[FieldTrip] fieldtrip question - help needed

[jan-mathijs schoffelen]

Cara Sara,

OK. Indeed, if you have run ft_rejectvisual and removed some channels,  
these are missing from your actual data. The original data.grad should  
however still contain the labels of all channels, and you can use  
grad.label as 'LABEL275', provided the order of the remaining channels  
did not change. Alternatively, you could define cfg.keepchannel =  
'nan', before calling ft_rejectvisual. The noisy channels will then  
still be 'present' in your data (but set to NaN). I believe that this  
would take away the need to do the first part of the script.

Best

JM

PS: please note that it is not optimal to do source reconstruction  
after replacing noisy channels with an average of their neighbours. At  
least you need to take this into account in the creation of the  
leadfields.


On Jan 14, 2011, at 11:03 AM, s.rombetto at cib.na.cnr.it wrote:

> Hello
> yes I'm referring to the example script 'fixing a missing sensor'.  
> Even if you consider the structure grad, you also load LABEL275.mat  
> and GRAD275.mat, that is the layout in someway.
>
> What I do is to process my data: I define trials and preprocess  
> them. Then I use ft_rejectvisual to remove artifacts and noisy  
> channels.
> After this my data are ok, but I have some missing channels.
> In order to solve the inverse problem, I'm using a software that  
> needs all the channels....and this is the reason why I'm trying to  
> "fix the missing sensor".
>
> Sara
>
>> Cara Sara,
>>
>> I am not sure whether I totally understand your question. Are you
>> referring to the example script 'fixing a missing sensor'?
>> When I look at it on our (renewed) wiki (still brings tears to my
>> eyes), I don't see any mention of a layout.
>>
>> I agree with you that the example however can be made a bit more  
>> clear.
>> This is my take on it, starting from the back of the pipeline:
>>
>> The 'fixing' algorithm replaces a 'dead' channel with the average
>> signal recorded at its neighbours. In the current implementation, the
>> description of the sensors and the data (=the grad-structure and the
>> data.trial) which go into ft_channelrepair need to contain a full
>> description of the sensor-array, including the faulty one. In other
>> words, if your sensor array has 10 sensors, one of which you want to
>> have repaired, the input to ft_channelrepair needs a data-structure
>> containing data from the spatial location of those 10 sensors (this  
>> is
>> in data.grad, and needed to determine which channels are the
>> neighbours). For consistency, the time course of the faulty channel
>> also needs to be there. The actual time courses of the faulty channel
>> will not be used, and these can take any value. Depending on the
>> hardware and recording protocol, a faulty channel is either present  
>> or
>> not in the data. In the example script, the faulty channel was not  
>> even
>> recorded, and this is the reason for the first part of the script:  
>> the
>> creation of a dummy channel, initialized with 0's in order to make
>> ft_channelrepair work. If, however, your dataset actually contains  
>> data
>> on the faulty channel (which may look very noisy), you can skip the
>> first part of the script, and directly go to the second part.
>>
>> Best wishes,
>>
>> Jan-Mathijs
>>
>>
>> On Jan 13, 2011, at 4:09 PM, s.rombetto at cib.na.cnr.it wrote:
>>
>>> Hello
>>> can anybody spend some more words on the chapter "Fixing a missing  
>>> sensor"?
>>> I'm trying to use it with a AtB (like ITAB-Chieti) device...but I
>>> cannot understand how to use it. I've several troubles when I've to
>>> declare the layout, since I don't have a layout file, but I  
>>> construct
>>> that starting from grad...
>>> Do you have any hint for me?
>>>
>>> Sara Rombetto
>>>
>>>
>>> -------------------------------------------
>>> * Dott.ssa Sara Rombetto                   *
>>> * Istituto di Cibernetica "E. Caianiello"  *
>>> * Via Campi Flegrei, 34                    *
>>> * 80078 Pozzuoli (NA)                      *
>>> * Italy                                    *
>>> * tel +390818675054                        *
>>> * fax +390818675326                        *
>>> -------------------------------------------
>>>
>>> ----------------------------------------------------------------
>>> This message was sent using IMP, the Internet Messaging Program.
>>>
>>>
>>> _______________________________________________
>>> fieldtrip mailing list
>>> fieldtrip at donders.ru.nl
>>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>>
>> Dr. J.M. (Jan-Mathijs) Schoffelen
>> Donders Institute for Brain, Cognition and Behaviour,
>> Centre for Cognitive Neuroimaging,
>> Radboud University Nijmegen, The Netherlands
>> J.Schoffelen at donders.ru.nl
>> Telephone: 0031-24-3614793
>>
>> _______________________________________________
>> fieldtrip mailing list
>> fieldtrip at donders.ru.nl
>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>
>
>
> -------------------------------------------
> * Dott.ssa Sara Rombetto                   *
> * Istituto di Cibernetica "E. Caianiello"  *
> * Via Campi Flegrei, 34                    *
> * 80078 Pozzuoli (NA)                      *
> * Italy                                    *
> * tel +390818675054                        *
> * fax +390818675326                        *
> -------------------------------------------
>
> ----------------------------------------------------------------
> This message was sent using IMP, the Internet Messaging Program.
>
>
>
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip

Dr. J.M. (Jan-Mathijs) Schoffelen
Donders Institute for Brain, Cognition and Behaviour,
Centre for Cognitive Neuroimaging,
Radboud University Nijmegen, The Netherlands
J.Schoffelen at donders.ru.nl
Telephone: 0031-24-3614793