Researchers:
Scott Eaton, Gary McDarby, Rob Burke, Ed Lalor +
collaborators (Tomas Ward – NUI Maynooth)
The Cerebus system is comprised of functional blocks that work together
to
implement a multi-functional brain-computer interface. The system is completely
modular and is built around object-oriented design patterns; each functional
block,
or node, encapsulates specific functionality independent of the function
of the
other nodes. There are six classes of nodes: host nodes, power nodes,
distribution
nodes, reference nodes, data nodes and mechanical nodes. These nodes slide
onto distribution rails that act first as a mechanical structure for positioning
the
nodes and second, as a housing for the power and communications bus that
links
the nodes.
The
object-orientated design is a central feature of the system. This modular
approach results in an extremely flexible platform for the brain research
and
interface development. Because the functionality of each data node is
entirely
independent of the configuration of the skeleton or the other data nodes
in the system, diverse types of sensors can be used simultaneously. For
example,
by combining EEG data nodes with DOT data nodes,it is possible to maximise
the
benefits of each modality while minimizing the overall limitations of
the system. In
this case, EEG nodes give the system high temporal resolution but limited
spatial
resolution. DOT nodes complement this by providing improved spatial resolution
though they lack the temporal resolution of the EEG nodes. Other data
nodes can
be added to provide additional application specific information (patient
location,
environmental noise, temperature, etc).
|
