Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Adding information collected in minutes #13

Open
wants to merge 3 commits into
base: master
Choose a base branch
from
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
133 changes: 133 additions & 0 deletions actions_summary.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,133 @@
# Preface

This file tries to summarise what [the group](https://www.incf.org/sig/incf-working-group-standardized-data.) has been discussing so far.
We hope to facilitate onboarding with this document.

We exists officially since the end of 2020.
We have focused the work on bringing animal data (expecially neuro-, electro-physiology data) into BIDS format.
This is highly related to nwb/nix format and openminds electrophysiology expansion projects.

Julien Colomb initiated this document from reading the [minutes googledoc files](minutes-links.md).
Details about the group members were saved in the [contributors.md](contributors.md) file.

# BIDS consideration

- BIDS currently does not support linkage between data files obtained at the same time on different subjects. (level 1 is subjects, that would be one unique animal).
- [BEP for microscopy data](https://bids-specification.readthedocs.io/en/stable/04-modality-specific-files/10-microscopy.html) might be interesting to look at (see also [issue](https://github.com/bids-standard/bids-specification/issues/510), [PR](https://github.com/bids-standard/bids-specification/pull/881)). For instance the sample issue [was sorted](https://github.com/bids-standard/bids-specification/pull/816) (together with some of us).

# Set of use cases

### Example

- Dandi [asset level metadata:](https://api.dandiarchive.org/api/dandisets/000055/versions/draft/assets/ccdc3ae1-e98d-4a17-93d8-5b7386689fe0/)
- Dandi [example of samples](https://api.dandiarchive.org/api/dandisets/000008/versions/draft/assets/13eb9838-8a09-431e-8724-06a12fabb2d8/):
- Dandi [example of probes](https://api.dandiarchive.org/api/dandisets/000045/versions/draft/assets/b7855fa3-b6ee-4545-86b4-b4413095876b/):  
- Dandi [dataset level metadata](https://api.dandiarchive.org/api/dandisets/000008/versions/draft/)
- openminds [data](https://search.kg.ebrains.eu/instances/33a53fd9-d8ae-4313-9c28-b5921e5a0fee)set
- [human eeg example](https://www.humanbrainproject.eu/en/follow-hbp/news/localize-mi-an-open-source-dataset-of-simultaneous-intracerebral-stimulation-and-hd-eeg-in-humans/)

\

# BEP032: Electrophy data

The group has been discussing how to use the power of nwb and nix format, while having a BIDS format.

## Type of data

patch clamp data is included in bep032 (work done during brainhack 2021), `Microelectrode Recordings` ( BEP0206) may be a special case of electrophysiology ?
BEP 026 and 032 to be merge ?

## Files combinations

- BIDS requires different data files for different modalities (electrophy data, behavior recording, ..., meaning different data collection equipment) , while nwb/nix combine them.
- Several sessions/run will be included in nix/nwb, BIDS should get session specific files
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
- Several sessions/run will be included in nix/nwb, BIDS should get session specific files
- BIDS should accept nwb and nix files that contain multi-part recordings (e.g. multiple stimulation & recording snippets). Metadata for these will be handled in the `events.tsv` file. Alternatively, these individual recording snippets can also be split into multiple nix/nwb files and described as different `runs`.


Proposed restrictions for NWB / NIX files as they would be used in BIDS: - can not contain multiple modalities - can not contain data from multiple (BIDS) sessions - should be accompanied by one sidecar file (which can contain a hierarchical structure)

## Dealing with runs

One session may include different experimental "runs".

# metadata

## Decisions

- one set of metadata files for each run (redundant but simpler)
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
- one set of metadata files for each run (redundant but simpler)
- BEP032 proposes one set of metadata files for each run (redundant but simpler). Depending on the BIDS inheritance principle the number metadata files can be reduced for sessions with shared metadata.


