[Eeglablist] erspboot & newtimef

[Arnaud Delorme]

Dear Rene,

yes, when comparing conditions, the statistical comparison is performed for each pixel.
Say if you have 200 time points and 100 frequencies, you will get a matrix of 200 x 100 x 2. As for the erspboot for frequencies (100 x 2), the first element in the last dimension contains the lower bound and the second element contains the upper bound.

If you are into hacking some functions, look up the applyboot subfunction of timecrossf. I believe it is applying the bootstrap mask for all possible configurations.

Note that the functionality of newtimef you are using is considered obsolete. It is better to build a STUDY with a single subject and 2 conditions, then compute the statistics using data trials. In this case, you can use many different methods for correcting for multiple comparisons. You can also compare more than 2 conditions or build a n x m design. 

Best,

Arno

ps: I think that correction for multiple comparison is not performed when comparing conditions in newtimef. It is only available when analyzing a single dataset.

On 30 Jan 2013, at 07:27, Rene Huster wrote:

> 
> Dear colleagues,
> 
> I struggle with the ERSP-masks when comparing conditions using newtimef(). 
> When comparing two conditions using the resampling options, erspboot provides me with the upper and lower boundaries for the time-frequency plots; these are the first two matrices in erspboot. 
> However, instead of a third matrix with lower/upper boundaries erspboot contains two full freq*time matrices. 
> 
> Unfortunately, I have no idea how to reconstruct the masked ERSP-image for the difference. 
> 
> Any help is much appreciated!
> 
> Best, 
> 
> 
> René
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to eeglablist-unsubscribe at sccn.ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to eeglablist-request at sccn.ucsd.edu