EEG dataset and OpenBMI toolbox for three BCI paradigms: an investigation into BCI illiteracy.

Min-Ho Lee,Hong-Kyung Kim,Siamac Fazli,O-Yeon Kwon,Yong-Jeong Kim,Young-Eun Lee,Seong-Whan Lee,John Williamson

doi:10.1093/gigascience/giz002

Min-Ho Lee, Hong-Kyung Kim + Show 6 more

Open Access

https://doi.org/10.1093/gigascience/giz002

Copy DOI

Journal: GigaScience	Publication Date: Jan 30, 2019
Citations: 300	License type: CC BY 4.0

Affiliation: Korea University, Nazarbayev University

Abstract

BackgroundElectroencephalography (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). Here, we present a BCI dataset that includes the three major BCI paradigms with a large number of subjects over multiple 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. We evaluated the decoding accuracies for the individual paradigms and determined performance variations across both subjects and sessions. Furthermore, we looked for more general, severe cases of BCI illiteracy than have been previously reported in the literature.ResultsAverage decoding accuracies across all subjects and sessions were 71.1% (± 0.15), 96.7% (± 0.05), and 95.1% (± 0.09), and rates of BCI illiteracy were 53.7%, 11.1%, and 10.2% for MI, ERP, and SSVEP, respectively. Compared to the ERP and SSVEP paradigms, the MI paradigm exhibited large performance variations between both subjects and sessions. Furthermore, we found that 27.8% (15 out of 54) of users were universally BCI literate, i.e., they were able to proficiently perform all three paradigms. Interestingly, we found no universally illiterate BCI user, i.e., all participants were able to control at least one type of BCI system.ConclusionsOur EEG dataset can be utilized for a wide range of BCI-related research questions. 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. Furthermore, our results support previous but disjointed findings on the phenomenon of BCI illiteracy.

Highlights

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)
We provide the BCI dataset with a laboratory developed toolbox to visualize EEG data in time-frequency domains and to validate baseline performance on the three paradigms by commonly used machine learning techniques such as common spatial pattern (CSP) [23], common spatio-spectral pattern (CSSP) [24], filter bank common spatial pattern (FBCSP) [25], Bayesian spatio-spectral filter optimization (BSSFO) [26], and canonical correlation analysis (CCA) [27]
The average rates of BCI illiteracy over the sessions were 53.7%, 11.1%, and 10.2% in the MI, ERP, and SSVEP data, respectively. These results indicate that exogenous BCI paradigms [38] (i.e., ERP and SSVEP), where external visual stimuli evoke brain responses, show a relatively small ratio of BCI illiteracy compared to the endogenous BCI paradigm [38] (i.e., MI) where a user induces the brain signals with a predefined mental task

Summary

Introduction

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). We present a BCI dataset that includes the three major BCI paradigms with a large number of subjects over multiple sessions. We evaluated the decoding accuracies for the individual paradigms and determined performance variations across both subjects and sessions. Results: Average decoding accuracies across all subjects and sessions were 71.1% (± 0.15), 96.7% (± 0.05), and 95.1% (± 0.09), and rates of BCI illiteracy were 53.7%, 11.1%, and 10.2% for MI, ERP, and SSVEP, respectively. Compared to the ERP and SSVEP paradigms, the MI paradigm exhibited large performance variations between both subjects and sessions. We found that 27.8% (15 out of 54) of users were universally BCI literate, i.e., they were able to proficiently perform all three paradigms. Our results support previous but disjointed findings on the phenomenon of BCI illiteracy

Objectives

Methods

Findings

Conclusion