|
Dear Jakob,
You can put the attached file in /external/fileio/private to remove
that warning and slightly improve the handling of your dataset. The
warning has to do with the fact that your EEGLAB dataset does not
contain fiducials which SPM needs to do proper coregistration. It is
possible to use electrode locations as fiducials if you can find the
corresponding points at the structural scan. That's what the attached
code will do. However, if the electrode locations you had in EEGLAB
are just standard locations, you should just assign default sensor
locations using the 'Prepare' interface. That will make things
somewhat more straightforward. Also if you have individually measured
electrode locations you can load them to SPM using for instance .sfp
format (see the manual).
Best,
Vladimir
On Thu, Jul 2, 2009 at 11:07 AM, Jakob Scherer<[log in to unmask]> wrote:
> Dear Vladimir,
> the conversion runs successfully on my workstation too, now.
> btw i still get the following warnings. but i think it's a minor
> issue, i'll get back to it later
> -----------------------------------------------------------------------
> Running job #1
> -----------------------------------------------------------------------
> Running 'M/EEG Conversion'
>
> SPM8: spm_eeg_channelselection (v2866) 12:01:28 - 02/07/2009
> ========================================================================
> Warning: Could not obtain fiducials automatically.
>> In spm_eeg_convert at 410
> In spm_cfg_eeg_convert>eeg_convert at 231
> In matlabbatch\private\cfg_run_cm at 29
> In cfg_util>local_runcj at 1363
> In cfg_util at 699
> In cfg_ui>MenuFileRun_Callback at 1139
> In gui_mainfcn at 96
> In cfg_ui at 53
> checkmeeg: no channel type, assigning default
> checkmeeg: no units, assigning default
> checkmeeg: transform type missing, assigning default
> checkmeeg: data scale missing, assigning default
> checkmeeg: data type is missing or incorrect, assigning default
> creating layout from cfg.elec
> creating layout for ext1020 system
> Done 'M/EEG Conversion'
> Done
>
>
> thanks a lot,
> jakob
>
> On Wed, Jul 1, 2009 at 6:06 PM, Vladimir
> Litvak<[log in to unmask]> wrote:
>> Dear Jakob,
>>
>> I could reproduce the error. It comes down to Arno's bug in
>> read_eeglabheader. I fixed the bug and the conversions completes
>> successfully. Put the attached file in /external/fileio/private. Let
>> me know if you have any additional problems.
>>
>> Best,
>>
>> Vladimir
>>
>> On Wed, Jul 1, 2009 at 4:32 PM, Jakob Scherer<[log in to unmask]> wrote:
>>> sorry, i messed it up, it should read (read_eeglabdata.m, line 75ff)
>>>
>>> try
>>> pathstr = fileparts(filename);
>>> % assuming the .dat and .set files are located in the same directory
>>> fid = fopen([ pathstr filesep header.orig.data]);
>>> catch
>>> fid = fopen([ header.orig.filepath filesep header.orig.data ]);
>>> end
>>>
>>> Then the next error encountered is:
>>> -----------------------------------------------------------------------
>>> Running job #1
>>> -----------------------------------------------------------------------
>>> Running 'M/EEG Conversion'
>>>
>>> SPM8: spm_eeg_channelselection (v2866) 17:28:17 - 01/07/2009
>>> ========================================================================
>>> Warning: Could not obtain fiducials automatically.
>>>> In spm_eeg_convert at 410
>>> In spm_cfg_eeg_convert>eeg_convert at 231
>>> In matlabbatch\private\cfg_run_cm at 29
>>> In cfg_util>local_runcj at 1363
>>> In cfg_util at 699
>>> In cfg_ui>MenuFileRun_Callback at 1139
>>> In gui_mainfcn at 96
>>> In cfg_ui at 53
>>> checkmeeg: no channel type, assigning default
>>> checkmeeg: no units, assigning default
>>> checkmeeg: transform type missing, assigning default
>>> checkmeeg: data scale missing, assigning default
>>> checkmeeg: data type is missing or incorrect, assigning default
>>> creating layout from cfg.elec
>>> creating layout for ext1020 system
>>> Failed 'M/EEG Conversion'
>>> Index exceeds matrix dimensions.
