Type: Software,

Keywords: Neurophysiology, Immunostaining imaging, Electrophysiology, Optophysiology, NWB, BIDS, Programmatic API, Streaming data

Resource ID: SCR_017571

Distributed Archives for Neurophysiology Data Integration

DANDI is the BRAIN Initiative supported data archive for publishing and sharing
neurophysiology data including electrophysiology, optophysiology, and
behavioral time-series, and images from immunostaining experiments. The archive
supports a broad range of users with different levels of expertise by providing
a spectrum from web-based to programmatic mechanisms to access and upload data
and helps improve the expertise through training of the scientific user base
through tutorials and workshops. All aspects of DANDI infrastructure are
developed with an adherence to open software, explicit licensing, and community

* Search, view, download, upload, and publish standardized neurophysiology data at any scale
* Create and collaborate with other researchers using open or embargoed (compliant with NIH policy) datasets and using compute near data
* DANDI’s API allows full programmatic access and enables integrating software directly with DANDI
* DANDI provides streaming access to parts of data using a combination of cloud technologies and storage formats, thus allowing for more scalable analysis software and visualization technologies.
* DANDI exposes all data as versioned DataLad datasets. This allows users to view an entire dataset without downloading any data in their local filesystem, and then selectively download specific files or folders
* Through DANDI Hub, a JupyterHub deployment next to the archive, DANDI provides an easy-to-use, interactive, parallel, graphical, and programmable data analysis and compute environment close to the large datasets.

* Find relevant neurophsyiology datasets
* Create, upload, and publish neurophysiology datasets in standardized format
* Collaborate using DANDI as a space for sharing public and embargoed data
(conforms to NIH data sharing policy).

* Provides standardized data with a search portal
* Computation near data in the cloud
* Full programmatic access via API and software libraries

* Metadata and data quality varies across datasets
* Some datasets are too large to download

* Data has to be in standardized formats using Neurodata Without Borders (NWB –
https://nwb.org) or Brain Imaging Data Structure (BIDS – https://bids-specification.readthedocs.io/en/stable/) as relevant.
* Data submitters convert, organize and validate data before upload.
* The DANDI command line interface requires a Python environment.


Satrajit Ghosh, Principal Research Scientist


Massachusetts Institute of Technology, Cambridge, MA



Yaroslav Halchenko, Research Associate Professor, Dartmouth College
Benjamin Dichter, Founder, Catalyst Neuro
Roni Choudhury, Kitware, Inc



* 5R24MH117295-04