## Schemas and ontologies

### age categories

<http://purl.obolibrary.org/obo/UBERON_0000105>

### Spatial position of probes

-  diversity of keys collected to be reduced
- issues
- differentiate between target location vs actual location

- consider location corrections or changes to be stored in different runs

- allow for uncertainties

- maybe consider start coordinate and end coordinate ?

- alignment methods are typically post-processes and should maybe be considered as derivatives

### Examples of schemas

Schemas and ontologies for dealing with probes, electrodes, nerve cuffs, etc. and the intersection between location (both semantic and spatial) and data acquisition and actuation/stimulation.

[GitHub dandi-cli issue related to issues with modelling probes](https://github.com/dandi/dandi-cli/issues/642#issuecomment-845623408)

[SPARC recommendations for physiology (work in progress)](https://docs.google.com/document/d/1-mADY7Rnev1jbtNBvlEX5nHNf8FD2YsNCTPGI-8k27w/edit#)

[BEP 32 animal physiology extension](https://docs.google.com/document/d/1oG-C8T-dWPqfVzL2W8HO3elWK8NIh2cOCPssRGv23n0/edit#heading=h.4k1noo90gelw)

<https://pynwb.readthedocs.io/en/stable/pynwb.ecephys.html>

<https://pynwb.readthedocs.io/en/stable/pynwb.icephys.html>

<https://github.com/HumanBrainProject/openMINDS_SANDS>

<https://github.com/HumanBrainProject/openMINDS_ephys> (work in progress)

<https://www.incf.org/sig/incf-working-group-electrophysiology-stimulation-ontology>

<https://www.incf.org/sig/incf-working-group-standardized-data>

# Relation and tools

- Openminds :links needed here: openminds for electrophysiology expansion.
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
- Openminds :links needed here: openminds for electrophysiology expansion.
- OpenMinds: [OpenMinds for electrophysiology](https://github.com/HumanBrainProject/openMINDS_ephys) expansion.


```{=html}
<!-- -->
```
- NWB: [best practice](ttps://www.nwb.org/best-practices/) -\> inspector (command line tool to check practices), different from a validation tool. See also [NeuroConv](http://neuroconv.readthedocs.io): library for converting data to NWB
- Conversion tool : <https://gin.g-node.org/RemiGau/example_ephys_bids_conversion>

```{=html}
<!-- -->
```
- other tools
- BEAVERDAM (build, explore, and visualize experiment databases of metadata; early version, in development), translate BIDS to \"database\"

- ALPACA (automated light-weight provenance tracking), provenance + ontology for analysis method

- <https://github.com/INCF/artem-is>

- [BEADL](https://beadl.org):  design, communicate, and disseminate behavioral task descriptions

\

##

## Current main TO DOs

1. BEP 032 finishing and merge

2. Keep on having these discussions to identify specific points that need to be harmonized between projects (e.g metadata in dandi and openminds) 

3. Gathering info about all conversion tools between the different formats into a single place could be a good practical contribution from our group!

4. Setting milestones on the development of tools to convert data into BIDS
65 changes: 65 additions & 0 deletions contributors.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,65 @@
From first meeting minutes:

Sylvain Takerkart (CNRS, Institut de Neurosciences de la Timone - INT, Marseille France)
I’m the head of the neuroinformatics core facility at INT, historically an fMRI person, an early BIDS believer and recently converted to dealing with organizing animal data at our 150-people institute…

Lyuba Zehl (Research Centre Jülich, Jülich, Germany | Human Brain Project / EBRAINS)
Together with Ida Aasebø I’m leading the task “Curation services, ontology engineering and services for access to data in protected storage” in the final financing phase of the HBP. Besides curating data for the HBP database, I’m focusing on ontology and in particular metadata schema engineering for general neuroscience (across modalities and species).

Thomas Brochier, (CNRS, Institut de Neurosciences de la Timone - INT, Marseille France), Head of the Mediterranean Primate Research Center and Head of the Cognitive Motor Control Team at INT. Neurophysiologist for more than 25 years, using chronically implanted multi-electrode arrays in macaque monkey.

