PubMed Health⌕ Search
Home/Archive/PubMed

PubMed · 11032032

Spike-driven synaptic plasticity: theory, simulation, VLSI implementation.

S Fusi·M Annunziato·D Badoni·A Salamon·D J Amit
View original source ↗Export metadata ↓

Abstract

We present a model for spike-driven dynamics of a plastic synapse, suited for aVLSI implementation. The synaptic device behaves as a capacitor on short timescales and preserves the memory of two stable states (efficacies) on long timescales. The transitions (LTP/LTD) are stochastic because both the number and the distribution of neural spikes in any finite (stimulation) interval fluctuate, even at fixed pre- and postsynaptic spike rates. The dynamics of the single synapse is studied analytically by extending the solution to a classic problem in queuing theory (Takacs process). The model of the synapse is implemented in aVLSI and consists of only 18 transistors. It is also directly simulated. The simulations indicate that LTP/LTD probabilities versus rates are robust to fluctuations of the electronic parameters in a wide range of rates. The solutions for these probabilities are in very good agreement with both the simulations and measurements. Moreover, the probabilities are readily manipulable by variations of the chip's parameters, even in ranges where they are very small. The tests of the electronic device cover the range from spontaneous activity (3-4 Hz) to stimulus-driven rates (50 Hz). Low transition probabilities can be maintained in all ranges, even though the intrinsic time constants of the device are short (approximately 100 ms). Synaptic transitions are triggered by elevated presynaptic rates: for low presynaptic rates, there are essentially no transitions. The synaptic device can preserve its memory for years in the absence of stimulation. Stochasticity of learning is a result of the variability of interspike intervals; noise is a feature of the distributed dynamics of the network. The fact that the synapse is binary on long timescales solves the stability problem of synaptic efficacies in the absence of stimulation. Yet stochastic learning theory ensures that it does not affect the collective behavior of the network, if the transition probabilities are low and LTP is balanced against LTD.

Explore related subjects

Action PotentialsArtificial IntelligenceComputer SimulationComputersElectric ConductivityLong-Term PotentiationModels, NeurologicalNeural InhibitionNeural Networks, ComputerNeuronal PlasticityProbabilityStochastic ProcessesSynapsesJournal ArticleResearch Support, Non-U.S. Gov't

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

S Fusi, M Annunziato, D Badoni, A Salamon, D J Amit. 2000. Spike-driven synaptic plasticity: theory, simulation, VLSI implementation.. https://doi.org/10.1162/089976600300014917

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2007

Relationship of complex spike synchrony bands and climbing fiber projection determined by reference to aldolase C compartments in crus IIa of the rat cerebellar cortex.

Synchronous complex spike (CS) activity occurs most often among cerebellar Purkinje cells located in a narrow longitudinal (parasagittal) strip of cortex (synchrony band). The relationship of the anatomical organization of the olivocerebellar projection to these synchrony bands has not been investigated in detail. Thus, we studied this relationship by using the aldolase C (zebrin II) expression pattern, another landmark for the cerebellar longitudinal organization, as a reference frame in rat crus IIa. Crus IIa consists of 10 aldolase C-positive and -negative longitudinal compartments. Aldolase C labeling after multiple-electrode recording of CSs indicated that in lateral crus IIa (compartments 5+ to 7+) synchrony bands were generally constrained to single compartments. In contrast, in medial crus IIa (compartments 4a- to 5a-) the synchrony within and across the compartments was much higher than in lateral crus IIa, resulting in wide synchrony bands covering multiple compartments. Retrograde labeling of olivary neurons by injections of biotinylated dextran amine into aldolase C compartments in crus IIa showed that compartments in medial crus IIa were all innervated by the caudal part of the medial accessory olive. On the other hand, each aldolase C compartment in the lateral crus IIa was innervated by a region in a different subnucleus in the rostral inferior olive. These regions in different subnuclei were located close to each other. These results suggest that CS synchrony bands reflect the olivocerebellar compartmental projection pattern and neuronal coupling within a particular olivary subnucleus, and that medial and lateral crus IIa may be functionally distinct.

Action Potentials↗
PubMed2007

Motoneurons of the flight power muscles of the blowfly Calliphora erythrocephala: structures and mutual dye coupling.

The morphologies of the motoneurons of the dorsolongitudinal and the three dorsoventral flight power muscles (DLM, DVM 1-3) of Calliphora were investigated by means of cobalt backfills and intracellular biocytin stainings. The DLM is innervated by four prothoracic motoneurons supplying the four ventral muscle fibers and one mesothoracic motoneuron supplying the two dorsal fibers. The three fibers of the DVM 1 and the two fibers of the DVM 2 are innervated by five mesothoracic motoneurons, whereas the two fibers of the DVM 3 are innervated by two prothoracic motoneurons. In general, the motoneurons of each muscle have a common ventral soma cluster located in a characteristic position on the ipsilateral side of the thoracic ganglion, show similar dendritic arborizations in the mesothoracic wing neuropil, and have the same axon pathway. Only the soma of the common motoneuron of two dorsal fibers of the DLM is situated dorsally in the contralateral hemiganglion. The motoneurons of each muscle were found to be strongly dye coupled with each other, indicating that they are connected by gap junctions. In addition, the motoneurons of each muscle establish characteristic coupling patterns with the motoneurons of the other flight power muscles on both sides of the thorax and with two bilateral groups of local mesothoracic interneurons. The revealed coupling patterns are assumed to be of major relevance for the generation the characteristic, rhythmic flight activity of the motoneurons described in previous studies.

Action Potentials↗
PubMed2007

Systemic signalling in barley through action potentials.

Using apoplastic voltage- and ion selective microprobes, in barley leaves action potentials (APs) have been measured, which propagate acropetally as well as basipetally from leaf to leaf or from root to leaf following the application of mild salt stress (e.g. 30-50 mM KCl or NH(4)Cl) or amino acids (e.g. 1 mM glutamic acid or 5 mM GABA). Voltage changes were biphasic, followed an 'all-or-none' characteristic, and propagated at 20-30 cm min(-1) irrespective of the direction. With the salt-induced APs, a strong initial depolarization is the main AP-releasing factor that first causes Ca(2+) influx and then anion efflux. Ca(2+) influx coincides with an initial slower depolarization, the rapid anion efflux causes the typical voltage 'break-through'. Subsequently, K(+)-efflux starts after the depolarizing voltage has passed the K(+) equilibrium potential (inversion of the K(+) driving force). Glutamic acid and GABA induce APs not through membrane depolarization, but presumably by binding to a putative receptor or to ligand-gated Ca(2+)-conducting channels, respectively, followed by Ca(2+) induced activation of anion efflux. APs are accompanied by transient apoplastic pH increase (about 1 unit), and by cytoplasmic pH decrease (about 0.5 units). The apoplastic pH change is interpreted as an indicator of stress, the cytoplasmic pH change as a prerequisite for defence related gene activation. Since APs are released by agents added in a moderate concentration range, it is suggested that they may serve as first and fast systemic signals following attack from pathogens.

Action Potentials↗