[Eeglablist] Cluster-based permutation tests

[Shingo Tokimoto]

Dear Dr. Mikołaj Magnuski,

Thank you for the reply. (Thank you, Makoto.) I have tried "dbstack if error" command. The messages in the command window are as follows.
-----
Reading all ERSP data:......
Reading all ERSP data:......
...
...
...
Reading all ERSP data:......
Reading all ERSP data:......
1 x 2, paired data, computing T values
the call to "ft_selectdata" took 1 seconds and required the additional allocation of an estimated 5 MB
-------------
The error messages in the window of EEGLAB error are as below.
--------
EEGLAB error in function ft_checkconfig() at line 148:

Error: ft_checkconfig (line 148)
The field cfg.neighbours is required

This is a problem with FIELDTRIP. The Fieldtrip version you downloaded is corrupted. Please manually replace Fieldtrip with an earlier version and/or email the Fieldtrip developers so they can fix the issue.
--------

I really appreciate your kindness. I am wondering the conversion of .set files of EEGLAB to FieldTrip is faster. Thank you in advance again.

************************************************
Shingo Tokimoto, Ph.D.
in Linguistics and Psychology
Department of Foreign Languages
Mejiro University
4-31-1, Naka-Ochiai, Shinjuku, Tokyo,
161-8539, Japan
tokimoto at mejiro.ac.jp
************************************************


On 2016/01/19, at 9:17, Mikołaj Magnuski wrote:

> I do not know how eeglab interfaces with fieldtrip to perform cluster-correction so my help is going to be limited:
> 
> In fieldtrip to perform cluster-correction you need to know which channels are considered direct neighbours - that is how activity can be clustered in channel space. This information is specified in cfg.neighbours. ft_prepare_neighbours and ft_neighbourplot can be helpful in defining this.
> You could try performing cluster-correction directly in fieldtrip - it's not difficult.
> 
> BTW - did you paste the full error stack? It would be easier to detect the problem if we knew where exactly it happens on the eeglab side. You can try turning on the breakpoints -> stop on errors option and once you catch the error type: "dbstack" to see the whole "nesting" of function calls that lead to the error.
> 
> 18 sty 2016 19:19 "Shingo Tokimoto" <tokimoto at mejiro.ac.jp> napisał(a):
> Dear EEGLA users,
> 
> I am trying to apply cluster-based permutation tests to my EEG data. I have found the relevant discussion on EEGLAB list, and I know the settings in "statistical parameters" in the STUDY function. With the setting as specified, cluster-based permutation tests work well for ERP but do not work for ERSP. The error message appears as below.
> 
> --------------
> EEGLAB error in function ft_checkconfig() at line 153:
> 
> Error: ft_checkconfig (line 153)
> The field cfg.neighbours is required
> 
> This is a problem with FIELDTRIP. The Fieldtrip version you downloaded is corrupted. Please manually replace Fieldtrip with an earlier version and/or email the Fieldtrip developers so they can fix the issue.
> ------------
> 
> I posted the question to Fieldtrip mailing list, but they say that the function for ERSP should work since it works well for ERP.
> 
> I am running EEGLAB (eeglab13_5_4b) on Matlab 2013b. I tried Fieldtrip-lite ver. 151020 first, and replaced it with fieldtrip-20160104. However, the same message appears. It is my pleasure if anyone tell me the proper "earlier version" of Fieldtrip. I appreciate any suggestions. Thank you in advance.
> 
> ************************************************
> Shingo Tokimoto, Ph.D.
> in Linguistics and Psychology
> Department of Foreign Languages
> Mejiro University
> 4-31-1, Naka-Ochiai, Shinjuku, Tokyo,
> 161-8539, Japan
> tokimoto at mejiro.ac.jp
> ************************************************
> 
> 
> _______________________________________________
> 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
>