Email updates

Keep up to date with the latest news and content from BMC Neuroscience and BioMed Central.

This article is part of the supplement: Twentieth Annual Computational Neuroscience Meeting: CNS*2011

Open Access Poster presentation

Spike distributions of a population based hierarchical network of the rat amygdaloid complex

Oliver Schmitt1*, Peter Eipert1, Andreas Wree1 and Klaus-Peter Schmitz2

Author Affiliations

1 Department of Anatomy, University of Rostock, Germany

2 Department of Biomedical Engineering, University of Rostock, Germany

For all author emails, please log on.

BMC Neuroscience 2011, 12(Suppl 1):P285  doi:10.1186/1471-2202-12-S1-P285


The electronic version of this article is the complete one and can be found online at: http://www.biomedcentral.com/1471-2202/12/S1/P285


Published:18 July 2011

© 2011 Schmitt et al; licensee BioMed Central Ltd.

This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Background

The amygdaloid nuclear complex (ANC) is a conglomerate of gray matter lying in the depth of the anteromedial temporal lobe ventral to the lentiform nucleus. ANC is critical for producing appropriate emotional and behavioral responses to biologically relevant sensory stimuli. It constitutes an essential link between sensory and limbic areas of the cerebral cortex and subcortical brain regions, such as the hypothalamus, brainstem, and striatum, that are responsible for eliciting emotional and motivational responses. The connectome of the rat ANC is modeled in neuroVIISAS (a platform for mapping, visualization and analysis of connectivity data from tracing studies) that is used to generate realistic large scale networks of all regions of the central nervous system. The models are transferred by a PyNEST interface to the simulation engine NEST and simulation results can be analyzed and visualized by neuroVIISAS.

Methods

700 Stereotactic tracing studies have been evaluated (metastudy) with regard to connection sources, targets and weights (density of a connection) to obtain a directed graph with weights of the ANC and further regions of the central nervous system (CNS) of the rat. neuroVIISAS is developed in JAVA using the Java Development Kit (JDK) (v1.6.0) and the development environment Eclipse (v3.2). The Visualization Toolkit (VTK) is used for 3D-visualization of brain region surfaces. The framework is used to analyze connectome at different levels of granularity by graph analytical methods. Based on results of local or region specific graph properties we are able to define spike injections in dependence of spatial properties of the region locations and their estimated total number of neurons. The definition of the network and simulation parameters is written into a PyNEST file that is processed by the simulation engine NEST. Simulation results are read by neuroVIISAS for analysis and visualization.

Results

So far, the connectome of the bilateral rat CNS consists of 6690 regions and 39310 connections. 274 connections of 91 subregions of the ANC are found in 700 stereotactic tracing studies. The Internal connectivity of the ANC has an average cluster coefficient (ACC) of 0.285 (mean ACC of 1000 Barabasi-Albert (BA1000) network simulations: 0.084) and an average path length (APL) of 3.064 (APL-BA1000: 3.091) indicating a scale-free network. The posteromedial cortical nucleus has the highest internal in- (13) and outdegree (19), a low Shapley-rate of -0.287 and highest value for cyclic connections. The graph motif statistics of the internal ANC network revealed abundant divergence, chain and motifs with reciprocal connections (p ≤ 0.001 for BA1000). In dependence of the volume of the regions the number of neurons (6*105) were estimated and downscaled by a factor of 1000 to allow simulation of leaky integrate and fire (LIF) neurons. 13069 LIF neurons (20% inhibitory, 80% excitatory) were generated and the simulations were performed for 100 ms with a spike injection of 5 ms within the PMCO. The number of connections between populations was correlated with the projection weights. Average firing frequencies of LIF neurons lie between 50 Hz and 475 Hz. The anterior basomedial nucleus and the medial amygdaloid nucleus anterodorsal part are first neighbors of PMCO and show the largest similarity (0.908) of their 0.3 ms binned spike distributions. Furthermore, in balanced randomizations of population internal connectivity as well as in small-world and scale free internal connection scheme, we observed oscillations that depend on the neighborhood distance to the spike injection region.

Conclusions

Our data provide evidence for a topographic dependence of spike dynamics in multi-population models with different internal connection schemes. Furthermore, a time dependence of the spike distributions of populations having a larger distance to the spike injection populations produce delayed spike responses indicating a spatiotemporal consistency of the connectional ANC model.