Yes, SPM can average sensor positions but for that you need to convert
to SPM before averaging. You could also do the averaging in Fieldtrip
code using the same low-level function (ft_average_sens) but you also
need to read the fiducials and average those as well. You can find the
SPM code that does it in spm_eeg_merge and do something similar.
Best,
Vladimir
On 8 Feb 2012, at 10:22, Stenner, Max-Philipp wrote:
> Hi Vladimir,
>
> thanks very much.
>
> Since the sensor and fiducial positions are session-specific I
> suppose the conversion makes sense only at an early analysis stage,
> i.e. before timelocked/grandaveraged data across sessions is
> computed (or am I still misunderstanding something? How does SPM
> deal with variations in sensor position across sessions?)
>
> Sorry if these are silly questions, it's my first MEG analysis.
>
> Thanks again,
> best
> Max
>
> ________________________________________
> Von: Vladimir Litvak [[log in to unmask]]
> Gesendet: Dienstag, 7. Februar 2012 19:51
> Bis: Stenner, Max-Philipp
> Cc: [log in to unmask]
> Betreff: Fieldtrip to SPM
>
> Hi Max,
>
> All the information necessary for creating a head model in SPM cannot
> be extracted from a Fieldtrip structure because it's not there
> (particularly fiducials). So what you need to do is to use the
> function 'Copy MEG sensors' in MEEGTools toolbox (spm_eeg_copygrad) to
> read this information from on of the raw files and add it to your SPM
> dataset.
>
> Best,
> Vladimir
>
> On 7 Feb 2012, at 18:37, Stenner, Max-Philipp wrote:
>
>> Dear Vladimir,
>>
>> I am trying to convert ERF data from fieldtrip structures to spm-
>> compatible objects using spm_eeg_ft2spm for source reconstruction.
>>
>> Information on sensor position and fiducials is not stored in the
>> object that results from spm_eeg_ft2spm when using either
>> grandaverages or individual timelock-data gathered across several
>> sessions. It might be a silly question but is conversion to the spm
>> format hence only reasonable at earlier stages, i.e. when
|