Hi,

Thanks a lot for this advice.

I have now been able to obtain a successful hpi fit both using the  
three good coils, and inputting the second digitisation with the  
recorded data from the first. In both cases this results in a much  
more sensible device-head transformation matrix. I think with this  
correct transformation matrix (close to identity for rotation and  
translation close to 0,0,40 (-7 0.5 51) that i can now trust the  
results of SSS and trans -default. Is this correct?

However, i still have a problem. When i try to perform maxmove  
compensation the hpi fits still fail. I guess from looking at the file  
that its because the incorrect initial digitised hpi information is  
still in there (though it tells me that its ignoring the incorrect  
coil information). So i'm still wondering if there is some way to copy  
the second set of isotrak points values into the raw fif file, such  
that maxfilter would then compute fits from them during maxmove?

Thanks again,

Marie

Quoting Dionyssios Theofilou <[log in to unmask]>:

> Hello Marie,
>
> Yes, it is possible to tell maxfilter to ignore the incorrect coil and
> work only with the other three. (In fact only three of the coils are  
>  always used).
> To tell the hpifit program to try a different selection of the   
> necessary coils
> you must use the –swap option. For example:
>
> … /dacq/bin/hpifit  -file  <my_troubled_file.fif>  -swap 1234
>
> will try to select first the electrode 1, then the 2, and at the end  
>  the electrode 4.
>
> … /dacq/bin/hpifit  -file  <my_troubled_file.fif>  -swap 3142
>
> will try to select first the electrode 3, then the 1, and at the end the 2.
>
> You will have to use also the show_fiff command, for example:
>
> … /bin/util/show_fiff  -v  -t  222  <my_hpi_coils.fif>
>
> The copy_trans_fiff command, for example:
>
> … /bin/util/copy_trans_fiff  -f  <my_hpi_coils.fif>  <my_test_coils.fif>
>
> And perhaps also the copy_pos_fif command.
>
> (Use the –help option with all these commands for more details).
>
> I hope that this has helped,
>
> My Best,
> Dionyssios.
>
>
>> Hi,
>>
>> I was wondering if anyone could suggest how to go about changing the
>> stored HPI coil positions from digitisation in a continuous fif data file?
>>
>> We recently recorded a data set in which the digitized coil location of
>> one coil is incorrect and this causes HPI fitting in maxfilter to fail
>> (the coil fits are fine, just the distances are too big 80-100mm). I've
>> tried maxfiltering with the -hpicons tag, and although this tried to
>> correct the deviant coil position, the HPI fits still fail. Luckily, we
>> re-digitised the positions of the coils after recording and would like to
>> use these correct coil location points in the original raw file. Does
>> anyone know if how to go about doing this? Or is there is any way to tell
>> maxfilter to ignore the one coil location thats incorrect and work with
>> the other three?
>>
>> Any advice and/or suggestions would be greatly appreciated.
>>
>> Thanks,
>>
>> Marie
>>
>>
>>
>> --
>>
>>
>



------------------------------------------------------------------
The University of Glasgow, Department of Psychology WebMail system
------------------------------------------------------------------