[Eeglablist] newcrossf coherence matrix

[Arnaud Delorme]

> Thanks, this was helpful.  We’re still not quite clear on the time  
> bins though – is it the case that the timesout matrix gives the  
> midpoints of the time bins, and if so, why is it that the time bins  
> are not evenly spaced?  (They seem to be shifted by 8 or 12 ms.)


Yes, this is correct. The time bins are not perfectly evenly spaced  
because if you have 933 original time points (in your data epoch) and  
generate 200 time estimates for the purpose of time-frequency  
decomposition, you have to find the closest time point (in the  
original time array) for each evenly spaced time-frequency time point.

There is an option in newtimef though where you enter a negative value  
for the 'timesout' input and it simply subsample the original time  
space (-4 is considering 1 time point out of 4). This technique does  
ensure uniform time sampling.

Arno

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20090821/8b09e05a/attachment.html>