>>> In file "P:\Science\Matlab\spm8\external\fieldtrip\private\prepare_layout.m"
>>> (???), function "prepare_layout" at line 608.
>>> In file "P:\Science\Matlab\spm8\external\fieldtrip\ft_prepare_layout.m"
>>> (???), function "ft_prepare_layout" at line 11.
>>> In file "P:\Science\Matlab\spm8\spm_eeg_project3D.m" (v3005), function
>>> "spm_eeg_project3D" at line 22.
>>> In file "P:\Science\Matlab\spm8\spm_eeg_prep.m" (v3005), function
>>> "spm_eeg_prep" at line 94.
>>> In file "P:\Science\Matlab\spm8\spm_eeg_convert.m" (v3088), function
>>> "spm_eeg_convert" at line 461.
>>> In file "P:\Science\Matlab\spm8\config\spm_cfg_eeg_convert.m" (v3059),
>>> function "eeg_convert" at line 231.
>>>
>>> The following modules did not run:
>>> Failed: M/EEG Conversion
>>> =======================================================================================
>>>
>>> i'll load up the dataset to reproduce the error
>>> thanks,
>>> jakob
>>>
>>>
>>> On Wed, Jul 1, 2009 at 5:13 PM, Jakob Scherer<[log in to unmask]> wrote:
>>>> dear vladimir,
>>>>
>>>> in read_eeglabdata.m, line 75:
>>>>
>>>> fid = fopen(header.orig.data);
>>>>
>>>> assumes that the files to be read in are located in the current matlab
>>>> directory.
>>>> one could change this into
>>>>
>>>> pathstr = fileparts(filename); % assuming the .dat and .set files are
>>>> located in the same directory
>>>> try
>>>> fid = fopen([ pathstr filesep header.orig.data]);
>>>> catch
>>>> fid = fopen([ header.orig.filepath filesep header.orig.data ]); %
>>>> which i personally find less intuitive, but more consistent with
>>>> eeglab
>>>> end
>>>>
>>>> to avoid this.
>>>>
>>>>
>>>> On Fri, Jun 26, 2009 at 7:08 PM, Vladimir
>>>> Litvak<[log in to unmask]> wrote:
>>>>> Dear Jakob,
>>>>>
>>>>> Could you send me the error message and the problematic dataset? You
>>>>> can use yousendit.com if it's too large.
>>>>>
>>>>> Best,
>>>>>
>>>>> Vladimir
>>>>>
>>>>> On Fri, Jun 19, 2009 at 3:08 PM, Jakob Scherer<[log in to unmask]> wrote:
>>>>>> hi,
>>>>>>
>>>>>> it seems to me that in
>>>>>>
>>>>>> spm_cfg_eeg_convert.m, line 37 should read:
>>>>>> usetrials.val = {1}; % instead of usetrials.val = {0},
>>>>>>
>>>>>> to match the default value defined in spm_eeg_convert.m
>>>>>> (from the help text:
>>>>>> % S.usetrials - 1 - take the trials as defined in the data [default]
>>>>>> % 0 - use trial definition file even though the data is
>>>>>> % already epoched)
>>>>>>
>>>>>>
>>>>>> Then, having corrected that, i get the following error:
>>>>>>
>>>>>> SPM8: spm_eeg_channelselection (v2866) 15:22:55 - 19/06/2009
>>>>>> ========================================================================
>>>>>> Failed 'M/EEG Conversion'
>>>>>> Invalid MEX-file
>>>>>> 'P:\Science\Matlab\spm8\@file_array\private\mat2file.mexw64': Invalid
>>>>>> access to memory location.
>>>>>>
>>>>>> .
>>>>>> In file "P:\Science\Matlab\spm8\@file_array\subsasgn.m" (v1355),
>>>>>> function "subfun" at line 132.
>>>>>> In file "P:\Science\Matlab\spm8\@file_array\subsasgn.m" (v1355),
>>>>>> function "subsasgn" at line 73.
>>>>>> In file "P:\Science\Matlab\spm8\spm_eeg_convert.m" (v3088), function
>>>>>> "spm_eeg_convert" at line 346.
>>>>>> In file "P:\Science\Matlab\spm8\config\spm_cfg_eeg_convert.m" (v3059),
>>>>>> function "eeg_convert" at line 231.
