Avalanches neuronais durante o ciclo sono-vigÃlia de ratos
AUTOR(ES)
Tiago Lins Ribeiro
DATA DE PUBLICAÇÃO
2009
RESUMO
Neuronal avalanches are spatiotemporal patterns of neuronal activity that occur spontaneously in supercial layers of the mammalian cortex under various experimental conditions. Previous studies in brain slices and cultured cells have shown that the distribution of sizes P(s) of neuronal avalanches obeys a power law relationship: P(s)  sÂa , with exponent a 3=2. This is compatible with the idea that neuronal networks in vitro operate in a critical regime, but little is known about the properties of neuronal avalanches in vivo. In this dissertation the statistics of neuronal avalanches, obtained from extracellular action potential recordings (spikes) in rats surgically implanted with multielectrode arrays, were investigated. The animals were continually recorded for many hours across the sleep-wake cycle, before, during and after a novel object exploration task. Anesthetized rats were also recorded. In particular, we sought statistical signatures in the in vivo recordings that allowed us to test the hypothesis of criticality in the brain. The data were divided in temporal bins Dt. An avalanche was dened as a sequence of bins with non-null activity. We have done statistical analysis regarding the size (number of spikes) s, duration d and interval t between avalanches, for a given Dt. We observed that the statistics of neuronal avalanches vary depending on the behavioral state. In any state, longer and larger avalanches become more frequent after the objects exploration experiment. We found distributions of size similar to the lognormal distribution: P(s)ÂsÂ1 exp[Âc(ln(s)Âb)2]. We studied an excitable cellular automaton model and showed that similar distributions can be obtained from a system in its critical point when there is undersampling. In other words, systems exhibiting power law size distributions when the recordings are made from all network sites, exhibit lognormal-like distributions when only a small fraction of the network elements are measured. In addition, we explored the temporal dynamics and observed that the family of distributions of intervals D(t ;sc) between consecutive avalanches with size s  sc obeys a scaling law: D(t ;sc) = R(sc)F(tR(sc)), with R(sc) = ht iÂ1 and F a scaling function. The results are similar to scaling laws found in solar ares, forest res, fractures, earthquakes and other systems. The presence of scaling in the temporal dynamics of the system is compatible with the idea of self-organizing criticality and was found independently of the behavioral state, brain region and experiment period
ASSUNTO(S)
cellular automata autÃmatos celulares statistical mechanics directed percolation self-organized criticality mecÃnica estatÃstica percolaÃÃo direcionada criticalidade auto-organizada neurociÃncia neuroscience fisica
