Dry-contact EEG Electrodes and Instant Donning Multichannel Headsets for Brain-Computer Interface

Abstract

Introduction: There are various applications that use electroencephalogram (EEG) signals, such as brain-computer interface (BCI), clinical diagnosis, polysomnography, and authentication. These applications require multichannel EEG measurement. However, as the process for the preparation of the multichannel EEG measurement is time-consuming and laborious, its implementation has been limited to hospital and laboratory environments. Although there have been diverse approaches introduced to overcome these limitations, it is not yet feasible to measure EEG signals in daily life. In this dissertation, new designs of dry-contact EEG electrodes and multichannel EEG headsets are introduced that overcome the limitations of existing dry-contact EEG electrodes, simplify the procedure, and reduce the time for the preparation of multichannel EEG measurement without assistance.

Methods: The proposed EEG electrodes have a reverse-curve-arch shape and were fabricated with sterling silver using a 3D printer. This unique structure was designed to reach the surface of the scalp by passing through the layer of hair, and to maximize the contact area between the skin and electrode. Therefore, its design could reduce discomfort and pain when being worn. In addition, instant-donning multichannel EEG headsets with electrodes were introduced. These headsets were designed to let the electrodes reach the surface of the scalp while being worn. It takes approximately ten seconds to put the headsets on, including the ground and reference electrodes, without any preparation or assistance. Several experiments were conducted to validate the proposed electrodes and headsets, and they were applied to the brain-computer interface system.

Results: The skin-electrode impedance of the proposed electrode was relatively less than existing spiky electrode and the correlation between signals using Ag/AgCl and proposed electrode are higher than the spiky electrode. As instant-donning headsets, it was proven that reliable EEG signals up to eight channels were acquired just after putting it on without any assistant. SSVEP-based BCI system with the proposed system showed the average accuracy of 95.70% and ITR of 20.34b/m. Additionally, ASSR-based BCI system with the system showed the average accuracy of 76%.

Conclusions: The electrodes and headsets introduced in this study overcome the drawbacks of the existing dry-contact electrodes and multichannel headsets. It was proven that reliable EEG signals were obtained using the proposed system. This study shows the feasibility of the instant-donning headsets for various EEG-based applications, including BCI. The proposed system makes it possible to measure EEG signals in daily life without assistance, and contributes to the development of various EEG applications.