Skip to content

ENH: Allow gradient compensated data in maxwell_filter#10554

Merged
larsoner merged 5 commits intomne-tools:mainfrom
larsoner:gradcomp
Apr 14, 2023
Merged

ENH: Allow gradient compensated data in maxwell_filter#10554
larsoner merged 5 commits intomne-tools:mainfrom
larsoner:gradcomp

Conversation

@larsoner
Copy link
Copy Markdown
Member

@larsoner larsoner commented Apr 20, 2022

I was interested to see if we could add support for maxwell_filter(raw_ctf.apply_gradient_compensation(3)). It looks like it's possible, but the performance is not as good as processing un-compensated data. Maybe I'm doing something wrong, though...

I'll leave this open and think about it for a bit.

@hoechenberger hoechenberger added this to the 1.2 milestone Jun 10, 2022
@larsoner larsoner modified the milestones: 1.2, 1.3 Sep 13, 2022
@larsoner larsoner removed this from the 1.3 milestone Dec 8, 2022
@larsoner
Copy link
Copy Markdown
Member Author

larsoner commented Apr 6, 2023

@britta-wstnr I hear you have been trying this :) Think it's working well enough to merge?

@larsoner larsoner added this to the 1.4 milestone Apr 6, 2023
@britta-wstnr
Copy link
Copy Markdown
Member

britta-wstnr commented Apr 6, 2023

Mhhh ... I agree with you that it's not making much of a difference whether 3rd order gradients are applied or not ... which is also puzzling to me (as it does make quite a difference on the pre-tSSS data). Let me investigate a little more next week 🕵️ only got to check it out on one of our datasets so far!

@larsoner larsoner removed this from the 1.4 milestone Apr 6, 2023
@britta-wstnr
Copy link
Copy Markdown
Member

britta-wstnr commented Apr 14, 2023

I extensively tested this on 50 CTF datasets with varying amounts of noise/contamination.

My findings:

  • the 3rd order synthetic gradient does not make much of a difference - probably a lot of the same noise gets caught by the tSSS as well?
  • a comparison between tSSS -> gradient compensation and gradient compensation -> tSSS (this PR) shows that the results are really similar (for most cases almost identical), except that what this PR implements seems to result in a little less noise overall.

@larsoner
Copy link
Copy Markdown
Member Author

larsoner commented Apr 14, 2023

Okay great, in it goes then!

