Image Processing Reference
In-Depth Information
information from image data at macro-, meso- and microscales. We also discuss data
integration and neural network modeling, as well as the visualization, analysis and
comparison of brain networks.
21.1 Introduction
Connectomics is a field of neuroscience that analyzes neuronal connections. A
connectome is a complete map of a neuronal system, comprising all neuronal con-
nections between its structures. The term 'connectome' is close to the word 'genome'
and implies completeness of all neuronal connections, in the same way as a genome is
a complete listing of all nucleotide sequences. The goal of connectomics is to create
a complete representation of the brain's wiring. Such a representation is believed to
increase our understanding of how functional brain states emerge from their under-
lying anatomical structure [
89
]. Furthermore, it can provide important information
for the cure of neuronal dysfunctions like schizophrenia or autism [
83
].
Different types of connectivity can be distinguished.
Structural
or anatomical
connectivity usually refers to the “wiring diagram” of physical connections between
neural elements. These anatomical connections range in scale from those of local
circuits of single cells to large-scale networks of interregional pathways [
87
].
Func-
tional
connectivity is defined as “the temporal correlation between spatially remote
neurophysiological events” [
32
]. This can be seen as a statistical property; it does
not necessarily imply direct anatomical connections. Finally,
effective
connectivity
concerns causal interactions between distinct units within a nervous system [
32
].
Sporns et al. [
89
] differentiate between macro-, meso- and microscale connec-
tomes. At the
macroscale
, a whole brain can be imaged and divided into anatomically
distinct areas that maintain specific patterns of interconnectivity. Spatial resolution
at the macroscale is typically in the range of millimeters. One order of magnitude
smaller is the
mesoscale
connectome that describes connectivity in the range of
micrometers. At this scale, local neuronal circuits, e.g., cortical columns, can be dis-
tinguished. At the finest
microscale
, the connectome involves mapping single neu-
ronal cells and their connectivity patterns. Ultimately, connectomes from all scales
should be merged into one hierarchical representation [
89
].
Independently of the scale, the connectivity can be represented as a
brain graph
G
with nodes
N
and weighted edges
E
representing anatomical entities and
the degree of structural or functional interactions, respectively. Associated to each
abstract graph is a graph in real space that connects real anatomical entities. Neural
systems can be investigated by analyzing topological and geometrical properties of
these graphs and by comparing them. An equivalent way of representing an undi-
rected or directed brain graph is a
connectivity
or
association matrix C
, whose entries
c
ij
represent the degrees of interactions. Thresholding and sometimes also binarizing
them reveals the essential interactions. A spatial connectivity graph can be depicted
in real space, showing the actual physical structure of the neural system. A connec-
(
N
;
E
)
Search WWH ::
Custom Search