Abstract
Partial Directed Coherence (PDC) is a powerful tool to estimate a frequency domain description of Granger causality between multivariate time series. One of the main limitation of this estimator, however, has been so far the criteria used to assess the statistical significance, which have been obtained through surrogate data approach or arbitrarily imposed thresholds. The aim of this work is to test the performances of a validation approach based on the rigorous asymptotic distributions of PDC, recently proposed in literature. The performances of this method, defined in terms of percentages of false positives and false negatives, were evaluated by means of a simulation study taking into account factors like the Signal to Noise Ratio (SNR) and the amount of data available for the estimation and the use of different methods for the statistical corrections for multiple comparisons. Results of the Analysis Of Variance (ANOVA) performed on false positives and false negatives revealed a strong dependency of the performances from all the factors investigated. In particular, results indicate an amount of Type I errors below 7% for all conditions, while Type II errors are below 10% when the SNR is at least 1, the data length of at least 50 seconds and the appropriate correction for multiple comparisons is applied.
Original language | English |
---|---|
Title of host publication | Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS |
Pages | 5016-5019 |
Number of pages | 4 |
DOIs | |
Publication status | Published - 2011 |
Event | 33rd Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS 2011 - Boston, MA, United States Duration: Aug 30 2011 → Sep 3 2011 |
Other
Other | 33rd Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS 2011 |
---|---|
Country | United States |
City | Boston, MA |
Period | 8/30/11 → 9/3/11 |
ASJC Scopus subject areas
- Computer Vision and Pattern Recognition
- Signal Processing
- Biomedical Engineering
- Health Informatics