>>>>>>
>>>>>> The following modules did not run:
>>>>>> Failed: M/EEG Conversion
>>>>>>
>>>>>>
>>>>>> Thanks for any help,
>>>>>> regards, jakob
>>>>>>
>>>>>
>>>>
>>>
>>
>
function [sens] = read_sens(filename, varargin)
% READ_SENS read sensor positions from various manufacturer specific files.
% Currently supported are ASA, BESA, Polhemus and Matlab for EEG
% electrodes and CTF and Neuromag for MEG gradiometers.
%
% Use as
% grad = read_sens(filename, ...) % for gradiometers
% elec = read_sens(filename, ...) % for electrodes
%
% Additional options should be specified in key-value pairs and can be
% 'fileformat' string
%
% An electrode definition contain the following fields
% elec.pnt Nx3 matrix with carthesian (x,y,z) coordinates of each electrodes
% elec.label cell-array of length N with the label of each electrode
%
% A gradiometer definition generally consists of multiple coils per
% channel, e.g.two coils for a 1st order gradiometer in which the
% orientation of the coils is opposite. Each coil is described
% separately and a large "tra" matrix (can be sparse) has to be
% given that defines how the forward computed field is combined over
% the coils to generate the output of each channel. The gradiometer
% definition constsis of the following fields
% grad.pnt Mx3 matrix with the position of each coil
% grad.ori Mx3 matrix with the orientation of each coil
% grad.tra NxM matrix with the weight of each coil into each channel
% grad.label cell-array of length N with the label of each of the channels
%
% See also TRANSFORM_SENS, PREPARE_VOL_SENS, COMPUTE_LEADFIELD
% Copyright (C) 2005-2008, Robert Oostenveld
%
% $Log: read_sens.m,v $
% Revision 1.13 2009/06/03 09:52:15 roboos
% added zebris_sfp
%
% Revision 1.12 2009/02/02 16:10:15 vlalit
% Provide the 'headertype' argument to the internal read_header call.
%
% Revision 1.11 2009/01/23 10:32:55 vlalit
% New reader for Neuromag fif format using the MNE toolbox (http://www.nmr.mgh.harvard.edu/martinos/userInfo/data/sofMNE.php) implemented by Laurence Hunt.
%
% Revision 1.10 2008/09/18 10:38:33 vlalit
% Added 4D formats to be recognized by read_sens
%
% Revision 1.9 2008/05/22 14:33:18 vlalit
% Changes related to generalization of fiducials' handling in SPM.
%
% Revision 1.8 2008/04/14 20:51:36 roboos
% added convert_units
%
% Revision 1.7 2008/04/11 16:17:22 roboos
% added polhemus_fil
%
% Revision 1.6 2008/03/20 13:43:14 roboos
% added support for besa_pos
%
% Revision 1.5 2008/03/18 12:34:30 roboos
% fixed bug: added varargin to input arguments, thanks to Juan
%
% Revision 1.4 2008/03/06 09:27:54 roboos
% updated documentation
%
% Revision 1.3 2008/03/05 11:06:11 roboos
% test the presence of the fileio toolbox, needed when this function is included in forwinv
%
% Revision 1.2 2008/03/05 10:54:05 roboos
% added optional argument for fileformat
% some documentation changes
%
% Revision 1.1 2008/01/28 20:10:11 roboos
% new functions based on existing fieldtrip code
%
% test whether the file exists
if ~exist(filename)
error(sprintf('file ''%s'' does not exist', filename));
end
% get the options
fileformat = keyval('fileformat', varargin);
% determine the filetype
if isempty(fileformat)
fileformat = filetype(filename);
end
switch fileformat
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% read the content from various files that contain EEG electrode positions
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
case 'asa_elc'
sens = read_asa_elc(filename);
case 'polhemus_pos'
sens = read_brainvision_pos(filename);
case 'besa_pos'
tmp = importdata(filename);
if ~isnumeric(tmp)
error('unexpected file format for fileformat=besa_pos')
end
[nchan,nrow] = size(tmp);
if nrow==3
sens.pnt = tmp;
elseif nrow==9
pnt1 = tmp(:,1:3); % bottom coil
pnt2 = tmp(:,4:6); % top coil
ori = tmp(:,7:9); % orientation of bottom coil
sens.pnt = [pnt1; pnt2];
sens.ori = [ori; ori];
sens.tra = [eye(nchan) -eye(nchan)];
else
error('unexpected file format for fileformat=besa_pos')
end
[p, f, x] = fileparts(filename);
