The Cybathlon:
An update on
machine learning
Written by Vinay Jayaram
ETHZ Cybathalon 2016
On October 8th, 2016 about a dozen people sat in an auditorium in Zürich and raced each other purely
with their thoughts. Hooked up to various computers and sensors, they used their brain activity to
control the avatars in a racing game, bypassing the entire physical body to talk to a computer. It
wasn’t the first time humans have communicated with computers solely through their thoughts, but
it may have been the most high profile. Thousands of onlookers watched the avatars displayed on
the jumbotron as the pilots gave new meanings to the phrase “with minds racing.”
Surprisingly enough, brain-computer
interfaces (BCIs) have existed since
computers were dumber than Ap-
ple Watches. Since the early 1990s,
people have known that the electrical
patterns of activity in the brain change
when people do specific tasks (imagin-
ing moving a part of the body is a good
example), and these changes aren’t
actually that hard to measure. Just
put some electrodes on the scalp and
that’s often enough. The natural ques-
tion then is, why are you only hearing
about this now? Well, they were just
not very consistent (or convenient). It’s
true that people could control cursors
with their minds about twenty years
ago; it was just an incredibly tedious
process. Imagine if every time you
wanted to use your mouse you had
to teach your mouse to understand
your movements for fifteen minutes,
say once every two hours. Then, if you
want to make it a bit more realistic,
imagine it can only be used once every
four seconds, misses every fifth com-
mand, and it has all the ergonomics of
an ox-drawn plough. But never fear! It
only took you six months to train your
brain to be able to use it at all.
These days things are far better —
good enough that some people can
actually type pretty quickly using a BCI,
though still nowhere near as fast as
with one’s fingers. It can only take ten
minutes to get the tech used to you,
but the ugly and often uncomfortable
8 | NEUROMAG |May 2017
hardware remains an issue. Worse,
they are really only effective in calm,
undistracting environments. All in all
much better than before, but still no-
where near something a healthy hu-
man would actively choose to use.
Even though healthy people may not
opt to trade their mice in for BCIs, there
are definitely groups that might, such
as those people no longer capable of
using a mouse. It is for these groups
that BCIs are geared. This competition
for brain-computer interfaces took
place at the Cybathlon, the world’s first
international competition for disabled
athletes with assistive technology.
In essence, it’s the Olympics for the
people who are unable to participate
in the current Special Olympics — for
now it consists only of short races, but
perhaps true to ancient form it will
one day encompass everything from
wrestling to marathons. In this first
iteration it had short races designed
to mimic everyday tasks that are dif-
ficult for people with prostheses of all
sorts, such as using the same artifi-
cial hand to both screw in a lightbulb
and grab a string, or navigating steps
with a wheelchair. In the case of the
BCI race, it was to highlight options for
individuals who are incapable of con-
trolling their muscles due to injury. Un-
like the Special Olympics, in which the
assistive technology is a tool to enable
humans to show off how well they
are able to perform, in the Cybathlon
the tech itself took center stage along
with the athletes. Which, if you think
about the Olympics, is a bold step in
and of itself.
Imagine a runner in the Special Olym-
pics. She requires prostheses to run
because both of her legs were ampu-
tated below the knee. For years before
the Olympics she trained herself to run
with these pieces of metal and plastic
strapped to her shins, and when the
day comes she runs better than she
ever has in her life. She wins.
Now imagine someone at the Cybath-
lon training for the race with her pow-
ered arm prosthesis. Every day she
practices using it to do the tasks of
the race, like picking up various objects
and opening doors and holding bags,
and she gets more and more used to
it. Once every month, the technicians
come in and upgrade it further so it
responds more quickly. On the day of
the race, she’s there on the floor of the
stage, nervous and waiting for the bell
to commence. But once it does, some-
thing just feels not quite right, and
she’s not sure why. She does pretty
well, but she loses.
The magic of the Olympics is that we
get to pretend that winners and los-
ers are both reflections of the human
condition. How you do reflects you,
and pretty much only you — your luck,
your training, and even your biology.