Stable Hebbian Learning from Spike Timing-Dependent Plasticity

Abstract

We explore a synaptic plasticity model that incorporates recent findings that potentiation and depression can be induced by precisely timed pairs of synaptic events and postsynaptic spikes. In addition we include the observation that strong synapses undergo relatively less potentiation than weak synapses, whereas depression is independent of synaptic strength. After random stimulation, the synaptic weights reach an equilibrium distribution which is stable, unimodal, and has positive skew. This weight distribution compares favorably to the distributions of quantal amplitudes and of receptor number observed experimentally in central neurons and contrasts to the distribution found in plasticity models without size-dependent potentiation. Also in contrast to those models, which show strong competition between the synapses, stable plasticity is achieved with little competition. Instead, competition can be introduced by including a separate mechanism that scales synaptic strengths multiplicatively as a function of postsynaptic activity. In this model, synaptic weights change in proportion to how correlated they are with other inputs onto the same postsynaptic neuron. These results indicate that stable correlation-based plasticity can be achieved without introducing competition, suggesting that plasticity and competition need not coexist in all circuits or at all developmental stages.

Temporal Characteristics of the Predictive Synchronous Firing Modeled by Spike-Timing-Dependent Plasticity

“…The rule induces no competition between synapses and consequently no activity regulation (Kistler and van Hemmen 2000;van Rossum et al 2000;Rubin et al 2001;Câteau and Fukai 2003). The results obtained (data not shown) were quite similar to those shown in Figure 4C; as expected, synchronous firing was not predictive in the absence of the activity regulation.…”

“…The standard multiplicative STDP rule (Kistler and van Hemmen 2000;van Rossum et al 2000;Rubin et al 2001) was incapable of organizing predictive synchrony. The results, however, do not exclude the possible emergence of predictive synchrony under multiplicative-type rules.…”

Time scales of memory, learning, and plasticity

“…Neural systems have several different mechanisms (Bi and Poo 1998;Turrigiano et al 1998;Royer and Paré 2003;Zhou et al 2003), which are also used in theoretical studies (Rochester et al 1956;Bienenstock et al 1982;Riedel and Schild 1992;Van Rossum et al 2000) to avoid these runaway dynamics.…”

“…This is known as spike-timing-dependent plasticity (STDP). Synapses in different brain areas under different conditions exhibit experimentally different weight change curves, which are then fitted mostly as exponential kernels (Van Rossum et al 2000;Froemke et al 2005). A possible mathematical description is given in Table 1-6.…”

Sensory experience modifies spontaneous state dynamics in a large-scale barrel cortical model

“…Most of the available spiking neuron models with STDP assume a default symmetric profile at all synapses with the same learning windows (e.g. Kitano & Fukai 2004, Rao & Sejnowski 2001, Van Rosssum et al 2000. Given that STDP learning rules can modify sensory processing, for example modify receptive fields and orientation preferences in adult cats (Yao & Dan 2001, Fu et al 2002, it is likely that the temporal window of learning plays an important role.…”