Full text
| |
| Author(s): |
Lucas Massaroppe
Total Authors: 1
|
| Document type: | Master's Dissertation |
| Press: | São Paulo. |
| Institution: | Universidade de São Paulo (USP). Escola Politécnica (EP/BC) |
| Defense date: | 2011-08-02 |
| Examining board members: |
Luiz Antonio Baccalá;
Birajara Soares Machado;
João Ricardo Sato
|
| Advisor: | Luiz Antonio Baccalá |
| Abstract | |
The purpose of this work is to present the development of methods for characterizing the connectivity between nonlinear neurophysiological time series. Methodologies from Information Theory Approximate and Sample Entropies are used to represent the complexity of the series in a period of time, which allows inferring on how its variability is transferred to other sequences, using partial directed coherence. Methods: For each system under consideration, (1) It is done a transformation in another, relating it to measures of entropy, (2) The connectivity is estimated by the use of partial directed coherence and (3) The robustness of the procedure is analyzed via Monte Carlo simulations and sensitivity analysis. Results: For the simulated examples, the proposed technique is able to offer plausible results, through the correct inference of the connectivity direction, in cases of nonlinear coupling (quadratic), with a reduced number of signals samples, where other approaches fail. Conclusion: The process proves to be an extension of the Granger causality to the nonlinear case. (AU) | |
| FAPESP's process: | 09/04397-7 - Brain connectivity characterization via Approximate Entropy and Granger causality |
| Grantee: | Lucas Massaroppe |
| Support Opportunities: | Scholarships in Brazil - Master |