Andrew Davison (CNRS, Institut des Neurosciences Paris-Saclay (NeuroPSI), France)
Leader of the Neuroinformatics research group at NeuroPSI, contributor to data curation task of HBP, with a focus on in-depth curation and data reuse. Co-maintainer of the Neo project. Interested in comparing experimental and simulation-derived data, mainly at the microscopic level (spikes, membrane potential, LFP). Previously a member of the INCF task force on electrophysiology data sharing.

Yaroslav Halchenko (AKA Yarik) (Dartmouth College, New Hampshire, USA)
A lead on elderly PyMVPA and NeuroDebian projects, and younger DataLad and DANDI archive projects. Participant and contributor to various “open neuroscience science” projects, BIDS included. More info could be found on http://centerforopenneuroscience.org

Mikkel Lepperød (CINPLA, department of biosciences, university of Oslo, Norway). Newly started postdoc, working with bio-inspired machine learning. During my PhD I was involved in both computational and experimental projects - mainly focusing on grid cells. To this project I can participate with experience from developing Exdir and Expipe - see Sec. 4c for more info.

Kévin Blaize (CNRS, Institut de Neurosciences de la Timone - INT, Marseille France).
Postdoc at the INT for 1.5 year, doing mainly optical imaging in V1 in awake macaca (VSDi and IOI). I did my PhD at the ‘institut de la vision’ (Paris, France) doing functional ultrasound imaging in visual areas in rodents and in non-human primates. I’m interested in data sharing so I would participate to this project as an experimentalist.

Emmanuel Procyk (Inserm,Stem Cell and Brain Research in Lyon, France). Team Leader Neurobiology of executive functions. Neurophysiologist working on frontal cortex anatomy, functions and dynamics in nonhuman primates. Work with ephys data from monkeys, humans and brain imaging data from monkeys and humans.

Charlie Wilson (Inserm,Stem Cell and Brain Research in Lyon, France). Researcher in Neurobiology of Executive Functions team. Neurophysiologist working in vivo in monkeys on frontal cortex function and dynamics. Work with longitudinal ephys data from monkeys, as well as some brain imaging data from monkeys and humans.

Samuel Garcia (CNRS, centre de recherche en neuroscience de Lyon). Leader of nothing.
I am in the core team (and sometimes alone) in following project : neo, spikeinterface, tridesclous, ephyviewer and pyacq.

Julian Budd. Computational neuroscientist interested in cortical structure and function. Recently working with and curating combined in vitro single cell anatomical and electrophysiological hippocampal datasets (Neo and Neuroshapes); in vitro electrophysiological recordings from isolated hippocampus (Elephant); and, analysing cortical axon putative connectivity in relation to in vivo intrinsic optical imaging feature maps.

Mikael Palner, preclinical PET and SPECT imaging scientist. Team leader and manager of a Milabs Vector4CT at Copenhagen University NMR center. Working on a neuroPET standard atlast for our preclinical tracers, as well as functional connectivity. PET, SPECT or MR we got rodent data to share. Also editing a special issue for Frontiers of Neuroscience Brain Imaging Methods, https://www.frontiersin.org/journals/neuroscience/sections/brain-imaging-methods, could maybe serve as a point of publication?