elpfile = fullfile(p, [f '.elp']);
elafile = fullfile(p, [f '.ela']);
if exist(elpfile, 'file')
warning(sprintf('reading channel labels from %s', elpfile));
% read the channel names from the accompanying ELP file
lbl = importdata(elpfile);
sens.label = strrep(lbl.textdata(:,2) ,'''', '');
elseif exist(elafile, 'file')
warning(sprintf('reading channel labels from %s', elafile));
% read the channel names from the accompanying ELA file
lbl = importdata(elafile);
lbl = strrep(lbl, 'MEG ', ''); % remove the channel type
lbl = strrep(lbl, 'EEG ', ''); % remove the channel type
sens.label = lbl;
else
% the file does not have channel labels in it
warning('creating fake channel names for besa_pos');
for i=1:nchan
sens.label{i} = sprintf('%03d', i);
end
end
case 'besa_sfp'
fid = fopen(filename);
tmp = textscan(fid, ' %[^ \t]%n%n%n');
fclose(fid);
sens.label = tmp{1};
sens.pnt = [tmp{2:4}];
case 'neuromag_mne'
hastoolbox('fileio');
hdr = read_header(filename,'headerformat','neuromag_mne');
sens = hdr.elec;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% gradiometer information is always stored in the header of the MEG dataset
% hence uses the standard fieldtrip/fileio read_header function
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
case {'ctf_ds', 'ctf_res4', 'neuromag_fif', '4d', '4d_pdf', '4d_m4d', '4d_xyz', 'yokogawa_ave', 'yokogawa_con', 'yokogawa_raw'}
% check the availability of the required low-level toolbox
% this is required because the read_sens function is also on itself included in the forwinv toolbox
hastoolbox('fileio');
hdr = read_header(filename, 'headerformat', fileformat);
sens = hdr.grad;
case 'neuromag_mne_grad'
hastoolbox('fileio');
hdr = read_header(filename,'headerformat','neuromag_mne');
sens = hdr.grad;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% This is for EEG formats where electrode positions can be stored with the data
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
case {'spmeeg_mat', 'eeglab_set'}
% check the availability of the required low-level toolbox
% this is required because the read_sens function is also on itself included in the forwinv toolbox
hastoolbox('fileio');
hdr = read_header(filename);
if isfield(hdr, 'grad')
sens = hdr.grad;
elseif isfield(hdr, 'elec')
sens = hdr.elec;
else
error('no electrodes or gradiometers found in the file')
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% these are created at the FIL in London with a polhemus tracker
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
case 'polhemus_fil'
[sens.fid, sens.pnt] = read_polhemus_fil(filename, 0);
% the file does not have channel labels in it
warning('no channel names in polhemus file, using numbers instead');
for i=1:size(sens.pnt, 1)
sens.label{i} = sprintf('%03d', i);
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% matlab files can contain either electrodes or gradiometers
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
case 'matlab'
matfile = filename; % this solves a problem with the matlab compiler v3
warning('off', 'MATLAB:load:variableNotFound');
tmp = load(matfile, 'elec', 'grad', 'sens', 'elc');
warning('on', 'MATLAB:load:variableNotFound');
if isfield(tmp, 'grad')
sens = getfield(tmp, 'grad');
elseif isfield(tmp, 'elec')
sens = getfield(tmp, 'elec');
elseif isfield(tmp, 'sens')
sens = getfield(tmp, 'sens');
elseif isfield(tmp, 'elc')
sens = getfield(tmp, 'elc');
else
error('no electrodes or gradiometers found in Matlab file');
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% these are created by a Zebris tracker, at CRC in Liege at least.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
case 'zebris_sfp'
[sens.fid, sens.pnt, sens.fid_label, sens.label] = read_zebris(filename, 0);
otherwise
error('unknown fileformat for electrodes or gradiometers');
end
if senstype(sens, 'eeg')
% only keep positions and labels in case of EEG electrodes
dum = sens;
sens = [];
sens.pnt = dum.pnt;
sens.label = dum.label;
end
% this will add the units to the sensor array
sens = convert_units(sens);
|
