Page 135 - Handbook of Biomechatronics

132 Domen Novak

1.1.1 EEG Paradigms
Before focusing on the technical aspects of EEG measurements, let us first
look at the waveforms of interest in the EEG signal as well as ways of eliciting
them. The most important waveforms for biomechatronics are steady-state
visually evoked potentials (SSVEPs), the P300, and motor/mental imagery,
all of which are used to actively send commands through a BCI (Novak and
Riener, 2015). However, BCIs can also measure a user’s mental workload or
error-related brain potentials without the user’s active participation or even
awareness, as we shall see in the following sections.

Steady-State Visually Evoked Potentials
SSVEPs are the brain’s natural responses to visual stimulation at different
frequencies (Nicolas-Alonso and Gomez-Gil, 2012). In brief, if a person
looks at a light that is flashing with a particular frequency, their visual cor-
tex responds with EEG activity at the same frequency. This principle is
used in BCIs as a gaze-tracking method: multiple symbols are shown to
the user on a screen, with each symbol flashing at a different frequency.
By measuring the SSVEP frequency using electrodes close to the visual
cortex, the machine can identify which symbol the user is looking at.
Depending on the number and complexity of possible commands, this
can be done either in a single stage (the final command is directly selected
from all possible ones) or in multiple stages (a subset of commands is first
selected from all possible ones, and the final specific command is then
selected from the subset).
SSVEPs are commonly used in biomechatronics to send commands to a
device. The user is presented with multiple commands on a screen (e.g.,
move robot forward, stop) and selects one by looking at it. The user can also
choose not to send a command by simply not focusing on the screen. The
approach is noninvasive and easy to use with little or no training, and the
number of possible commands can be quite high—the main limitations are
keeping the symbols on the screen far enough apart so that the user is not
looking at two flashing lights at once as well as keeping the different symbols
flashing at sufficiently different frequencies that they can be separated in the
EEG. The main disadvantage of the SSVEP approach is that a screen must be
added to the device, which may not be optimal for all situations (e.g., por-
table devices). Furthermore, it is prone to false positives since users still see
the screen at the edge of their vision even if they do not wish to control the
device (Ortner et al., 2011).