DIRECTLY COUPLING PRESYNAPTIC SHORT-TERM DYNAMICS WITH POSTSYNAPTIC LONG-TERM PLASTICITY SHAPES CONNECTIVITY AND DYNAMICS IN BRAIN NETWORKS

Activity-dependent synaptic plasticity is a fundamental learning mechanism that shapes connectivity and activity of neural circuits. Existing computational models of Spike-Time-Dependent Plasticity (STDP) model long-term synaptic changes with varying degree of biological details. A common approach is to neglect the influence of short-term dynamics on long-term plasticity, which may represent an oversimplification for certain long-term plasticity mechanisms. Thus, there is a need for new models to investigate how short-term dynamics influence long-term plasticity. To this end, we introduce a novel model, the Short-Long-Term STDP (SL-STDP) rule, which directly integrates short-term dynamics with postsynaptic long-term plasticity. We fit the new model to layer 5 visual cortex recordings (Sjostrom et al. 2001) and study how the short-term plasticity affects the firing rate frequency dependence of long-term plasticity. Our analysis reveals that the pre- and postsynaptic frequency dependence of the long-term plasticity plays a crucial role in shaping the self-organization of recurrent neural networks and their information processing through the emergence of sinks and source nodes. We applied the SL-STDP rule to RNNs and found that the neurons of SL-STDP network self-organized into distinct firing rate clusters and stabilized the dynamics. Finally, we evaluated how the modified connectivity affects networks’ information capacities in reservoir computing tasks and found improved memory capacity. Our study demonstrates that short-term dynamics–induced changes in the frequency dependence of long-term plasticity play a pivotal role in shaping brain network dynamics and link synaptic mechanisms to information processing.