International Conference on NonLinear Theory and its Applications NOLTA2004 PowerPoint PPT Presentation

presentation player overlay
1 / 13
About This Presentation
Transcript and Presenter's Notes

Title: International Conference on NonLinear Theory and its Applications NOLTA2004


1
International Conference on NonLinear Theory and
its Applications NOLTA2004
  • Macroscopic Manifestation of
  • Neural Connectivity in EEG Signals
  • Cristian CARMELI, Oscar DE FEO, Giorgio INNOCENTI
  • 30 November 2004

2
Outline
  • Why and How do neurons cooperate?
  • The Binding Problem
  • Assessment of cooperative activity
  • Synchrony
  • Its Drawbacks
  • Alternative approach
  • Application to EEG signals
  • Results
  • Conclusions

3
A Simple Example The Visual Pathway
4
  • The Binding Problem
  • Binding is the process responsible for
    functionally linking the distributed activity of
    large neuronal populations.
  • Problem identification of the mechanisms of such
    process.

5
The Traditional Binding Solution
  • Cooperativeness
    Synchronous Activity
  • Interactions Synchronization

6
The Common Synchronization Measures
  • Correlation, Spectral Coherence, Mutual
    Information, Phase Synchronization, etc.
  • Widely applied
  • Limitations
  • They address bivariate measures
  • They are symmetric
  • Weak or non-synchronizing interactions may not be
    observable

7
An Alternative Approach
  • The cooperation among neuronal assemblies is
    simpliciter given by interactions that
  • not necessarily imply synchronization.
  • A new method to assess the functional
    interactions (connectivity) among multiple
  • measurements
  • Today at 1630, Room D
  • Session Tu4-D Nonlinear Time Series Analysis
  • Identifying Dependencies among Multivariate Time
    Series
  • Oscar DE FEO, Cristian CARMELI

8
Brief Description of the Method
  • Hp. 1. Brain as a network of weakly interacting
    dynamical oscillators
  • 2. A local oscillator behind each
    measurement
  • The method proceeds along two steps
  • Local self-model identification
  • Linear modeling of the interactions by
    cross-relation of all the measurements to the
    dynamics unjustified by the local models

TODAY 16.30 ROOM D
9
Proposal
  • Application of the method to high density EEG
    signals induced by visual stimuli

10
Experimental Framework
11
Previous Analyses
  • 70-71

    84-90

12
Result
71
84
84
71
70
90
90
70
71
70
13
Conclusions
  • The result emphasizes the complexity of the
    changes in the macroscopic manifestation of
    functional connectivity induced by very simple
    stimuli
  • First step toward the assessment of the
    macroscopic manifestation of functional
    cortico-cortical connectivity
  • The study on more subjects is needed
Write a Comment
User Comments (0)
About PowerShow.com