

    EEG signals usually have a high dimensionality which makes it difficult for classifiers to learn the difference of various classes in the underlying pattern in the signal. This paper investigates several evolutionary algorithms used to reduce the dimensionality of the data. The study presents electrode and feature reduction methods based on Genetic Algorithms (GA) and Particle Swarm Optimization (PSO). Evolution-based methods are used to generate a set of indexes presenting either electrode seats or feature points that maximizes the output of a weak classifier. The results are interpreted based on the dimensionality reduction achieved, the significance of the lost accuracy, and the possibility of improving the accuracy by passing the chosen electrode/feature sets to alternative classifiers. © 2012 IEEE.

  • Dimension reduction in EEG data using particle swarm optimization

    2 July 2018

    EEG data contains high-dimensional data that requires considerable computational power for distinguishing different classes. Dimension reduction is commonly used to reduces the necessary training time of the classifiers with some degree of accuracy lost. The dimension reduction is usually performed on either feature or electrode space. In this study, a new dimension reduction method that reduce the number of electrodes and features using variations of Particle Swarm Optimization (PSO) is used. The variation is in terms of parameter adjustment and adding a mutation operator to the PSO. The results are assessed based on the dimension reduction percentage, the potential of selected electrodes and the degree of performance lost. An Extreme Learning Machine (ELM) is used as the primary classifier to evaluate the sets of electrodes and features selected by PSO. Two alternative classifiers such as Polynomial SVM and Perceptron are used for further evaluation of the reduced dimension data. The results indicate the potential of variations of PSO for reducing up to 99% of the data with minimal performance lost. © 2012 IEEE.

  • The use of evolutionary algorithm-based methods in EEG based BCI systems

    2 July 2018

    Electroencephalogram (EEG) based Brain Computer Interface (BCI) is a system that uses human brain-waves recorded from the scalp as a means for providing a new communication channel by which people with limited physical communication capability can effect control over devices such as moving a mouse and typing characters. Evolutionary approaches have the potential to improve the performance of such system through providing a better sub-set of electrodes or features, reducing the required training time of the classifiers, reducing the noise to signal ratio, and so on. This chapter provides a survey on some of the commonly used EA methods in EEG study. © 2013, IGI Global.

  • Cross subject mental work load classification from electroencephalographic signals with automatic artifact rejection and muscle pruning

    2 July 2018

    � Springer International Publishing AG 2016. Purpose of this study was to understand the effect of automatic muscle pruning of electroencephalograph on cognitive work load prediction. Pruning was achieved using an automatic Independent Component Analysis (ICA) based component classification. Initially, raw data from EEG recording was used for prediction, this result was then compared with mental work load prediction results from muscle-pruned EEG data. This study used Support Vector Machine (SVM) with Linear Kernel for cognitive work load prediction from EEG data. Initial part of the study was to learn a classification model from the whole data, whereas the second part was to learn the model from a set of subjects and predict the mental work load for an unseen subject by the model. The experimental results show that an accuracy of nearly 100% is possible with ICA and automatic pruning based pre-processing. Cross subject prediction significantly improved from a mean accuracy of 54% to 69% for an unseen subject with the pre-processing.