Back to the main page.

Bug 3072 - Document (on the wiki) how balancing coefficients are accounted for in the forward computation

Reported 2016-02-19 01:35:00 +0100
Modified 2016-05-05 20:30:16 +0200
Product: FieldTrip
Component: documentation
Version: unspecified
Hardware: PC
Operating System: Windows
Importance: P5 normal
Assigned to: Jan-Mathijs Schoffelen
Depends on: 1445
See also:

Fran├žois Tadel - 2016-02-19 01:35:10 +0100

Add comments in ft_dipolefitting (or ft_datatype_sens?) to indicate how to pass the information about the 3rd order gradient compensation information to the leadfield computation (needs to be applied to the leadfield), when estimating dipoles.

Jan-Mathijs Schoffelen - 2016-02-19 09:12:34 +0100

In general, balancing is applied upon computation of the leadfield. This takes place in ft_compute_leadfield. The sensor-array structure contains a 'tra' field that contains balancing information. The balancing can pertain to higher order gradients, but also creates first-order gradiometers from the single coils (as in e.g. the ctf sensor array), or can do other fancy stuff, such as taking linear combinations of the magnetometer coils if spatial components have been removed from the data. Re. the 3d order gradients, the coefficients are represented in grad.balance, and are added to this structure upon reading the header information of the data. The function ctf2grad (or bti2grad) adds the system specific info to the gradiometer array. ft_denoise_synthetic allows for toggling the ctf-data to higher order gradiometer representation. Both the time series data, and the 'tra'-matrix in the gradiometer array are updated, to ensure consistent forward models, given the channel data.

Jan-Mathijs Schoffelen - 2016-02-19 10:48:22 +0100

Fran├žois Tadel - 2016-02-20 00:06:10 +0100

I added the support of the SSP and CTF compensation matrices in the Brainstorm function that builds the .elec and .grad fields. They are not correctly applied to the .tra field. Thanks