Satra Ghosh, a computer scientist and a computational neuroscientist by training, but these days we do a lot of plumbing for science: workflows (Nipype, Pydra), standards (BIDS, NIDM), software (neurodocker, nobrainer, testkraken, etc.,.) and infrastructure (neuroscout, dandi). Together with @Yaroslav Halchenko I am involved in several neuroinformatics projects. One such project is the DANDI archive (https://dandiarchive.org), a place to store, share, and collaborate on NWB-based cellular neurophysiology data. When not plumbing, we try to understand the neurobiology of human communication and use machine learning for precision medicine. For more projects that my group does see https://satra.cogitatum.org/group).
(shameless plug - co-editing this special issue on the neurocommons: https://www.springer.com/journal/12021/updates/17978796)

Julien Colomb, ex-neurobiologist, very bad, nerdy R developer and open FAIR data advocate. I have been interested in animal descriptors (especially transgenic animals, from nematodes to mammals) and ways to implement them for years (www.reagents.io) but never got to spend enough time on this. Data curator at the SFB1315, I finally got a position where working on animal description standards could be part of my day job. I love spreadsheets, because that is the only format which is really human writable.

Thomas Wachtler (LMU Munich, Department of Biology), computational neuroscientist and former neurophysiologist, Scientific Director of the German Neuroinformatics Node (G-Node) of the INCF. At the G-Node we work on tools and infrastructure for research data management including formats for Metadata collection and Data organization (odML, NIX) and github-like services for research data (gin.g-node.org).

Jeremy Garcia (CNRS, Institut de Neurosciences de la Timone - INT, Marseille France) doing a Master degree in IT / software engineering in co-op program with neuroinformatics core facility at INT , main contributor /developer of the AnDO Checker with Sylvain Takerkart, First time on an open source project .

Fred Barthelemy (Forschungszentrum Jülich, Germany / INT, Marseille France) Less and less neuroscientist / more and more data manager and developer of data management solutions. I currently use Neo and odML (and NIX soon) to handle data and metadata in the field of electrophysiology in NHP since 5 years, trying to be a bridge between the computer scientists and the experimentalists.

David Meunier (Institut de Neurosciences de la Timone - INT, Marseille France), Research Engineer AMU, recently developed Neuropycon (ephypype and graphpype) for workflows involving MEG and graph analysis and Macatools (macapype) for PNH MRI brain segmentation. Interested in BIDS, Docker, and Open source.

David Pucelle (Institut de Neurosciences de la Timone - INT, Marseille France) Engineer IT AMU, software engineering in neuroinformatics core facility at INT.

Chris van der Togt (Netherlands Institute for Neuroscience) Neuroscience researcher, Institute Data manager. I develop tools for ca imaging (Matlab, JAVA, C++), also a webapp in VUE-Cli (node/Javascript) for data management to expose metadata (FAIR) of research projects at the institute.

Robert Oostenveld (Donders Institute, Radboud, NL) head of FieldTrip MATLAB toolbox development; worked with Pascal Fries on monkey electrophys (acquisition and analysis) and cat electrophys (analysis), now mainly doing methods development for MEG, EEG and human iEEG; member of BIDS steering group and involved in many BEPs; architect for the Donders Research Data repository; Contributed to HCP young adult dataset.

Eric Burguière (CNRS, ICM, Paris) head of the NERB group (Neurophysiology of Repetitive Behaviors, https://icm-institute.org/en/team/team-burguiere/ ); we are using both mice and humans in a translational approach with behavior, ephys, opto/chemogenetic tools. We recently developed an innovative behavioral operant box completely automatized where animals are exposed for several weeks to a task, performing thousands of trials in a self-paced manner. We are continuously upgrading this setup by making it compatible with other open-access hardware such as open ephys, miniscope, deeplabcut technologies.

Michael Denker (Forschungszentrum Juelich, Germany) heads the team for Data Management in Electro- and Optophysiology in the Group for Statistical Neuroscience headed by Sonja Gruen. He is involved with the development of tools such as Elephant, Neo and NetworkUnit for the analysis of electrophysiological data and model validation in the context of the EBRAINS infrastructure of the Human Brain Project, and is involved with the development of data acquisition workflows for electrophysiological experiments.

