Bug 1247 - (Ronald) get information about electrode types in EEG structure
(Ronald)
: get information about electrode types in EEG structure
Status: UNCONFIRMED
Product: EEGLAB
General
: _other
: All All
: P3 enhancement (vote)
Assigned To: Bugzilla Manager
:
:
:
  Show dependency treegraph
 
Reported: 2012-03-26 23:49 PDT by Ronald
Modified: 2012-04-24 02:59 PDT (History)
1 user (show)

See Also:


Attachments


Note

You need to log in before you can comment on or make changes to this bug.


Comment 2 Ulrich Reinacher 2012-04-24 02:59:45 PDT
This would help streamline procedure a lot and help with refactoring existing
functions as well as writing functions aimed at multimodal analysis.
   Many emails on the eeglab list came up on workarounds for omitting selected
channels from filtering, baselining or other inappropriate processing. So far
pop_runica is the only function I know of that already checks for channel type.

Following that train of thoughts it would be good to specify physical units (to
differentiate between Voltage and e.g. Brightness, Temperature, Position, Heart
rate) and their significantly recorded digits and scaling.
   Units together with scaling could help make eegplot and other visualizations
easier to handle.
   Setting digits beyond recorded range in a DOUBLE to zero reduces file size
more than just converting to SINGLE. And of course it is good praxis to be
honest about recording accuracy.
Comment 1 Ronald 2012-03-26 23:57:02 PDT
Addressing different datatypes by index then becomes as easy as

>> ix = find( strcmpi( {EEG.chanlocs.type} , '%DESIRED_TYPE%'));
>> EEG_of_type = EEG.data( ix, :);

The curly brackets are needed to temporarily convert the structured output
EEG.chanlocs.type to a cell, allowing for strcmpi to return a logical array.
Description Ronald 2012-03-26 23:49:57 PDT
Many among us are moving toward multimodal signals, the combination of EEG and
EOG being a readily available extension. To easily address electrodes as per
type, it would be nice to have an additional obligatory field in the EEG
structure, namely

EEG.chanlocs(channel_index).type

containing a EEGlab homogenised character string 'EEG', 'EOG', 'ref', ... This
would then facilitate the search for electrodes in reference based filtering,
regression, active noise cancellation, etc. in which relations between
different "modalities" are sought.

Kind regards,

Ronald