The Human Brain Network - Biological Networks - page 203

et al ., 2007), the creation of a map of the entire human brain at cellular resolution
still faces seemingly insurmountable methodological obstacles.
While a microscopic map of the human brain would provide connectivity
information at an unprecedented level of detail, much can be learned about the
global organization of the human brain from applying methods that deliver
mesoscopic or macroscopic datasets. Currently among the most promising
approaches are those that attempt to construct maps of the entire cerebral cortex
that detail the interconnections of segregated cortical regions. Several such maps
have been created in the past few years. Increasingly, investigators rely on
noninvasive imaging techniques such as diffusion tensor or diffusion spectrum
imaging (DTI and DSI, respectively) to obtain in vivo structural connectivity.
Diffusion imaging records signals from the white matter of the brain and allows
the computational reconstruction and estimation of fiber trajectories across the
brain (Conturo et al ., 1999; LeBihan, 2003).
Outside of diffusion imaging, researchers have utilized observed cross-
correlations in cortical thickness or volume across individuals to infer structural
connectivity patterns. Although the precise mechanism for this effect is still
unknown, previous research indicates that such gray-matter thickness
correlations can serve as reliable indicators for the presence or absence of
structural connections. He et al . (2007) generated a structural connection matrix
on the basis of such measurements obtained from 124 individual brain data sets.
An analysis of the resulting connection matrices using graph theory methods
demonstrated small-world connectivity and the existence of local communities of
areas forming structural modules. A more recent study of the connection
matrices obtained from cortical thickness correlations has focused in more detail
on the modularity structure of cortex (Chen et al ., 2008). The study generated
modules of brain regions from connectivity data and showed a high degree of
overlap between such modules and previously described functional brain
systems, e.g. those composed of primarily visual, sensorimotor and
auditory/language-related areas. While these studies have provided some of the
very first data sets mapping the human cortex in its entirety, it must be noted that
the methodological approach of inferring cortico-cortical connections on the
basis of cortical thickness or volume correlations has several shortcomings. By
its very nature, the method provides relatively indirect information about cortical
connection patterns. Furthermore it requires the aggregation of data from large
numbers of individuals to derive a single connection data set across a subject
group - connection patterns or individual brains cannot be accessed.
Other investigators have attempted to build whole-brain connection matrices
from diffusion imaging data. Iturria-Medina et al . (2007; 2008) have constructed