Detection of nonlinearity in the neural activity (Q4377230)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 1115876
Language Label Description Also known as
default for all languages
No label defined
    English
    Detection of nonlinearity in the neural activity
    scientific article; zbMATH DE number 1115876

      Statements

      Detection of nonlinearity in the neural activity (English)
      0 references
      0 references
      0 references
      0 references
      8 February 1999
      0 references
      neural activity
      0 references
      deterministic and stochastic phase-locking
      0 references
      sinusoidal forcing
      0 references
      low-dimensional deterministic chaos
      0 references
      neural networks
      0 references
      nonlinear dynamics
      0 references
      nonlinear forecasting
      0 references
      Individual neurons are characterised by nonlinear behaviours like the threshold for firing or refractoriness and therefore they are good candidates for exhibiting complex dynamics. Moreover it has been shown that external perturbations can induce responses ranging from deterministic to stochastic phase-locking up to chaotic activity; these results support the view that single neural cells are able, due to their intrinsic nonlinearity, to exhibit a variety of behaviours. On the other hand, the deterministic process underlying neural activity could be obscured by the noise always present in biological systems. So further insight into the generation mechanism of these behaviours can be derived by analysing single neuron models of different complexity, with the aim at reproducing the experimental responses. These problems will be examined herein, with particular reference to the complex responses observed in snail neurons under weak periodic stimulation. At least two kinds of dynamical behaviour are known to occur in neurons, the deterministic phase-locking and the stochastic one, but intermediate situations can also be observed, as demonstrated by our results. So the simple inspection of the interspike interval (ISI) histogram cannot be sufficient to classify the kind of behaviour. In this paper, both the experimental responses and those of single neuron mathematical models have been analysed in the framework of nonlinear dynamics. The models under examination herein are the Leaky Integrate and Fire (LIF) and the Fitzhugh-Nagumo (FHN), both driven by a weak sinusoidal signal plus noise. The analysis is based on the well established nonlinear forecasting methods, able to detect the short-term structure of the interspike interval time series. The underlying idea is that, if there is any level of determinism in a system, it should manifest itself as some predictability of its evolution. Moreover, to exclude predictability that can be accounted for by a linear stochastic process, the statistical significance of the results was assessed using surrogate data. These techniques allow the nonlinearity present in the analysed time series to be detected and quantified.
      0 references

      Identifiers