The cerebral cortex exhibits highly complex dynamic regimes during spontaneous activity. A plethora of parameters were tried to capture this complexity focusing on different features. One of the most relevant, showed by the spontaneously running cortical networks, is represented by synchronies. While the instantaneous higher order interactions, that incorporate also higher order synchronies, have been well described by weak pairwise correlations [1], their temporal dynamics has not yet been thoroughly analyzed. In a previous paper we showed that multiunit firing activity exhibits intermittent chaotic behavior [2]. We therefore focused on predictability of higher order loose synchronies (LSs), i.e. firing events jointly occurring within 30-50ms temporal windows. We analyzed extracellular simultaneous multiple recordings of spontaneously active Somatosensory Primary cortices of lightly gas-anesthetized rats. We first developed a statistical method based on a hypothesis test combined to a data clustering to extract and classify synchronous and non-synchronous events. The resulting symbolic sequence represents the multiunit spiking activity where some symbols are associated with LSs and others with non-synchronous events. We approximated the Kolmogorov complexity of these sequences within fixed length sliding windows by the compressed sequence length (CSL) computed with a set of Unix compressors (zip, gzip, bzip2) [3]. On comparing the real sequences (RS) with surrogate sequences obtained through random permutations, we found long strings of significantly low CLS regions in comparison with the surrogated sequences (SS) (Fig A). The rate of LS occurrences showed high positive correlation with CLS values. LS predictability was analyzed with Variable Order Markov Model techniques estimating both short and long range sequence dependencies [4]. We found that the LSs in RS were 10 to 100% more predictable than LSs in SS, only the last 5 to 15 symbols were relevant for prediction (Fig B). Unexpectedly, the rate of correct LS predictions wasn’t significantly correlated with CLS. Finally, the rate of LS prediction and the rate of LS occurrence resulted positively correlated. These results deliver important cues on the events leading to the occurrence of LS. The high variability in predictions suggests that the cortical LSs may potentially endorse diverse tasks merged in the shared functional state of spontaneous activity.

Higher order synchrony predictability in somatosensory cortex during spontaneous activity / A.G. Zippo, R. Storchi, J. Lin, G. Caramenti, M. Valente, G.E.M. Biella. - In: FRONTIERS IN COMPUTATIONAL NEUROSCIENCE. - ISSN 1662-5188. - 4:(2010), pp. 1-1. ((Intervento presentato al convegno Bernstein Conference on Computational Neuroscience tenutosi a Berlin nel 2010 [10.3389/conf.fncom.2010.51.00140].

Higher order synchrony predictability in somatosensory cortex during spontaneous activity

A.G. Zippo
Primo
;
J. Lin
Secondo
;
2010

Abstract

The cerebral cortex exhibits highly complex dynamic regimes during spontaneous activity. A plethora of parameters were tried to capture this complexity focusing on different features. One of the most relevant, showed by the spontaneously running cortical networks, is represented by synchronies. While the instantaneous higher order interactions, that incorporate also higher order synchronies, have been well described by weak pairwise correlations [1], their temporal dynamics has not yet been thoroughly analyzed. In a previous paper we showed that multiunit firing activity exhibits intermittent chaotic behavior [2]. We therefore focused on predictability of higher order loose synchronies (LSs), i.e. firing events jointly occurring within 30-50ms temporal windows. We analyzed extracellular simultaneous multiple recordings of spontaneously active Somatosensory Primary cortices of lightly gas-anesthetized rats. We first developed a statistical method based on a hypothesis test combined to a data clustering to extract and classify synchronous and non-synchronous events. The resulting symbolic sequence represents the multiunit spiking activity where some symbols are associated with LSs and others with non-synchronous events. We approximated the Kolmogorov complexity of these sequences within fixed length sliding windows by the compressed sequence length (CSL) computed with a set of Unix compressors (zip, gzip, bzip2) [3]. On comparing the real sequences (RS) with surrogate sequences obtained through random permutations, we found long strings of significantly low CLS regions in comparison with the surrogated sequences (SS) (Fig A). The rate of LS occurrences showed high positive correlation with CLS values. LS predictability was analyzed with Variable Order Markov Model techniques estimating both short and long range sequence dependencies [4]. We found that the LSs in RS were 10 to 100% more predictable than LSs in SS, only the last 5 to 15 symbols were relevant for prediction (Fig B). Unexpectedly, the rate of correct LS predictions wasn’t significantly correlated with CLS. Finally, the rate of LS prediction and the rate of LS occurrence resulted positively correlated. These results deliver important cues on the events leading to the occurrence of LS. The high variability in predictions suggests that the cortical LSs may potentially endorse diverse tasks merged in the shared functional state of spontaneous activity.
computational neuroscience
Settore INF/01 - Informatica
Settore ING-INF/05 - Sistemi di Elaborazione delle Informazioni
Article (author)
File in questo prodotto:
Non ci sono file associati a questo prodotto.
Pubblicazioni consigliate

Caricamento pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/2434/249915
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact