Comparison of methods for estimation of time-varying transmission in multichannel data

The directed transfer function was estimated from the time-varying multivariate autoregressive model computed by means of Kalman filter. The performance of the method was tested by means of simulations. The Kalman filter method was compared with the sliding window approach used in estimation of Short-time Directed transfer Function. Both method gave similar results, however the application of Kalman filter method becomes problematic for bigger number of channels.