[Eeglablist] Comparison of statistical and correction methods in Eeglab

[Michal Vavrecka]

Dear Arno,

thanks for useful comment and sorry for the delay.


On 27.9.2013 22:02, Arnaud Delorme wrote:
> Dear Michal,
>
> very nice.
>
> On thing you would do is to use the cluster correction algorithm on you 2-D matrix (time x channels), otherwise you p-value are not corrected for multiple comparison across time points.
I am not aware of cluster correction algorithms either in Eeglab or in 
general EEG processing. Can you provide me some info (Matlab function or 
code example) how to apply it to time x channel matrix?


>
> You might also want to fix the scale for the ERSP. You could also avoid plotting the contour lines for the values in topoplot, as this could be distracting during the animation. Finally, you could try to apply a 2-D convolution (conv2 function) to smooth values before rendering them and avoid sharp changes
Thanks for advices, I am going to incorporate them in the code.

> (bear in mind, the movie will be not statistically perfectly valid though).
The video is proposed to help to "look where and when" interesting 
things happens for further visualisations. What do you consider as "not 
perfect"? Are there problems with short time periods or there are some 
mathematical limitations of this method?

Thanks for your answer

Michal

> Arno
>
> On Sep 26, 2013, at 4:42 AM, Michal Vavrecka <vavrecka at fel.cvut.cz> wrote:
>
>> Dear Eeglab users,
>>
>> I wrote a short code that allows to visualize p-values for specific
>> statistical and correction methods in Eeglab. You can watch their power.
>> The data came from experiment based on presentation of affective stimuli.
>>
>> Video - http://youtu.be/vH1G4P1BGBI - (best in HD)
>>
>> Looking forward to read your comments
>>
>> Michal
>>
>> -- 
>> Michal Vavrecka
>> assistant professor
>> Biodat Research Group
>> Incognite Research Unit
>> FEE CTU
>> Karlovo nam. 13
>> Prague 2
>> phone: +420224357609
>> cell: +420608661977
>> personal: http://bio.felk.cvut.cz/~vavrecka/
>> groups: http://incognite.felk.cvut.cz/
>> http://bio.felk.cvut.cz/
>>
>>
>>
>> _______________________________________________
>> 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


-- 
Michal Vavrecka
assistant professor
Biodat Research Group
Incognite Research Unit
FEE CTU
Karlovo nam. 13
Prague 2
phone: +420224357609
cell: +420608661977
personal: http://bio.felk.cvut.cz/~vavrecka/
groups: http://incognite.felk.cvut.cz/
http://bio.felk.cvut.cz/