![]() Conventional methods for estimating modeling order include Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC) and Final Prediction Error (FPE). Lower model orders poorly represent the signal while higher orders increase noise. Identifying correct AR’s modeling order is an open challenge. The results indicate the important role of having an optimal mixture of expertise in the subjects’ data.Īutoregressive (AR) models are of commonly utilized feature types in Electroencephalogram (EEG) studies due to offering better resolution, smoother spectra and being applicable to short segments of data. The study is focused on finding the optimal mixture of subjects in either of the proposed frameworks in addition to investigating the impact of various electrode and features selections. In the second framework, the preprocessed EEG of multiple subjects is concate- nated into a single ”super subject”, from which PSO selects electrodes and features for use on the new subject. In the first framework, electrodes and features selected by PSO from individual subjects are com- bined into a single ”meta-mask” to be applied to the new subject. This paper investigates the feasibility of two frameworks for enhancing subject transfer through a 90%+ reduction of EEG features and electrodes using Particle Swarm Optimization (PSO). Furthermore, our results support previous but disjointed ndings on the phenomenon of BCI illiteracy.Subject transfer is a growing area of research in EEG aiming to address the lack of having enough EEG samples required for BCI by using samples originating from individuals or groups of subjects that previously performed similar tasks. All methods for the data analysis in this study are supported with fully open-source scripts that can aid in every step of BCI technology. Our EEG dataset can be utilized for a wide range of BCI-related research questions. all participants were able able to control at least one type of BCI system. Interestingly, we found no universally illiterate BCI user, i.e. ![]() they were able to pro ciently perform all three paradigms. Furthermore, we found that 27.8% (15 out of 54) of users were universally BCI literate, i.e. Compared to the ERP and SSVEP paradigms, the MI paradigm exhibited large performance variations between both, subjects and sessions. Furthermore, we looked for more general, severe cases of BCI illiteracy than have been previously reported in the literature.Īverage decoding accuracies across all subjects and sessions were 71.1% (☐.15), 96.7% (☐.05), and 95.1% (☐.09), and rates of BCI illiteracy were 53.7%, 11.1%, and 10.2% for MI, ERP, and SSVEP, respectively. We evaluated the decoding accuracies for the individual paradigms and determined performance variations across both, subjects and sessions. In addition, information about the psychological and physiological conditions of BCI users was obtained using a questionnaire, and task-unrelated parameters such as resting state, artifacts, and electromyography of both arms were also recorded. In this paper, we present a BCI dataset that includes the three major BCI paradigms with a large number of subjects over multiple sessions. Electroencephalography (EEG)-based brain-computer interface (BCI) systems are mainly divided into three major paradigms: motor-imagery (MI), event-related potential (ERP), and steady-state visually evoked potential (SSVEP).
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |