Back to the main page.

Bug 3253 - Integrate TSSS toolbox and SSS-beamformer in FT

Status NEW
Reported 2017-02-16 12:27:00 +0100
Modified 2017-11-14 10:05:51 +0100
Product: FieldTrip
Component: core
Version: unspecified
Hardware: PC
Operating System: Windows
Importance: P5 enhancement
Assigned to:
Depends on:
See also:

Vladimir Litvak - 2017-02-16 12:27:11 +0100

Elekta agreed to make the SSS prototype code developed by Samu Taulu freely available. We would like to integrate the use of this code in FT including the possibility of source analysis in SSS moment space. I made a short code snipped applying TSSS to a timelock struct and transforming it and grad to moment space. This code should be tested on phantom data at the first instance. Robert should decide what the most suitable function is for interfacing with the code. Vladimir

Robert Oostenveld - 2017-04-12 13:20:01 +0200

I suggest to call this ft_denoise_sss. That puts it besides the existing ft_denoise_pca and ft_denoise_synthetic functions (which are for 4D and CTF respectively). The low-level code shared by Samu would go into fieldtrip/external/sss (or similar). This involves extending ft_hastoolbox to detect the sss external dependency and add to to the path when needed. If there is need for further private code (only used in FT), that can go in fieldtrip/private.

Robert Oostenveld - 2017-04-12 13:21:53 +0200

I also suggest making a fieldtrip/test/test_ft_denoise_sss test function. It could start from Subject01.ds and demonstrate a simple pipeline.

Vladimir Litvak - 2017-04-12 13:27:31 +0200

Hi Sarang, Have you made any improvements to the FT function I sent you? That function might be a good starting point to work on. For now I don't have a version of Samu's code that I can put on open Github, but we all have a copy of it so we can go forward and develop ft_denoise_sss and once we sort out a public version of Samu's code we can release it as well. Happy Easter, Vladimir

Sarang Dalal - 2017-04-27 19:00:22 +0200

My apologies, I haven't had a chance to experiment much with it yet... Hopefully after MEG Nord!