Biomedical Engineering Reference
In-Depth Information
or between neural networks affected by disease. In view of the amazing plasticity of the brain, these
man-made devices may, with appropriate sensory feedback, be assimilated and become part of the
user's cognitive space, at par with the rest of the body. Because of their man-made qualities, neural
interfaces carry special qualities. Potentially, neural interfaces can scale human's natural reaction
time, force, and abilities through engineered devices that are much faster and more powerful than
biological tissue. In particular, they may enable a higher bandwidth between the human brain and
the digital computer, which has conditioned the present research direction and metaphors for the
use of computers in our society.
1.1 TyPES oF BRaIN-MaChINE INTERFaCES
In the last decade, neural prosthetic technologies have commonly been grouped under the title of
brain-computer interfaces (BCIs) or brain-machine interfaces (BMIs), and the names are often
used interchangeably.
3
BMIs can be divided into three basic types, depending upon the applica-
tion: sensory, cognitive, or motor. Sensory BMIs activate sensory systems with artificially generated
signals that translate physical quantities. The most common type of sensory BMI, with more than
50,000 implanted devices [
6
], are cochlear implants that use miniature microphones and signal pro-
cessing to translate sound wave pressure (physical quantities) in the ear into neuronal firings (neural
representation) applied to the auditory nerve's tonotopic organization, allowing deaf people to lis-
ten. The same basic concept is being developed for retinal prosthesis, which can deliver to the visual
system the appropriate stimulation that indicates the morphology of objects [
7
]. Cognitive BMIs
attempt to reestablish proper neural interactions among neural systems that have damaged internal
functioning within the network or do not have the ability to communicate with other networks.
Cognitive BMIs have been applied for the treatment of the limbic system that may have long- and
short-term memory deficits as in Alzheimer's disease [
8
]. Motor BMIs, on the other hand, seek to
translate brain activity from the central or peripheral nervous system into useful commands to ex-
ternal devices. The idea is to bypass the damaged tissue (as in a spinal cord transection) and deliver
the motor control commands either directly control a computer/prosthetic limb or deliver natural
motor control of muscles via functional electrical stimulation [
9
]. Reestablishing the control of lost
movement with the peripheral nervous system shares many of the neurotechnology and challenges
of BMIs for prosthetic control.
Motor, cognitive, and sensory BMIs for direct neural control have far reaching impact in re-
habilitation. If the whole brain or some of its parts are spared from disease or injury and the path of
3
Historically, BCI has been associated with noninvasive approaches of interfacing with the nervous system as with
the electroencephalogram (EEG) (described later). Alternatively, BMI commonly refers to invasive techniques.