@larsoner larsoner merged commit fc981bd into mne-tools:main Apr 14, 2023
@larsoner larsoner deleted the gradcomp branch April 14, 2023 14:59
larsoner added a commit to cbrnr/mne-python that referenced this pull request Apr 21, 2023
@larsoner
Merge remote-tracking branch 'upstream/main' into precommit
1770b7b 
* upstream/main: (50 commits)
  BUG: Fix bug with paths (mne-tools#11639)
  MAINT: Report download time and size (mne-tools#11635)
  MRG: Allow retrieval of channel names via make_1020_channel_selections() (mne-tools#11632)
  Fix index name in to_data_frame()'s docstring (mne-tools#11457)
  MAINT: Use VTK prerelease wheels in pre jobs (mne-tools#11629)
  ENH: Allow gradient compensated data in maxwell_filter (mne-tools#10554)
  make test compatible with future pandas (mne-tools#11625)
  Display SVG figures correctly in Report (mne-tools#11623)
  API: Port ieeg gui over to mne-gui-addons and add tfr gui example (mne-tools#11616)
  MAINT: Add token [ci skip] (mne-tools#11622)
  API: One cycle of backward compat (mne-tools#11621)
  MAINT: Use git rather than zipball (mne-tools#11620)
  ENH: Speed up code a bit (mne-tools#11614)
  [BUG, MRG] Don't modify info in place for transform points (mne-tools#11612)
  [BUG, MRG] Fix topomap extra plot generated, add util to check a range (mne-tools#11607)
  ENH: Add mne-bids-pipeline to mne sys_info (mne-tools#11606)
  MAINT: `coding: utf-8` is implicit in Python 3 (mne-tools#11599)
  ENH: Read eyetracking data (Eyelink) (Fork of mne-tools#10855 ) (mne-tools#11152)
  MAINT: In Python 3, do not prefix literals with `u` (mne-tools#11604)
  MAINT: object is an implicit base for all classes (mne-tools#11601)
  ...
larsoner added a commit to georgeoneill/mne-python that referenced this pull request Apr 21, 2023
@larsoner
Merge remote-tracking branch 'upstream/main' into hfc
c55101f 
* upstream/main:
  BUG: Fix bug with paths (mne-tools#11639)
  MAINT: Report download time and size (mne-tools#11635)
  MRG: Allow retrieval of channel names via make_1020_channel_selections() (mne-tools#11632)
  Fix index name in to_data_frame()'s docstring (mne-tools#11457)
  MAINT: Use VTK prerelease wheels in pre jobs (mne-tools#11629)
  ENH: Allow gradient compensated data in maxwell_filter (mne-tools#10554)
  make test compatible with future pandas (mne-tools#11625)
larsoner added a commit to larsoner/mne-python that referenced this pull request Apr 25, 2023
@larsoner
Merge remote-tracking branch 'upstream/main' into cola5
7c44d72 
* upstream/main: (152 commits)
  FIX: missing channels/fiducials can be np.nan (mne-tools#11634)
  use py3.10 in precommit config (mne-tools#11648)
  MAINT: Unify GH Actions pytest (mne-tools#11644)
  MRG: Rename "Discourse" link in top navigation to "Forum" [ci skip] (mne-tools#11649)
  ENH: Add support for Harmonic Field correction (mne-tools#11536)
  Add pre-commit (mne-tools#11541)
  BUG: Fix bug with paths (mne-tools#11639)
  MAINT: Report download time and size (mne-tools#11635)
  MRG: Allow retrieval of channel names via make_1020_channel_selections() (mne-tools#11632)
  Fix index name in to_data_frame()'s docstring (mne-tools#11457)
  MAINT: Use VTK prerelease wheels in pre jobs (mne-tools#11629)
  ENH: Allow gradient compensated data in maxwell_filter (mne-tools#10554)
  make test compatible with future pandas (mne-tools#11625)
  Display SVG figures correctly in Report (mne-tools#11623)
  API: Port ieeg gui over to mne-gui-addons and add tfr gui example (mne-tools#11616)
  MAINT: Add token [ci skip] (mne-tools#11622)
  API: One cycle of backward compat (mne-tools#11621)
  MAINT: Use git rather than zipball (mne-tools#11620)
  ENH: Speed up code a bit (mne-tools#11614)
  [BUG, MRG] Don't modify info in place for transform points (mne-tools#11612)
  ...
larsoner added a commit to larsoner/mne-python that referenced this pull request Apr 25, 2023
@larsoner
Merge remote-tracking branch 'upstream/main' into kernel
bbc55af 
* upstream/main: (117 commits)
  FIX: missing channels/fiducials can be np.nan (mne-tools#11634)
  use py3.10 in precommit config (mne-tools#11648)
  MAINT: Unify GH Actions pytest (mne-tools#11644)
  MRG: Rename "Discourse" link in top navigation to "Forum" [ci skip] (mne-tools#11649)
  ENH: Add support for Harmonic Field correction (mne-tools#11536)
  Add pre-commit (mne-tools#11541)
  BUG: Fix bug with paths (mne-tools#11639)
  MAINT: Report download time and size (mne-tools#11635)
  MRG: Allow retrieval of channel names via make_1020_channel_selections() (mne-tools#11632)
  Fix index name in to_data_frame()'s docstring (mne-tools#11457)
  MAINT: Use VTK prerelease wheels in pre jobs (mne-tools#11629)
  ENH: Allow gradient compensated data in maxwell_filter (mne-tools#10554)
  make test compatible with future pandas (mne-tools#11625)
  Display SVG figures correctly in Report (mne-tools#11623)
  API: Port ieeg gui over to mne-gui-addons and add tfr gui example (mne-tools#11616)
  MAINT: Add token [ci skip] (mne-tools#11622)
  API: One cycle of backward compat (mne-tools#11621)
  MAINT: Use git rather than zipball (mne-tools#11620)
  ENH: Speed up code a bit (mne-tools#11614)
  [BUG, MRG] Don't modify info in place for transform points (mne-tools#11612)
  ...
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@britta-wstnr britta-wstnr britta-wstnr approved these changes

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

3 participants

@larsoner @britta-wstnr @hoechenberger