Karim NDIAYE (PhD, IR-CNRS, Institut du Cerveau - ICM, Paris, France)
I’m the managing director of the PRISME facility at the ICM which is dedicated to human behavioral and cognitive studies, trained in mathematical physics & engineering then in cognitive neuroscience, I am also quite familiar with MEG-EEG (used to be a Brainstorm dev) and fMRI. Curious of how much BIDS for animal behavior could apply to human behavioral data (and vice-versa).

Ben Dichter (Neurodata Without Borders Community Liaison) I am the founder of CatalystNeuro, a company that facilitates data interoperability between neuroscience labs. We specialize in Neurodata Without Borders (NWB), and we are interested in connecting with other groups to converge on data organization more broadly.

Lizbeth Mondragon (Research Engineer, Institut du Cerveau – ICM, Paris, France)
I am a research engineer in the Neurophysiology of Repetitive Behaviors Lab (Head Eric Burguière). I have a background in robotics and IA and from my PhD I gained experience in computational neuroscience, in-vivo electrophysiology in freely moving rodents and real-time analysis of ephys data. I’m an enthusiast for new technologies in the neuroscience space and also support open-source tools and data sharing.

Ida Aasebø (Dept of Medicine, University of Oslo, Norway). Curation team leader EBRAINS/the Human Brain Project with Lyuba Zehl. The work from my PhD focused primarily on rodent electrophysiology and behavior, analytics and experiments. Multimodal and heterogeneous data clearly makes data sharing problematic in EBRAINS and other repositories, interested in hearing about and discussing alternatives.
Tom Gillespie (Department of Neurosciences, University of California, San Diego). I wear a number of different hats. Maintainer of/ontologist for the Neuroscience Information Framework Standard Ontology (NIFSTD/NIF-Ontology) and InterLex (successor to NeuroLex). Co-maintainer of the Research Resource Identifier (RRID) system. Principal developer for the automated curation pipelines for SPARC data structure. I participate in the development of the data models for SPARC, BICCN, and HBP. I became interested in neuroinformatics while I was still an experimentalist working in an electrophysiology lab and wanted to build a system to completely automate the data and provenance collection for my experiments that would cover everything from mouse breeding to workflow automation for patching and recording cells. My PhD work focuses on representation and communication of scientific protocols.
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
Tom Gillespie (Department of Neurosciences, University of California, San Diego). I wear a number of different hats. Maintainer of/ontologist for the Neuroscience Information Framework Standard Ontology (NIFSTD/NIF-Ontology) and InterLex (successor to NeuroLex). Co-maintainer of the Research Resource Identifier (RRID) system. Principal developer for the automated curation pipelines for SPARC data structure. I participate in the development of the data models for SPARC, BICCN, and HBP. I became interested in neuroinformatics while I was still an experimentalist working in an electrophysiology lab and wanted to build a system to completely automate the data and provenance collection for my experiments that would cover everything from mouse breeding to workflow automation for patching and recording cells. My PhD work focuses on representation and communication of scientific protocols.
Tom Gillespie (Department of Neurosciences, University of California, San Diego). I wear a number of different hats. Maintainer of/ontologist for the Neuroscience Information Framework Standard Ontology (NIFSTD/NIF-Ontology) and InterLex (successor to NeuroLex). Co-maintainer of the Research Resource Identifier (RRID) system. Principal developer for the automated curation pipelines for SPARC data structure. I participate in the development of the data models for SPARC, BICCN, and HBP. I became interested in neuroinformatics while I was still an experimentalist working in an electrophysiology lab and wanted to build a system to completely automate the data and provenance collection for my experiments that would cover everything from mouse breeding to workflow automation for patching and recording cells. My PhD work focuses on representation and communication of scientific protocols.

