Figure 7.
Model of barrel map development. A, Schematic of model. Initially, 20 principal whisker afferents (PW) and 10 adjacent whisker afferents (AW) contact a layer IV neuron, reflecting the initial somatotopic bias established by molecular genetic mechanisms. Each TCA has several “synapses” or active zones with the layer IV neuron, each of which will release a vesicle with probability Pr after an afferent action potential. TCAs fire either with bursts synchronized across same-whisker afferents or with uncorrelated spontaneous action potentials. B, Spike timing-dependent learning rule. Bursts of presynaptic activity result in a range of times that a vesicle will be released for a given action potential, which is dependent on the PPR. Lowering PPR changes the balance between synaptic strengthening (LTP, left side of green arrow) and synaptic weakening (LTD, right side of green arrow) during a burst, with lower PPRs favoring LTP relative to LTD. C, Simulation results for PPR = 0.4 with the synaptic weights of 30 afferents (20 PW/10 AW) plotted over simulation time and the PW initially favored by more contacts. The simulation shows robust whisker segregation (elimination of AW and stabilization/strengthening of PW) over a wide range of model parameters. D, Summary of the mean synaptic weight of the PW and AW for a range of PPRs plotted over simulation time. For PPR >0.6 segregation fails, and the afferents are no longer able to drive postsynaptic spiking. With lower PPR values (<0.6), TCAs reliably release neurotransmitter before the postsynaptic action potential, resulting in strengthening of the PW and weakening of the AW because of uncorrelated postsynaptic activity. Thus, high PPRs, or low neurotransmitter release efficacy, can prevent afferent segregation. E, The final synaptic weights of the simulations shown in D, with error bars demonstrating the final distribution of synaptic strengths for each afferent. F, The final synaptic weights in simulations run with a range of different LTP/LTD ratios [i.e., the area under the LTP curve (B; blue) divided by the area under the LTD curve (B; red)]. Changing the PPR is equivalent to directly modifying the LTP/LTD ratio, and segregation fails for LTP/LTD ratios less than ∼0.3.