Figure 1.
Scheme of the structure and topology of the network.
A. Schematic representation of network connections between granule cells (GC), basket cells (BC), hilar perforant-path associated cells (HC) and mossy cells (MC). GC synapses are indicated by black lines, BC synapses are indicated by blue lines, HC synapses are indicated by orange dashed lines and MC synapses are indicated by green lines. Mossy fiber sprouting (MFS) is indicated by a black dashed line and the perforant-path input to GC and BC is indicated by a bold black line. The segments into which the GC dendritic tree are divided are indicated in yellow (granule cell layer dendrites), orange (proximal dendrites), green (medial dendrites) and blue (distal dendrites). B. Positioning of the GC within the molecular layer with its four subdivisions: granular layer (GL) in blue, inner molecular layer (IML) in orange, middle molecular layer (MML) in green and outer molecular layer (OML) in blue. The figure shows that perforant-path synapses to GCs are located in OML dendrites and mossy fiber sprouting synapses to GCs are located in IML dendrites regardless of the type of dendritic segment in these layers. C. Schematic representation of the dentate gyrus model with control GCs. D. Schematic representation of the dentate gyrus model with a fraction of the control GCs replaced with cells from PILO-treated animals (in red).
Table 1.
List of the three-dimensional GC reconstructions from neuromorpho.org used to build the mature DG model0.
Figure 2.
Schematic representation of the strategies used to classify the dendrites and place the synapses.
A. Classification of a dendrite as proximal (PD), medial (MD) and distal (DP) based on the distance from soma up to the branching point of the dendrite (yellow line). B. Location of the synapses depending on the sublayer in which a dendrite section is located.
Table 2.
List of the different families of models developed in the present study.
Figure 3.
Comparison of the intrinsic excitability of the mature and newborn GCs from control (YOUNG) and PILO groups.
The cases for model cells without spine loss are indicated in blue and the cases for model cells with spine loss are indicated in red. A. Average rheobase current (measured as explained in the text) for the three model cell groups. B. Average number of spikes evoked by the train of synaptic-like pulses as explained in the text.
Figure 4.
Examples of raster plots for different DG network models.
A. Response of a mature network without mossy fiber sprouting. B. Response of a mature network with 10% mossy fiber sprouting. C. Response of a mature network with 50% mossy fiber sprouting. D. Response of a mature network with 50% newborn control GCs without mossy fiber sprouting. E. Response of a mature network with 10% newborn control GCs and 10% mossy fiber sprouting. F. Response of a mature network with 50% newborn control GCs and 10% mossy fiber sprouting. G. Response of a mature network with 50% newborn PILO GCs without mossy fiber sprouting. H. Response of a mature network with 10% newborn PILO GCs and 10% mossy fiber sprouting. I. Response of a mature network with 50% newborn PILO GCs and 10% mossy fiber sprouting. J. Response of a mature network with 10% newborn PILO GCs with spine loss and without mossy fiber sprouting. K. Response of a mature network with 10% newborn PILO GCs with spine loss and 10% mossy fiber sprouting. L. Response of a mature network with 50% newborn PILO GCs with spine loss and 10% mossy fiber sprouting. The latter three figures for the cases with spine loss were obtained for the SL2 case.
Figure 5.
Overall frequency of the network (total spike count over simulation time) for the different network families as a function of the proportion of newborn GCs.
All cases are for 10% of mossy fiber sprouting. Each point in the graph corresponds to an average over 20 randomly generated models of the corresponding family. The different network families are indicated by different colors and their codes are the same as defined in Table 2. A. Cases in which spine loss was represented by a 50% reduction in the probability of connections with newborn GCs. The MPSL1 network type only had results up to the insertion of 80% of newborn cells because the insertions of 90% and 100% of newborn GCs did not allow the maintenance of the convergence and divergence factors of the network. The error bars represent the standard error. The overall frequency is significantly different over the different proportions () and types of altered cells and spine loss inserted (
) and their interaction (
). B. Cases in which spine loss was represented by a reduction of 25%, 50% and 75% in the probability of newborn GCs receiving connections. The coding is the same as in A but followed by the corresponding reduction in the probability of connection.