DTF

The repetitions of the experiment can be used to improve the MVAR model fit. It is worth noting that the information from many repetitions is used during the model fitting stage, the data are not averaged. The correlations between input data channels are calculated within each realization and then the correlation matrix averaged over realizations is used to calculate MVAR model parameters. Using ensemble avaeraging much shorter data segments can be processed. By dividing the original data epoch into short segments and calculating DTFs for every segment the dynamics of transmissions can be presented in form of time-frequency maps. This technique is called Short-time DTF (SDTF) [2].