Topological Effects of Synaptic Time Dependent Plasticity

We show that the local Spike Timing-Dependent Plasticity (STDP) rule has the effect of regulating the trans-synaptic weights of loops of any length within a simulated network of neurons. We show that depending on STDP’s polarity, functional loops are formed or eliminated in networks driven to normal spiking conditions by random, partially correlated inputs, where functional loops comprise weights that exceed a non-zero threshold. We further prove that STDP is a form of loop-regulating plasticity for the case of a linear network comprising random weights drawn from certain distributions. Thus a notable local synaptic learning rule makes a specific prediction about synapses in the brain in which standard STDP is present: that under normal spiking conditions, they should participate in predominantly feed-forward connections at all scales. Our model implies that any deviations from this prediction would require a substantial modification to the hypothesized role for standard STDP. Given its widespread occurrence in the brain, we predict that STDP could also regulate long range synaptic loops among individual neurons across all brain scales, up to, and including, the scale of global brain network topology.

💡 Research Summary

The paper investigates how the local spike‑timing‑dependent plasticity (STDP) rule influences the formation and elimination of synaptic loops of arbitrary length in neuronal networks. Using two complementary approaches—a large‑scale spiking network simulation and an analytical treatment of a linear network—the authors demonstrate that STDP acts as a loop‑regulating mechanism. In the simulations, a network of 1,000 integrate‑and‑fire neurons receives random, partially correlated inputs while each synapse follows a conventional STDP learning rule. Functional loops are defined as directed paths that return to the originating neuron, with all constituent synaptic weights exceeding a preset threshold. When the standard STDP polarity (pre‑post potentiation, post‑pre depression) is applied, loops of all lengths progressively weaken and eventually disappear, especially those longer than five synapses. Conversely, reversing the polarity (pre‑post depression, post‑pre potentiation) leads to a systematic strengthening of loop weights, resulting in a network dominated by strong recurrent circuits and increased firing variability. These results indicate that STDP does not merely adjust pairwise correlations but can globally shape the network’s cyclic topology.

The analytical part models a linear system with weight matrix (W) driven by Gaussian input with covariance (\Sigma). By expressing the STDP update as a continuous-time differential equation, the authors derive (\dot{W}=A W - W A^{\top}), where (A) encodes the convolution of the input covariance with the STDP kernel. This dynamics inherently suppresses antisymmetric components of (W). Eigenvalue analysis shows that antisymmetric (loop‑supporting) components generate eigenvalues with positive real parts, leading to instability. The STDP‑induced flow drives these eigenvalues toward the left half‑plane, eliminating unstable loops and stabilizing the system in a predominantly feed‑forward configuration. Thus, STDP can be formally classified as a “loop‑suppressing” form of plasticity, even in a purely linear setting.

From a neurobiological perspective, the findings generate a clear prediction: in brain regions where canonical STDP operates under normal spiking conditions, synapses should be organized mainly in feed‑forward pathways across all spatial scales. Observations of robust recurrent loops would therefore imply either a modification of the STDP rule (e.g., polarity reversal, broadened timing windows) or the presence of additional modulatory mechanisms such as neuromodulator‑gated plasticity, homeostatic scaling, or structural remodeling. The authors suggest experimental tests involving in‑vivo imaging of synaptic connectivity combined with precise manipulation of STDP timing windows, to quantify the ratio of loop to feed‑forward connections.

Overall, the study bridges a gap between local synaptic learning rules and global network topology, proposing that a ubiquitous plasticity mechanism can dictate large‑scale circuit architecture. This insight has implications for understanding developmental wiring, learning‑induced reorganization, and pathological rewiring in disorders where STDP dynamics may be altered.