Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    Mainak Jas
    @jasmainak
    does anyone else see a problem in the formatting of artifacts on Circle: https://1924-89170358-gh.circle-artifacts.com/0/html/index.html ?
    Alexandre Gramfort
    @agramfort
    I've seen this pb for another repo lately
    it's a circleci issue i think
    Stefan Appelhoff
    @sappelhoff
    are there gsoc 2020 students for MNE? @agramfort
    Alexandre Gramfort
    @agramfort
    no we screwed up the application deadline this year :(
    Stefan Appelhoff
    @sappelhoff
    damn =(
    Dominik Welke
    @dominikwelke
    hey. bids-derivatives standards are finalized atm (bids-standard/bids-specification#265).. are there plans to support saving derivative data using mne-bids? shouldn't be too difficult to add this functionality using the existing tools..
    Mainak Jas
    @jasmainak
    is there something relevant to ephys?
    Dominik Welke
    @dominikwelke
    you mean, whether there is smth specifically on ephys data in the proposal i linked? i didnt see anything, but the general principles are simple and cover ephys: its basically a specific location in the bids folder tree (various options), plus a required new entry "GeneratedBy" and recommended link to "SourceDataset" in dataset_description.json for the derived dataset
    Stefan Appelhoff
    @sappelhoff

    These are the "common derivatives" which lay out the foundation for other derivatives to build on. That being said, there is a lot of MRI stuff on top though.

    I think we can wait with the implementation of this within MNE-BIDS at least until the PR is merged and a new release of BIDS is out (~2 to 3 weeks).

    I'd prefer to wait a bit longer and see how the situation evolves ... but if it's important to you @dominikwelke I'd be happy to review a PR :)

    Dominik Welke
    @dominikwelke

    yes, i'd also wait at least until the new BIDS is out.

    it's not particularly important to me, atm. i just think it would be a nice and easy to implement feature, that could also simplify interaction with e.g. EEGLAB that has some standardized preprocessing pipelines.
    i might end up writing something like this anyway, if so I'lld release a PR :)

    Adam Li
    @adam2392

    does anyone here work w/ iEEG data and electrode localization that wants to work on simplifying the error-prone and tedious process?

    Was looking for some collaborators w/ open-source experience to help kick off a few ideas I had after OHBM. Hoping to migrate a chat over email and video if someone is...

    2 replies
    Stefan Appelhoff
    @sappelhoff

    question re:bids on brainhack mattermost:

    https://mattermost.brainhack.org/brainhack/pl/tfs4d3iqe7fr3qjasddyd6f86e

    @agramfort I'm trying to bring ds248 in bids-examples in line with the spec. Some of the channels types are listed as MEGGRAD this needs to be MEGGRADPLANAR or MEGGRADAXIAL do you know off the top of your head which it should be?

    asked by @rwblair
    Mainak Jas
    @jasmainak
    MEGGRADPLANAR
    Alexandre Gramfort
    @agramfort
    +1
    Alexandre Gramfort
    @agramfort
    @sappelhoff @jasmainak how would you name evoked data according to BIDS convention (anticipating derivatives): /bids_root/sub-01/ses-01/meg/sub-01_ses-01_evoked.fif hence extending kind or rather /bids_root/sub-01/ses-01/meg/sub-01_ses-01_proc-evoked_meg.fif
    cc @hoechenberger
    Richard Höchenberger
    @hoechenberger

    @sappelhoff @jasmainak how would you name evoked data according to BIDS convention (anticipating derivatives): /bids_root/sub-01/ses-01/meg/sub-01_ses-01_evoked.fif hence extending kind or rather /bids_root/sub-01/ses-01/meg/sub-01_ses-01_proc-evoked_meg.fif

    +1 for proc-evoked_meg.fif

    might be good to have your inputs/thoughts there. It was made up in one or two afternoons :)
    Alexandre Gramfort
    @agramfort
    I see that the _proc- entity is not used at all. Is it becomes it's meant for otherthings?
    Mainak Jas
    @jasmainak
    Yes "proc" actually means "processed on device" according to the entity table here: https://bids-specification.readthedocs.io/en/stable/99-appendices/04-entity-table.html
    Alexandre Gramfort
    @agramfort
    can you give an example of proc usage?
    Mainak Jas
    @jasmainak
    I think SSS files: https://bids-specification.readthedocs.io/en/stable/99-appendices/06-meg-file-formats.html#neuromagelektamegin. I know you'll say it's a derivative and I don't have an answer for that ... proc was kind of a hack to allow sharing both raw and SSS files when designing MEG-BIDS
    Richard Höchenberger
    @hoechenberger

    I followed https://bids-specification.readthedocs.io/en/stable/04-modality-specific-files/02-magnetoencephalography.html, which says:

    "The proc label is analogous to rec for MR and denotes a variant of a file that was a result of particular processing performed on the device. This is useful for files produced in particular by Elekta’s MaxFilter (e.g. sss, tsss, trans, quat, mc, etc.), which some installations impose to be run on raw data because of active shielding software corrections before the MEG data can actually be exploited."

