IEEE Consumer Electronics Magazine - March 2018 - 7

The tutorial was rated favorably
based on the IEEE Continuing Engineering Unit (CEU) evaluation scale,
although regrets were expressed for not
allocating enough time. Two CEU units
were granted to those who returned their
evaluation sheet after completion of the
challenge project. Not many international members understood the value of the
CEUs, but a few did sign up to obtain it.
The keynote delivered by Dr. Chu on
27 June served as the opening remarks
for SPCN 2017, held in the auditorium
decorated with the bust of the famous
Nobel laureate Ivan Petrovich Pavlov.
Her keynote inspired the linking of quantum physics and early neuroscience
research following the realization of the
statistic nature of neurons firing, triggering chain reactions. It also endorsed the
physical and psychological interaction in
human brain scenarios and the multidisciplinary necessity in grasping the randomness and sensitiveness shown in
brain experimentation. Digital intelligence transforming brain signals would
become the future trend due to the ubiquity of radio communications, i.e., cell
phones, with a surprisingly high penetration rate of 150% in Russia. Dr. Chu
concluded by reminding those in brain
research and practice of their social obligation, raising awareness of painful
impacts resulting from high-tech expansion, such as nuclear power plant accidents and terrorists' hacking over Internet
connectivity. Digital intelligence should
be upfront in addressing the human fear
over who owns whose brain.

notED PRoJECts
Awards were announced after Dr. Chu's
keynote speech, and first place was
earned by Team NeuroImaging. Their
winning project, Multiuser BCI, was
conducted by Nicolai Smetanin, Sacha
Kuznetsova, and Dimitrii Altukhov.
Eldar Kalmykov, chair of the board of
directors, DNK Corporation, headquartered in St. Petersburg, presented a cash
prize in addition to the first-place certificate issued by the IEEE Brain Initiative.
Two other teams and their respective
projects are also worthy of citing. All
three projects are summarized in the following sections.

MULTIUSER BCI
The winning team was Team NeuroImaging, which included Nicolai Smetanin,
Sacha Kuznetsova, and Dimitrii Altukhov, from the Centre for Cognition and
Decision Making, Higher School of
Economics (HSE), RAS, Moscow, Russia. The team described their project
briefly as follows (a short video was created to demonstrate their results [7]).
Application of the BCI technology is
not limited merely to the enhanced
human-machine interaction. The use of
BCIs can also broaden the means of
human-to-human communication and
interaction. Multiuser BCIs come into
play to offer new intriguing opportunities. They also enable various gaming
scenarios where individuals can compete
or collaborate with each other, improving
their BCI skills, and potentially optimizing some specific brain functions.
This type of game application might
be of a particular interest for motor
function recovery after brain injuries.
Electrophysiology-based rehabilitation
techniques usually comprise three components: motor-imagery-based BCI, realtime feedback, and functional electric
stimulation. Motor-imagery BCIs imply
heavy use of the sensorimotor system
coupled with visual feedback and are
often facilitated by the electrical stimulation of the corresponding muscles.
A multiuser-BCI-based game was
developed in which two people were
competing for robotic control over a
drink bottle feeder. Both players could
send to the robotic device one of the
three commands, go left, go right, or rest,
linked to imaginary movements of the
left and right hands or relaxation. Players
were sitting in front of each other with
the robotic device controlling the inclination angle of a bottle with liquid in
between. Each player was trying to pour
a drink in a glass to his or her own right
side and prevent the opponent from
doing the same. A player who was more
successful at reaching the state of concentration on his or her right hand gained
more control time over the device and, as
a result, won. On the contrary, if the player, by mistake, thought of his or her left
hand or did nothing, he or she would
help his or her opponent.

The bottle-controlling robot was
assembled at the Challenge site using
the LEGO MINDSTORMS EV3 education kit. As Team NeuroImaging spent
time well into the evening of 26 June,
two teammates' ongoing brain activity
was recorded via a Smart-BCI wireless
EEG cap with 24 electrodes.
The recorded brain signal was then
transmitted to a server, where it was
preprocessed and classified to one of
the three states to form a command
from the set described previously. The
ongoing sequence of the decoded commands for each player was streamed via
the lab streaming layer (LSL) protocol
to the robot-control server, where it was
accumulated over a window of N last
samples in the streamed sequence. Then
the robot-control server calculated the
difference in the number of occurrences
of go right and go left commands over
the window. If the resulting number was
lower than the predefined threshold, the
stop command would be sent to the
device. Otherwise, the angular speed of
the bottle was adjusted to be proportional to this difference and oriented to the
dominantly represented direction. The
game stopped when the bottle orientation reached a critical angle, and the liquid was poured into a cup situated in the
direction the winner's brain manifested.
The brick running Linux-based operating system ev3dev was used, which
allowed programming the robot's behavior in Python for a signal processing and
decoding pipeline. The control signals
followed three decoded player's states:
rest, left wrist rotation, and right wrist
rotation. To train the classifier, a training
session of brain activity was recorded
for both players prior to the game. Brain
activity for the classified states was
recorded in a sequence of three blocks,
three 20-s trials each (one trial per state),
adding up to a 180-s training EEG
record for each player. Such collected
EEG signals were used as training data
for the BCI state decoder built on top of
the open-source tool box developed in
the NeuroFeedBack-Lab at its school
center in HSE, Moscow.
For brain data processing, first, the
recorded EEG signals were split into
eight frequency bins by means of the

march 2018

IEEE Consumer Electronics Magazine

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - March 2018