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

List of potentially useful data sets #7

Open SylvainTakerkart opened 3 years ago

SylvainTakerkart commented 3 years ago

Let us list here the data sets that could be used for this standardization initiative...

SylvainTakerkart commented 3 years ago

from @satra in this post https://github.com/INCF/neuroscience-data-structure/issues/4#issuecomment-741804263:

you can look at the datasets on https://dandiarchive.org (there are about 38 dandisets covering 4 species, intra/extracellular and optical recordings)

SylvainTakerkart commented 3 years ago

from @jmbudd in this post: https://github.com/INCF/neuroscience-data-structure/issues/4#issuecomment-741459222

The Buzsaki lab, for example, has for many years shared their rodent hippocampal recording datasets many of which are available at crcns.org. In the process they have created a custom (session-based) data structure for their increasingly complicated combined ephys and behavioural experiments (see https://buzsakilab.com/wp/data-structure-and-format/).

SylvainTakerkart commented 3 years ago

from @SylvainTakerkart in this post: https://github.com/INCF/neuroscience-data-structure/issues/4#issuecomment-741832906

This dataset (utah array ephys, several sessions, several animals, exhaustive metadata) could also be useful: https://doi.org/10.1038/sdata.2018.55 https://doi.gin.g-node.org/10.12751/g-node.f83565/

tgbugs commented 3 years ago

A general pointer to the published sparc datasets, I'll pull out the stranger ones later https://sparc.science/data?type=dataset.

JuliaSprenger commented 3 years ago

from @SylvainTakerkart in this post: #4 (comment)

This dataset (utah array ephys, several sessions, several animals, exhaustive metadata) could also be useful: https://doi.org/10.1038/sdata.2018.55 https://doi.gin.g-node.org/10.12751/g-node.f83565/

There exists also an attempt to restructure this dataset according to AnDO specifications here: https://gin.g-node.org/sprenger/multielectrode_grasp/src/ando

chrisvdt commented 3 years ago

I would like to upload some empty datasets, but I can not push changes at the moment. For example I have an empty dataset for this study I would like to deposit here; Shape perception via a high-channel-count neuroprosthesis in monkey visual cortex. Chen X, Wang F, Fernandez E, Roelfsema PR. Science. 2020 Dec4;370(6521):1191-1196. doi: 10.1126/science.abd7435. PMID: 33273097.

Remi-Gau commented 3 years ago

I have told @SylvainTakerkart and @JuliaSprenger but I figured that I might as well put this here, but I am "stress testing" the BIDS BEP with an old dataset of mine that has both single cell recording and microscopy data:

https://gin.g-node.org/RemiGau/5HT_pain_ephys_raw

SylvainTakerkart commented 3 years ago

https://gin.g-node.org/RemiGau/5HT_pain_ephys_raw

Thanks @Remi-Gau ! But for info, I get a 404 error on this link...

Also, one question: is this a joint stress-test for BEP031 and BEP032?

Remi-Gau commented 3 years ago

https://gin.g-node.org/RemiGau/5HT_pain_ephys_raw

Thanks @Remi-Gau ! But for info, I get a 404 error on this link...

Had forgotten to make the repo public :facepalm:

Also, one question: is this a joint stress-test for BEP031 and BEP032? Yup it is: so some files already in there relate to the microscopy aspect.

If this is a problem for readability, let me know and I can try to "split things" while we iron the details of each BEP.

yarikoptic commented 3 years ago

from @satra in this post #4 (comment):

you can look at the datasets on https://dandiarchive.org (there are about 38 dandisets covering 4 species, intra/extracellular and optical recordings)

note that since then we have them all exposed also as DataLad datasets: https://github.com/dandi/dandisets

SylvainTakerkart commented 3 years ago

from @satra in this post #4 (comment):

you can look at the datasets on https://dandiarchive.org (there are about 38 dandisets covering 4 species, intra/extracellular and optical recordings)

note that since then we have them all exposed also as DataLad datasets: https://github.com/dandi/dandisets

yes, clearly! we'd need manpower to try to convert some of them (selected to be interesting from some point of view to be discussed) to BIDS... any idea @satra @yarikoptic ?