Image Processing Reference
In-Depth Information
Fig. 21.6
Reverse engineering of a cortical column. Reconstructed dendrites (
a
) are replicated and
inserted into the column reference frame according to a given neuron density (
b
). By determining
the local structural overlap with axons projecting into the column (
c
), the number of synapses for
different post-synaptic cell types can be estimated. (
d
) Shown are synapse densities for two cell
types. Figure created from data published in [
65
]
density and replicating and inserting the dendrite morphologies into the reference
frame according to the given cell type frequency (see Fig.
21.6
). Differences in
synaptic densities between cell types can be quantified and visualized [
65
]. Based on
the estimated number of synapses per cell, a complete network wiring is established
to study network function using numerical simulation [
57
].
Extracting relevant neurobiological knowledge from such network models is a
challenging task. Whereas computation of specific quantities for comparison with
literature results in order to validate the model is straightforward, exploratory knowl-
edge discovery within such large, complex networks is not. Easy-to-use tools are
needed to let the neurobiologist query and visualize the structural and functional
properties of such networks or ensembles of network realizations. As network mod-
els are increasing in size, large data handling will be a challenging issue as well.
21.8 Network Analysis and Comparative Visualization
A recent innovation in neuroimaging is connectivity analysis, in which the anatom-
ical or functional relation between different (underlying) brain areas is calculated
from data obtained by various modalities, allowing researchers to study the resulting
networks
of interrelated brain regions. Of particular interest are
comparisons
of func-
tional brain networks under different experimental conditions and between groups
of subjects.
Search WWH ::
Custom Search