Comments on: Tues Apr 26

By: David Zhou

David Zhou — Thu, 26 Apr 2012 14:07:02 +0000

Can you talk about Granger causality? Might this be possible to use this for EEG channels?

By: Rich Truncellito

Rich Truncellito — Thu, 26 Apr 2012 13:30:09 +0000

Do the rules governing bivariate time series data extend to multivariate cases?

By: Thomas Kraynak

Thomas Kraynak — Thu, 26 Apr 2012 13:29:12 +0000

If any, what are the differences to take in consideration between standard bootstrap procedures and bootstrapping the smoothed periodogram?

By: Scott Kennedy

Scott Kennedy — Thu, 26 Apr 2012 13:15:31 +0000

I usually read about spike-triggered averages used to identify the functional relationship between a motor cortical neuron and the EMG of a muscle. Could Granger causality also be used in this situation to show if the neuron is causally related to the EMG?

By: Rob Rasmussen

Rob Rasmussen — Thu, 26 Apr 2012 13:07:49 +0000

What would be a good statistical method to analyze the location of peaks in the periodograms? It seems like in figure 18.15 that the peaks can be relatively hidden within the 95% confidence bands.

By: Shubham Debnath

Shubham Debnath — Thu, 26 Apr 2012 05:28:53 +0000

Is there any particular smoothing method for periodograms that would be better for interpreting coherence?

By: Jay Scott

Jay Scott — Thu, 26 Apr 2012 05:22:53 +0000

Can you explain the bar notation in formula 18.51 on p554?

By: Sharlene Flesher

Sharlene Flesher — Thu, 26 Apr 2012 03:05:41 +0000

When wouldn’t you be able to use the periodogram to propagate uncertainty, and have to use the methods discussed in 18.4.2? Why wouldn’t you be able to use it, and the uncertainty obtained from it?

By: Noah

Noah — Thu, 26 Apr 2012 02:22:59 +0000

Is the propagation of uncertainty more complicated in bi-variate series?

By: Amanda Markey

Amanda Markey — Wed, 25 Apr 2012 17:06:40 +0000

I’m not sure I entirely understand what the significance tests based on the integrated squared difference between two smooth curves is really testing.
Suppose we had two curves that were exactly the same shape, but one had been translated x units up. It seems that this significance test would miss the similarity. Is there a way to test for whether the shapes are significantly different?