Biomedical Engineering Reference
In-Depth Information
A BCI may be an assistive technology (AT) that allows communication in people that are
paralyzed or have severe motor deficits. Most ATs usually require control by using muscles.
For this reason, people with disabilities leading to progressive motor degeneration, such
as amyotrophic lateral sclerosis (ALS), brainstem stroke, or severe cerebral palsy, lack any
communication device. In recent years, thanks to refined technologies, different systems
that enable a connection between the brain and a machine (e.g., a computer) have been
developed.
A BCI uses the electrical, magnetic, and metabolic activity generated by neurons to
give an input to an AT. This usually happens by measuring brain activity with brain
imaging techniques. Different electrophysiological recordings are possible: invasive and
noninvasive (Lebedev and Nicolelis 2006). The invasive method uses electrocorticography
(ECoG), which is characterized by intracranial recordings of electrical activity, or
direct activity of single neurons or neural assemblies. The noninvasive methods use
electroencephalography (EEG), magnetoencephalography (MEG), functional magnetic
resonance imaging (fMRI), and near-infrared spectroscopy (NIRS) to allow control over
a personal computer or peripheral device. This kind of method is mostly used to enable
paralyzed patients to develop a communication channel with the outside world (Wolpaw
et al. 2002). BCI research usually focuses on EEG because of the portability and the low cost
of this technique. To focus on patients, we will mainly present EEG-based BCI, providing
some information on BCIs that are based on other techniques.
A century has passed since the discovery of the EEG (Swartz and Goldensohn 1998),
and more than 30 years have gone by from the attempt to create an interface that is able
to establish a direct connection between the brain and machines. Over the past 20 years,
BCIs have seen a remarkable increase involving several research groups all over the world.
BCI systems consist of two separate functional blocks: a transducer, which translates
the person's brain activity into usable control signals, and the peripheral device (Mason
et al. 2005). The transducer is made of different parts (see Figure 17.1). Generally, there are
sensors that record brain activity, which usually has a very small amplitude and needs to
be amplified. To be usable, this signal needs to be cleaned from artifacts. Finally, often by
using complex mathematical algorithms, the specific feature of the brain activity that will
be used as an input can be extracted and translated. The assistive device component uses
these control signals to perform a desired activity or function.
Assistive device
Artifact
processor
Feature
extractor
Feature
translator
Amp
Device
controller
Sensors
BI transducer
FIgUre 17.1
(See color insert.) Generally, the sensors record brain activity, which needs to be amplified. Then, the signal
must be cleaned from artifacts. Finally, the feature of the brain activity that will be used as an input can be
extracted and translated by an algorithm. The assistive device component uses these control signals to perform
a desired activity or function.
Search WWH ::




Custom Search