[FieldTrip] Fwd: interpolation

[Eelke Spaak]

Dear Omer and Shlomit,

Data matrices (e.g. each element of a data.trial{} cell-array)
correspond to activity measured in *channels* over time. A channel in
a data matrix does not necessarily correspond to an electrode (or
gradiometer/magnetometer) which was present during the recording
session. For instance, an EEG dataset can be expressed as a bipolar
montage, which means that each channel corresponds to the difference
between two electrodes.

The positions of the physical electrodes are stored in sens.elecpos;
these don't change with referencing. The sens.tra matrix describes how
to convert from physical electrodes into data channels. If no
rereferencing is applied to the data, sens.tra will typically be the
identity matrix. In that case, sens.chanpos should also be identical
to sens.elecpos. For the case of e.g. a bipolar montage, sens.elecpos
will not change, while sens.chanpos might be updated to reflect points
halfway two electrodes (and sens.tra will then be different from
identity).

Having access to the physical sensors' positions and the sens.tra
matrix is needed for accurate source modelling, while sens.chanpos is
used in other cases such as determining neighbours or creating a
layout for topoplots.

Best,
Eelke

On 1 March 2015 at 11:38, shlomit beker <shlomitbeker at gmail.com> wrote:
>
> Hi Fieldtrippers,
>
> We would like to use the ft_channelrepair(cfg, data) function.
> However we don't understand how to load or convert the electrode positions
> into FieldTrip.
>
> We tried FT_READ_SENS and we looked in FT_DATATYPE_SENS, however there are
> two positions variables described in this structure:
>
> The structure for EEG or ECoG channels contains
>      sens.label    = Mx1 cell-array with channel labels
>      sens.chanpos  = Mx3 matrix with channel positions
>      sens.tra      = MxN matrix to combine electrodes into channels
>      sens.elecpos  = Nx3 matrix with electrode positions
>
> what do the third field mean? what is the difference between the second and
> the fourth?
>
>
> our format is the attached.
>
> Thanks a lot!
>
> Omer and Shlomit
>
>