Page 147 - Handbook of Biomechatronics

144 Domen Novak

1.4.4 EEG/fNIRS and Autonomic Nervous System Responses for
Workload Analysis
As previously mentioned, both EEG and fNIRS can be used as indicators of
mental workload. Since the mental workload estimate is obtained by
extracting several features from multiple EEG or fNIRS channels and input-
ting those features into a classification algorithm, it would be possible to
increase the classification accuracy using additional signals whose features
would provide complementary information about mental workload. One
popular type of signal are autonomic nervous system responses such as heart
rate, respiration, and peripheral skin conductance, all of which are correlated
with both physical and mental workload. Features from these signals can be
combined with features from the EEG and/or fNIRS using standard classi-
fication algorithms such as linear discriminant analysis or neural networks, as
reviewed in a survey paper by the author of this book chapter (Novak
et al., 2012).

2 BIOMECHATRONIC APPLICATIONS

Regardless of the exact sensor(s), BCI paradigm, and signal-processing
methods, the outputs of a BCI are essentially the commands that the user
wants to send to a biomechatronic device (for most BCIs) or an estimate
of the user’s mental state (for passive BCIs such as those mentioned in
“Mental Imagery” section). Currently, BCIs are primarily used in assistive
applications by people with disabilities who are unable to use other control
methods. For example, people with tetraplegia are paralyzed from the neck
down and thus cannot use devices such as keyboards, but can still control
biomechatronic devices using BCIs since this requires no movement below
the neck. However, nonassistive applications of BCIs also exist, and we pre-
sent a few examples of each application in the following sections.

2.1 Control of Powered Wheelchairs

Millions of people worldwide suffer from mobility impairments, and many
of them rely on powered wheelchairs to perform everyday activities. Such
powered wheelchairs are equipped with strong motors that allow them to
drive around quickly and climb ramps or even stairs. However, many
patients who could benefit from powered wheelchairs are not able to use
them since severe impairments (e.g., tetraplegia) prevent them from using
conventional wheelchair interfaces such as joysticks. Instead, such patients