You shouldn't be able to merge before epoching at all. I think SPM
allowed you to do that because your datasets were exactly identical in
length. But in principle you should only merge epoched data. Usually
in healthy subjects the head locations are quite consistent so people
at the FIL historically have ignored the differences. I had to develop
some machinery for my study with patients but eventually found that I
could get similar results no matter what I did about the head
locations. However, that might not be true for all cases.
Vladimir
On Tue, Oct 5, 2010 at 2:50 PM, Holly Rossiter <[log in to unmask]> wrote:
> Thanks Vladimir, I tried to merge before epoching so will now epoch
> everything and then merge and hopefully this will be fine. I am just using
> this to analyse CC's data but plan to use just one dataset for my study
> which will avoid the problem of the different head locations between
> datasets.
>
> Holly
>
> -----Original Message-----
> From: [log in to unmask] [mailto:[log in to unmask]] On Behalf
> Of Vladimir Litvak
> Sent: 05 October 2010 14:10
> To: Holly Rossiter
> Cc: [log in to unmask]
> Subject: Re: Merging MEG datasets
>
> Hi Holly,
>
>> On 05/10/10 10:26, Holly Rossiter wrote:
>>> I am trying to run analysis on work
>>> previously recorded at UCL and it is in 4 different datasets. I tried
>>> using the merge function in SPM but then I appear to lose all the
>>> markers in the data. Can you think why this might be happening?
>
> This is a known issue which comes down the the shortcoming of how the
> SPM format was originally defined. Since usually people only use
> events for epoching then after epoching when you merge the data it was
> safe to remove the events and no one has complained until now. But if
> there is some reason you want to retain them, let me know and I'll
> think of a workaround. In the long term I'm planning to slightly
> change the way the events are represented to fix this problem.
>
>>> Also
>>> what happens about the different positions of the head in the different
>>> datasets when they are merged together?
>>>
>
> The head location in the first dataset is used for the merged file.
> This is not ideal and there is a function in MEEGTools that you can
> use for more sophisticated things. This function is CTF-specific so
> for now there is no way to put it in main SPM code. Write me if you
> need more details.
>
> Best,
>
> Vladimir
>
>
>
|