Consiglio Nazionale delle Ricerche

Tipo di prodottoAbstract in rivista
TitoloPartial synchronization in diluted neural networks
Anno di pubblicazione2009
Autore/iSimona Olmi (1,3); Roberto Livi (1,3); Antonio Politi (2); Alessandro Torcini (2,3)
Affiliazioni autori1) Department of Physics, University of Florence, Sesto Fiorentino, I-50019, Italy 2) Istituto dei Sistemi Complessi, CNR, Sesto Fiorentino, I-50019, Italy 3) Istituto Nazionale di Fisica Nucleare, Sezione di Firenze, Sesto Fiorentino, I-50019, Italy
Autori CNR e affiliazioni
  • inglese
AbstractPartial synchronization (PS) is a collective dynamics discovered in fully coupled networks of leaky integrate-and-fire (LIF) neurons [1,2]. This dynamical state is quite peculiar, since the mean-field is characterized by a periodic dynamics, while the single neurons do not show any synchronization and behaves quasi-periodically in time. PS arises from the destabilization of a regime characterized by a constant mean field and by periodic motion of the single neurons. This regime, termed "Splay State," as been widely studied in several context, including computational neuroscience [3]. Our aim is to investigate the role of network dilution on the stability properties of PS and splay states and the influence of dilution on the transition scenario leading from the splay state to partial synchronization. In particular, we have examined a pulse-coupled excitatory network of LIF neurons with 20% of randomly broken links in two different configurations: quenched, where the network topology is fixed, and annealed, where the links to break are chosen at each spike emission. For both configurations, we have been able to rewrite the dynamics of the network (described by a set of differential equations) as an exact event driven map by extending the approach developed in [4]. We observe that the splay state and the PS state, as well as the transition scenario, are quite robust with respect to dilution, both for the quenched and annealed configurations. Moreover the dynamics for finite number of neurons is chaotic at variance with the fully coupled case. The dependence of the degree of chaotic behavior (characterized in terms of the maximum Lyapunov exponent) on the number of neurons N is analyzed, as a first results we notice that finite-N effects are more relevant in the quenched configuration with respect to the annealed one. Finally, the possibility to reproduce the deterministic dynamics of the diluted network for finite N, by introducing an additive colored noise term in the equations describing the fully coupled case, is investigated. The noise term amplitude and correlation are determined by the analysis of the mean field properties associated with finite diluted networks.
Pagine daP274
Pagine a-
Pagine totali-
RivistaBMC neuroscience (Online)
Attiva dal 2000
Editore: BioMed Central, - London
Paese di pubblicazione: Regno Unito
Lingua: inglese
ISSN: 1471-2202
Titolo chiave: BMC neuroscience (Online)
Titolo proprio: BMC neuroscience (Online)
Titolo abbreviato: BMC neurosci. (Online)
Titolo alternativo: Neuroscience (Online)
Numero volume della rivista10
Fascicolo della rivistaSuppl1
Verificato da referee-
Stato della pubblicazionePublished version
Indicizzazione (in banche dati controllate)-
Parole chiavesynchronization, neural networks
Link (URL, URI)
Titolo parallelo-
Scadenza embargo-
Data di accettazione-
Note/Altre informazioniPoster. Eighteenth Annual Computational Neuroscience Meeting: CNS*2009, Berlin, Germany, 18-23 July, 2009.
Strutture CNR
  • ISC — Istituto dei sistemi complessi
Moduli/Attività/Sottoprogetti CNR
  • MD.P02.017.001 : comportamento dinamico di sistemi complessi
Progetti Europei-
Poster pubblicato
Descrizione: Open access
Tipo documento: application/x-pdf

Dati associati a vecchie tipologie
I dati associati a vecchie tipologie non sono modificabili, derivano dal cambiamento della tipologia di prodotto e hanno solo valore storico.
Tipologia di comunicazione in rivistaAltra comunicazione (comment, in brief, forum)

Dati storici
I dati storici non sono modificabili, sono stati ereditati da altri sistemi (es. Gestione Istituti, PUMA, ...) e hanno solo valore storico.
Area disciplinarePhysics
Area valutazione CIVRScienze fisiche
ConvegnoEighteenth Annual Computational Neuroscience Meeting: CNS*2009
NoteBMC Neuroscience from Eighteenth Annual Computational Neuroscience Meeting: CNS*2009, Berlin, Germany, 18–23 July, 2009. BMC Neuroscience 2009, 10(Suppl 1):P274. BioMed Central Ltd. doi:10.1186/1471-2202-10-S1-P274