Anmerkungen:
Diese Datenquelle enthält auch Bestandsnachweise, die nicht zu einem Volltext führen.
Beschreibung:
In order to better understand the relationship of connectivity and function in the brain at different scales, in this work we show the results of using point neuron network simulations to complement connectivity information of whole brain simulations based on a dynamic neuron mass model. In our multiscale approach, we simulate a whole brain parcellated into 68 regions using a similar setup as described in Deco et al. 2014. Each region is modeled as a dynamic neuron mass and, in parallel, we also model each region as small point neuron populations in NEST. Structural plasticity in NEST is then used to calculate inner inhibitory connectivity required to match experimentally observed firing rate behavior. An interactive tool was developed in order to steer the structural plasticity algorithm and take all the regions, which are also highly interconnected, to their ideal firing activity. An inner inhibition fitting was first proposed in the work by Deco 2014, using an iterative tunning method. In our work, we allow the point neuron network to self generate the connectivity using simple homeostatic rules and then we feed this information to the dynamic mass model simulation. With the resulting connectivity data from the NEST simulations and experimentally obtained DTI inter region connectivity, simulations of the whole brain producing results comparable to experimental fMRI data are possible. Using this approach, the fitting and parameter space exploration times are reduced and a new way to explore the impact of connectivity in function at different scales is presented.