The “developer edition” of the Emotiv Epoc headset — if you think
you can handle it. Photo: drbakker/Flickr
Don’t you dare even think about your banking account password
when you slap on those fancy new brainwave headsets.
Or at least that
seems to be the lesson of a new study which found that sensitive personal
information, such as PIN numbers and credit card data, can be gleaned from the
brainwave data of users wearing popular consumer-grade EEG
headsets.
A team of security
researchers from Oxford, UC Berkeley, and the University of Geneva say that they
were able to deduce digits of PIN numbers, birth months, areas of residence and
other personal information by presenting 30 headset-wearing subjects with images
of ATM machines, debit cards, maps, people, and random numbers in a series of
experiments. The paper, titled “On
the Feasibility of Side-Channel Attacks with Brain Computer Interfaces,”
represents the first major attempt to uncover potential security risks in the
use of the headsets.
“The correct answer was found by the first guess in 20% of the
cases for the experiment with the PIN, the debit cards, people, and the ATM
machine,” write the researchers. “The location was exactly guessed for 30% of
users, month of birth for almost 60% and the bank based on the ATM machines for
almost 30%.”
To detect the first digit of the PIN, researchers presented the
subjects with numbers from 0 to 9, flashing on the screen in random order, one
by one. Each number was repeated 16 times, over a total duration of 90 seconds.
The subjects’ brainwaves were monitored for telltale peaks that would rat them
out.
The EEG headsets,
made by companies such as Emotiv Systems and NeuroSky, have become increasingly
popular for gaming and other applications. For the study, the researchers used
the Emotiv
Epoc Neuroheadset, which retails for $299.
The researchers —
Ivan Martinovic of Oxford University; Doug Davies, Mario Frank, Daniele Perito,
and Dawn Song of UC Berkeley; and Tomas Ros of the University of Geneva —
analyzed P300
peaks, an important component of event-related
potentials — electrical potentials that happen after the user is presented
with a stimulus.
The P300 “occurs approximately 300 milliseconds after an event
happens,” said Frank, a postdoctoral researcher at Berkeley, in a phone
interview with Wired. “The potential arises if you already prime your thoughts
toward a particular event…. An attacker could try to prime the thoughts of the
victim towards a particular secret that a victim has in mind. For instance, if
you know the face of some person, you might be able to observe a brainwave
pattern that is evidence of the user thinking about the face.”
“Brain Spyware”
Emotiv and NeuroSky both have “app stores,” where users of the
devices can download third-party applications. The applications use a common API
for access to the EEG device.
“In the case of the EEG devices, this API provides
unrestricted access to the raw EEG signal,” write the researchers. “Furthermore,
such applications have complete control over the stimuli that can be presented
to the users.”
The researchers envision a scenario in which a potential
malicious attacker could write “brain spyware” to harvest private information
from the user, which could be legitimately downloaded as an app.
“We simulated a scenario where someone writes a malicious app,
the user downloads it and trusts the app, and actively supports all the
calibration steps of the device to make the software work,” said Frank. In these
seemingly innocuous calibration steps, which are standard for most games and
other applications using the headsets, there could be the potential to harvest
personal information.
“We realized that
these devices are becoming increasingly popular — maybe in five, 10 years, it’s
very likely that many households will have one,” Frank said. “At the same time,
you can use all kinds of third-party apps for these devices. In this setting, as
security researchers, we identified that there is a potential to make some bad
stuff, to turn this technology against the user.” He said, however, that there
was no immediate threat in using the devices. But the experiments
devised by the researchers point to the devices’ darker potential.
“The simplicity of our experiments suggests the possibility of
more sophisticated attacks,” write the researchers, warning that “with the
ever-increasing quality of devices, success rate of attacks will likely
improve.”