search

INCF / neuroscience-data-structure

Space for discussion of a standardized structure (directory layout + metadata) for experimental data in systems neuroscience, similar to the idea of BIDS in neuroimaging
21 stars 3 forks source link

Question about microscopy data types and modalities #18

Open JoeZiminski opened 11 months ago

JoeZiminski commented 11 months ago

Hi Everyone,

We have a question about microscopy data, apologies if this is the wrong place for this discussion (if so it would be much appreciated if someone could point us to the correct place).

We are currently thinking about how to store microscopy data of histological type (i.e. post-mortem) and what we call 'functional' imaging, such as 2-photon with calcium indicators. Currently we have the data-type names histology and funcimg although we would like to move histology to micr to align with BEP031. However, in this case it is not entirely clear what data should go under micr, from reading the specification, any 2-photon data should be stored in micr. However, it seems unintuitive to mix both functional imaging and post-mortem histology within the same data-type. Similarly this would not reflect the distinction made in neuroimaging BIDS, that has anat and func to distinguish between anatomical and functional imaging.

Does anyone have any further insight as to this question? We are considering using micr for histology and funcimg for functional microscopy. However, this naming is somewhat confusing (as funcimg is a form of microscopy), and we do not want to break BIDS compatibility if indeed micr is intended to contain functional microscopy.

Thanks! Joe (from the SWC Neuroinformatics Unit working on NeuroBlueprint)

SylvainTakerkart commented 11 months ago

Hi Joe,

I think your question is really related to the BIDS specifications themselves, and therefore, I would re-post it there: https://github.com/bids-standard/bids-specification/issues (and I'll answer there with what I know ;) )

Cheers,

Sylvain

yarikoptic commented 11 months ago

FWIW, an example of BIDS dataset with micr stained data https://dandiarchive.org/dandiset/000108/draft/files?location=sub-MITU01%2Fses-20210521h17m17s06%2Fmicr it has only post-mortem/no func though. Indeed feels odd to place mix functional and anatomical data within the same micr/ - for MRI indeed there is a clear separation but I don't think it was envisioned for micr. I guess (?) it would be a different session though and likely a different suffix (2PE vs SPIM)? But indeed in any case sounds more of a generic BIDS question, not for this BEP

SylvainTakerkart commented 11 months ago

FWIK, there was no functional data around when the BEP microscopy was put together and when it was integrated officially in the BIDS specs... so it's lacking, and @JoeZiminski , your post reflects this... the official specs for microscopy are here now: https://bids-specification.readthedocs.io/en/stable/04-modality-specific-files/10-microscopy.html, and correspond to what @yarikoptic explains: different suffixes are used for func (e.g 2PE) and anat (e.g CONF) data types; but indeed, this seperation between func and anat (using suffixes) is not consistent with how it's done for MRI (with different "BIDS modalities" sitting in different directories)...

Remi-Gau commented 11 months ago

@JoeZiminski this indeed an inconsistency between modalities

my suggestion to help keep them separate would be to use different session folders

ses-histology/micr/sub-*png
ses-funcimg/micr/sub-*2PE.png