The work presented in this document
describes the use of multi-site interaction with large cortical networks
developing ex-vivo, in a culture dish, to study basic biophysical
aspects of synchronization, adaptation, and modulation in neuronal assemblies.
This model of a "generic" neural network was examined both under
spontaneous self-organized activity as well as under several classes of
We describe a basic mode of activity of
neuronal assemblies. This basic mode of activity is analogous to an action
potential and a model based on the logistic equation is offered to describe it.
The distribution of firing rates of the neurons participating in the rising
phase of the network spike hints that the functional topology of connectivity
between neurons participating in network spikes is scale-free. Such a
scale-free topology enables rapid synchronization of activities within and
between coupled assemblies.
We also “looked within" the
assemblies in an attempt to examine processes related to adaptation and
learning at the level of the single neuronal network or assembly. This is done
by defining activation pathways between synchronic and diachronic neuronal
activities using a modified correlation measure and examining the evolution of
pathways under several types of perturbations to the system:
Changes in activation
pathways in response to a continuous electrical drive, simulating the effect of
a drive on the exploration processes in context of a learning task.
Selective adaptation to
stimuli from different sources and frequencies.
The effect of a
neuromodulator (Dopamine) on changes in activation pathways. We showed that
dopamine disperses correlations between individual neuronal activities while
preserving the global distribution of correlations at the network level. This
dispersion of correlations is very similar to the dispersion seen after
stimulating the network with an effective electrical drive.