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
Tom Gillespie (Department of Neurosciences, University of California, San Diego). I wear a number of different hats. Maintainer of/ontologist for the Neuroscience Information Framework Standard Ontology (NIFSTD/NIF-Ontology) and InterLex (successor to NeuroLex). Co-maintainer of the Research Resource Identifier (RRID) system. Principal developer for the automated curation pipelines for SPARC data structure. I participate in the development of the data models for SPARC, BICCN, and HBP. I became interested in neuroinformatics while I was still an experimentalist working in an electrophysiology lab and wanted to build a system to completely automate the data and provenance collection for my experiments that would cover everything from mouse breeding to workflow automation for patching and recording cells. My PhD work focuses on representation and communication of scientific protocols.
Tom Gillespie (Department of Neurosciences, University of California, San Diego). I wear a number of different hats. Maintainer of/ontologist for the Neuroscience Information Framework Standard Ontology (NIFSTD/NIF-Ontology) and InterLex (successor to NeuroLex). Co-maintainer of the Research Resource Identifier (RRID) system. Principal developer for the automated curation pipelines for SPARC data structure. I participate in the development of the data models for SPARC, BICCN, and HBP. I became interested in neuroinformatics while I was still an experimentalist working in an electrophysiology lab and wanted to build a system to completely automate the data and provenance collection for my experiments that would cover everything from mouse breeding to workflow automation for patching and recording cells. My PhD work focuses on representation and communication of scientific protocols.
Julia Sprenger (Institut de Neurosciences de la Timone, Marseille, France). Research software engineer and main contributor to the BEP032 (so far). Background in physics, ephys data analysis, Python. Some experience in publishing datasets using Nix, odML and NWB.

15 changes: 15 additions & 0 deletions minutes-links.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,15 @@
## links to minutes files
- 2023-04: https://docs.google.com/document/d/1L2Jaboy6hhvp0kdYfqZ4XM8Zs2mbvwEs-OUDd9YqRsk/edit#heading=h.c69yh664n6mr
- 2023-01: https://docs.google.com/document/d/1LCsv8ADHjpcKekyOWN-GdK5UHLqfHU5zlfltF9GRy_E/edit#
- 2022-12:https://docs.google.com/document/d/1sMriiVU2krFXelqokvfm-pySPPZEN9JwUId9rdXYLdk/edit
- 2022-09: https://docs.google.com/document/d/1jwY6TQuqQS6JUxSlATlZGKwMit6PdMSI4T783cioWOc/edit
- 2022-06: https://docs.google.com/document/d/1wO3sHujMj7TLCdX_Gyl8CpIv1huJ1TEyFumsNJFj3mE/edit
- 2022-05: https://docs.google.com/document/d/1_HC4Yo7JkR9jhMSRCXhbyI19Cf4SVO5x0L6eZiaKTvg/edit
- 2022-04: https://docs.google.com/document/d/1U20ficIxX3Vo8sZmvkXX9UBePVf5LEpY5ss5U4WRGuM/view#heading=h.vg9nmr6q6cin
- 2022-03: https://docs.google.com/document/d/1PgBKGd6kh13vPoR7UK2Z1j0D1kddGXu16Qjnn29sQn4/edit
- 2022-02: https://docs.google.com/document/d/1XNxNnd3Mvk4wy9XyPEMJ589vwnVLCnQkWIA2cP2T7sI/edit#heading=h.vg9nmr6q6cin
- 2022-01: https://docs.google.com/document/d/102DKnNvS06ru8UoSzJtCUagkYv-y3u7HQwZ_POPDlq4
- 2021-10: https://docs.google.com/document/d/1O-Xun_2tDVjXw9BFH_w57L7jyeXPYqS0YxMubZMpIYk/edit
- 2021-07: https://docs.google.com/document/d/1cZroLI9ySQGMxw9k8lMKBMYVoE7IQqnUVc18KDEZ6bU/edit
- 2020-12: https://docs.google.com/document/d/1RsFNrqPhkHwrzz4QOJWirqvDNsrqRZDqfgOg6RAi4yk/edit
- 2020-06 https://docs.google.com/document/d/1NTDGBpdAIXVL5VI4Pyjzo-I783-OuCBzrhTFwDMJ5xU/view