Bids-specification: [iEEG] DCOffsetCorrection does not apply to raw data

[robertoostenveld]

This is listed as "RECOMMENDED. A description of the method (if any) used to correct for a DC offset. If the method used was subtracting the mean value for each channel, use "mean"."

It is not specified for MEG, nor for EEG. I don't consider it appropriate for raw iEEG data, which is what the specification refers to. De-meaning is a procedure used in offline processing, not during recording. I believe this should be specified in the derivatives instead.

[robertoostenveld]

I propose to remove the field from the specification. Of course we should also check the validator.

@choldgraf @dorahermes @sappelhoff @ftadel, what is your opinion about this?

[dorahermes]

This was added after discussion with @voytek in the google doc because he records with Biosemi. I personally have never encountered this, but also have never specifically looked into it.
@choldgraf do you remember the details?

[robertoostenveld]

this relates to https://www.biosemi.com/faq/adjust_filter.htm and to https://www.biosemi.com/faq/adjust_filter_activeone.htm. I don't see why that would not be represented in the SoftwareFilters specification.

[dorahermes]

The DCOffsetCorrection was added before we detailed the Hardware/SoftwareFilter specifications, if it is redundant now it should be removed.