Conference PaperPDF Available

Towards Mobile and Wearable Brain-Computer Interfaces

Authors:

Abstract

Brain-Computer Interfaces (BCIs) have not been adopted as a control paradigm for mainstream use because most BCI systems are cumbersome, difficult to set up, and do not generally perform well enough in mobile settings to replace existing input modalities. However, BCIs may have promise as part of multi-modal systems that augment interactions when the user’s hands are not free and/or voice commands are not possible, often a requirement in highly mobile application domains. With recent advances in electrode capabilities, and improvements in the processing power of mobile devices and head-worn displays, it is now possible to acquire, send and process EEG signals in real-time on mobile devices. These improvements make it possible to build a wearable mobile BCI, which could provide alternate interaction methods for mainstream users as well as the disabled population. This abstract describes two pilot studies in our ongoing work designing and evaluating wearable-mobile BCI components.
Towards Mobile and Wearable
Brain-Computer Interfaces
1N. Arora, 1I. Walker, 1L. Freil, 2J. Thompson, 1T. Starner, 1M. Jackson
1Georgia Institute of Technology, Atlanta, Georgia, USA, 2United Sciences, Atlanta, Georgia, USA
*85 5th Street, Atlanta, Georgia 30308, USA. E-mail: melody@cc.gatech.edu
Introduction: Brain-Computer Interfaces (BCIs) have not been adopted as a control paradigm for mainstream use
because most BCI systems are cumbersome, difficult to set up, and do not generally perform well enough in
mobile settings to replace existing input modalities. However, BCIs may have promise as part of multi-modal
systems that augment interactions when the user’s hands are not free and/or voice commands are not possible,
often a requirement in highly mobile application domains. With recent advances in electrode capabilities, and
improvements in the processing power of mobile devices and head-worn displays, it is now possible to acquire,
send and process EEG signals in real-time on mobile devices. These improvements make it possible to build a
wearable mobile BCI, which could provide alternate interaction methods for mainstream users as well as the
disabled population. This abstract describes two pilot studies in our ongoing work designing and evaluating
wearable-mobile BCI components.
Material, Methods and Results: In our first study, our aim was to
design a BCI to detect SSVEP with all wearable components. Google
Glass [2], was used to present two flashing visual stimuli to the
participant, at 13 Hz and 17 Hz frequency simultaneously. Our EEG
amplifier was an OpenBCI board that we clipped to the participant’s
belt using a custom 3D printed clip. We used three electrodes: occipital
(Oz) as signal, mastoid for ground, and the earlobe for reference, to
detect the SSVEP signal. We recorded the EEG data for offline
analysis. Over 10 sessions, using the apparatus illustrated in Figure 1,
we could detect to which of the two stimuli our participant was
attending to with 76%-84% accuracy for 13 Hz and 67%-72% accuracy
for 17 Hz, for amplitude spectra from PSD as feature for 1 second long
sliding window SSVEP using 10 cross-fold RF classifier trained on
each stimuli individually. We extended the experiment for walk-stop-
watch stimuli scenario and found the accuracy to be 93% for single
stimuli 1 second long sliding window SSVEP.
The aim of our second study was to determine if we could replace the scalp electrodes with easily made
customized in-ear electrodes adapted from the ear-electrode design discussed by Looney [1]. We used an eFit
scanner to create a model of the participant’s left ear. We then 3D printed an earpiece, and placed 3 pre-gelled
Ag/AgCl ground plate electrodes covered with silver foil so they would contact the walls of the ear canal in the
outer ear. Resulting in-ear electrode and Oz for comparison was attached to the wearable OpenBCI system and a
flashing 13Hz LED located 6 cm away from the user. As demonstrated in Fig 2, the peak SSVEP amplitude for
the occipital region is higher than ear canal, but SNR increased as well thus resulting in comparable accuracy of
detection of 80-90% from ear and scalp using a wearable BCI.
Figure 2. (above) 13 Hz SSVEP responses from LED with occipital and ear electrodes. (right) Custom made earpiece with labeled
electrode placements
Discussion and Significance: The first prototype demonstrated that SSVEP signals could be collected from a
fully mobile and wearable BCI system. Because the display was worn on the face and the bioamplifier was
small enough to clip to a belt while still being capable of effectively detecting SSVEP, we conclude that it is
now feasible to make a fully wearable BCI system using commercial components.
The second experiment demonstrated that we could substitute in-ear electrodes, making the BCI system
smaller and less obtrusive. We have developed a quick, easy, inexpensive way to create custom ear-electrodes,
which will enable our ongoing study to test a much wider range of users. Though there is more work to be done
before wearable BCIs can be used in everyday life as simple control systems, these studies have shown the
feasibility of the mobile and wearable approach.
References
[1] Looney, D., Kidmose, P., & Mandic, D. P. (2014). Ear-EEG: user-centered and wearable BCI. In Brain-Computer Interface Research
(pp. 41-50). Springer Berlin Heidelberg.
[2] Starner, T. Project glass: An extension of the self. Pervasive Computing, IEEE, 12(2), 14-16. 2013
Figure1. Google glass with SSVEP stimuli
and OpenBCI Board
ResearchGate has not been able to resolve any citations for this publication.
Chapter
We present a radically new solution for EEG-based brain computer interface (BCI) where electrodes are embedded on a customized earpiece, as typically used in hearing aids (Ear-EEG). This provides a noninvasive, minimally intrusive and user-friendly EEG platform suitable for long-term use (days) in natural environments. The operation of Ear-EEG is illustrated for alpha-attenuation and responses to auditory stimuli, and its potential in BCI is evaluated on an SSVEP study. We show that Ear-EEG bitrate performances are comparable with those of on-scalp electrodes, thus promising a quantum step forward for wearable BCI.
Article
The article discusses how the wearable computers such as Google Glass and head-up displays (HUD) have revolutionized the way people interact. People can Google search on Glass in a wide variety of contexts, both personal and professional. Google's search displays excerpts of the most related content for each of its hits, which means that the result of the search itself often provides people with the information they need without even clicking on a link. When the time between intention and action becomes small enough, the interface becomes an extension of the self. When the time between intention and action becomes small enough, the interface becomes an extension of the self. An example is a man riding a bicycle. If a car cuts him off, the rider does not think, 'If I squeeze this lever on the handlebar, it will pivot about its axle, pulling a cable. Using a sheaf, that cable will direct the force under the bicycle and pull on a pair of calipers that will then squeeze against the sides of his back tire.