Fig. 7
From: Coupled symmetric and asymmetric circuits underlying spatial orientation in fruit flies
The impact of synaptic kinetics on spatial orientation memory in the full model. The NMDA receptors play a crucial role in the dynamics of the model. We tested how NMDA-specific kinetics influences the maintenance of the activity bump using the same task as in Figs. 5b–i and demonstrated the result in region-based firing rate plots. a The model circuit worked with a synaptic time constant as small as 50 ms (left) but failed to produce any activity bump with a much smaller time constant such as 10 ms (right). Note that reducing the time constant leads to a smaller synaptic current; hence we needed to compensate the reduced current by scaling up the synaptic weight. The best scaling factor was found to be 1.2 when the time constant equals 50 ms. b We removed the mechanism of magnesium block from NMDA and found that the removal produced little impact to the activity bump. c We turned off the saturation mechanism and discovered that, although a bump presented initially, it quickly disappeared after the initiation of body rotation. d We tested the model with different NMDA synaptic weights in the random walk task and measured the mean deviation (as in Fig. 6e). Each color denotes data for the mean deviation produced by the specified synaptic strength, e.q., ×1.3 for 130% of the original strength used in Fig. 6. The model worked reasonably well within a moderate range (×1.0–×1.2) of the NMDA synaptic strength. e For comparison, we further tested how the strength of the cholinergic synapse (Ach) influences the model performance and found that the model was insensitive to the change of the synaptic weight. To visualize the data in d and e, we fit the data to a square-root function as represented by the curves