    I understand that it was assumed that SSS is performed on the recording device?
    I would love to abuse proc for all kinds of derivative processing. SSS is producing a derivative to me :)
    Mainak Jas
    @jasmainak
    I think what is a derivative is driven more by usecase rather than a strict definition. In the case of SSS, almost everyone (until a couple of years ago) used Maxfilter with probably default parameters. And you wouldn't analyze the data until you Maxfiltered it. So you could consider it "on device" for practical purposes
    That being said, I agree there is conflicts in how this is defined in the standard. You might have already seen the related issue that robert raised on the bids specification github page?
    Alexandre Gramfort
    @agramfort
    which github page?
    Mainak Jas
    @jasmainak
    Alexandre Gramfort
    @agramfort
    thx ! where do you stand now?
    Mainak Jas
    @jasmainak
    no strong feelings but I think it's best not to guess what the derivatives will be like. Unless you want to get more heavily involved in that effort?
    Alexandre Gramfort
    @agramfort
    @jasmainak I need to be an end user to have opinions. I am learning a lot by using the study template on multiple datasets now
    Adam Li
    @adam2392
    What’s the likelihood 0.5 is released sometime soon? I want to transition my code to use some of the latest simplifications and improvements :)
    Mainak Jas
    @jasmainak
    maybe @sappelhoff or @hoechenberger can help answer that question
    Stefan Appelhoff
    @sappelhoff

    I think it makes sense to release mne-bids shortly after MNE-Python (which is scheduled for September 15)

    --> then we have some short time to make sure mne bids 0.5 works well with the latest MNE release (0.21) ... and perhaps even one or two releases prior to that.

    And then for 0.6.dev we can (if necessary at all) adopt MNE master features again

    That said, I think it'd be a good idea to start tagging issues and PRs that we want to have finished prior to 0.5 release.

    @hoechenberger do you want to make a release this time around? I wrote pretty extensive docs on what to consider, so it should be fairly easy :)

    Alexandre Gramfort
    @agramfort
    yes let's try to release in september !
    Adam Li
    @adam2392
    sounds good! thanks for the info :)
    eort
    @eort
    Hey, when I tried to use write_raw_bids the procedure failed due to some issues with the events in my file (ValueError: You have 1 events shorter than the shortest_event.). I know this has nothing to do with mne-bids and should not happen in the first place. In mne-python I can fix this issue and work with the raw file alright, but it seems there is currently no way to fix/bypass the issue with mne-bids. So, my question would be, whether it would make sense/was possible, to increase the flexibility of write_raw_bids somehow, so that users have the opportunity to do something about issues inside their raw files? If not, I would have to temporarily change my local installation of mne-bids to allow this specific file to be read.
    Mainak Jas
    @jasmainak
    You can actually provide the events using an argument called "events_data". Do what's needed in mne-python to bypass the error and then provide the extracted events to mne-bids.
    1 reply
    Clemens Brunner
    @cbrnr
    Hi all! I just followed along the EEG example (https://mne.tools/mne-bids/dev/auto_examples/convert_eeg_to_bids.html#sphx-glr-auto-examples-convert-eeg-to-bids-py) and was puzzled why the annotations were read in separately in v0.4. Of course this has already been fixed in dev (annotations are now automatically loaded by MNE functions), so I wanted to ask if maybe v0.5 could be released soon or if there are still some other things to do before the new release.
    Stefan Appelhoff
    @sappelhoff
    @hoechenberger is doing the release this time, and to my knowledge, he's on it :-) see mne-tools/mne-bids#573
    Clemens Brunner
    @cbrnr
    OK nice, thanks!
    timonmerk
    @timonmerk
    Hey everyone! First, thanks for keeping up providing support for such a great tool! It helped our lab a lot already transferring the data in BIDS format!
    I have a question regarding writing the used reference in hindsight using write_raw_bids.
    In the tutorial https://mne.tools/mne-bids/stable/auto_examples/convert_mne_sample.html#sphx-glr-auto-examples-convert-mne-sample-py it get's explained that it needs to be added afterwards. Is there a way around that? Especially when working with ECOG grid with many channels it would be a useful functionality to just provide it as a keyword
    Mainak Jas
    @jasmainak
    I would support a function to update the sidecars similar to here: mne-tools/mne-bids#458. But final decision is up to @agramfort and @hoechenberger since they maintain the repo these days
    Alexandre Gramfort
    @agramfort
    @timonmerk tell me more. MNE-Python stores some information about the reference applied. So it means it should be possible to write it. I don't know the iEEG specificities
    timonmerk
    @timonmerk
    @agramfort Currently I am using the ´write_raw_bids´ function, and have to add a reference column afterwards for every "run" .tsv file where I hard code the used reference, this would be very useful to add rather as a keyword to the ´write_raw_bids´ function, or specify in a sidecar file?
    I am not sure also if I miss a way if it could be done otherwise, so I asked here :)
    Alexandre Gramfort
    @agramfort
    can you share with me one file you had to annotate manually?