Information Technology Reference
In-Depth Information
analyses will likely include more comprehensive coverage of subcortical regions
and pathways and may aim at achieving higher spatial resolution to capture
smaller fiber bundles and anatomical subdivisions. Another advance would be a
definition of ROIs (network nodes) according to functional criteria, for example
based on the detection of boundaries in functional connectivity patterns. Another
particularly promising avenue would be the combination of structural and
functional imaging in the same participants, which may provide insight into how
structural connectivity of the human brain shapes dynamic brain networks that
underlie cognitive function. These latter approaches are discussed in more detail
in the next section.
9.3. Dynamic Brain Networks
Structural brain connectivity provides a scaffold for the ongoing dynamics of the
brain that unfolds within it. In the adult brain, structural connections are likely to
be relatively stable across time (at least over periods of hours and days), while
even the most causal analysis of brain dynamics reveals that functional couplings
across the brain are highly variable, on a time scale of tens to hundreds of
milliseconds. It is these rapid fluctuations of functional connectivity (and by
extension, effective connectivity) that are associated with changes in perceptual
or cognitive state. Clearly, the static pattern of structural connectivity cannot
fully explain these rapid fluctuations, although it may more reliably shape
functional connectivity over longer time scales, for example when the brain is
cognitively “at rest”. The relationship between structural and functional
(effective) connectivity represents a major theoretical challenge in cognitive
neuroscience.
Network approaches begin to shed light on this relationship. One of the
useful attributes of network analysis tools is their applicability to both structural
and functional/effective connectivity data sets. For example, the centrality of a
vertex can be measured on the basis of its structural connection pattern, as well
as on the basis of the pattern of functional or effective connections it maintains.
This universality of network analysis approaches invites comparisons between
structural and functional data sets.
A rapidly increasing number of studies use network approaches in the
analysis of functional or effective connectivity patterns represented as graphs
(e.g. Dodel et al ., 2002; Salvador et al ., 2005a). There are numerous applications
of connectivity analyses to EEG, MEG and fMRI data sets. In most of these
approaches patterns of cross-correlation or coherence are represented as
undirected graphs with edges that represent the existence and, in some cases, the
Search WWH ::




Custom Search