Information Technology Reference
In-Depth Information
Fig. 12.6. Four motifs of spatio-temporal propagation of activity in a clustered network. The four
figures show the spread of activity during SBEs taken from a single culture that had architecture of
four interconnected sub-networks. The SBEs were clustered into groups of similar propagation
motifs, and for each SBE the neurons were color-coded according to their order of firing (blue -
first firing neuron, red - last neuron). The plots show the overlapped neuron order of firing of all
SBEs belonging to each motif. Figure taken with permission from Baruchi
et al
., in press.
2000a; Latham
et al
., 2000b; Shein
et al
., in press) or neurons with an enhanced
sensitivity (Eytan and Marom, 2006; Shein
et al
., in press), interplay between
inhibition and excitation activity in the network (Streit
et al
., 2001), or the
presence of localized initiation zones characterized by high neuronal density and
by recurrent and inhibitory network connections (Feinerman
et al
., 2007). The
fact that cultures spontaneously generate repeating motifs is of significant
importance as it proves that neuronal networks in-vitro are capable of
maintaining long-term memory. In a recent paper it was shown that new motifs
can also be imprinted into the culture by localized chemical stimulation (Baruchi
and Ben-Jacob, 2007). The number of motifs and how they correlate among
themselves may shed light on the morphology and architecture of the culture.
This may become apparent through the relationship between excitatory and
inhibitory neurons (Volman
et al
. 2005), the type of connectivity within the
culture, or the geometry of the culture sections (Blinder
et al
., 2005; Raichman
et al
., 2006).
One modeling approach to study the emergence of different SBEs motifs
utilizes the “function follows form” principle, based on the realization that
different network architectures might support different types of dynamics.
Consequently, it was shown in Volman
et al
. (2005) how synchronized bursting
events with different internal spatio-temporal organization arise in a network
with non-homogeneous connectivity. A network composed of two homogeneous
sub-networks that had a small overlap, supported two types of SBEs, each one
corresponding to the activity initiating in one sub-networks and spreading,
through the overlap region, to the other one. Moreover, the temporal appearance
of different SBEs could be controlled by imposing regulatory constraints on the
