Library of examples using DANDI archive datasets ('Dandisets') authored as MATLAB® live scripts
The DANDI Archive hosts and catalogs openly licensed cellular neurophysiology datasets.
DANDI Archive datasets ('Dandisets') predominantly use the Neurodata without Borders (NWB) data standard. The NWB standard allows neuroscientists to exchange cellular neurophysiology data in a common format.
This library contains reproduced and/or reworked Dandiset analyses authored as MATLAB live script examples.
Reproduced and reworked analyses can serve to advance the quality and impact of original scientific datasets. Supporting and encouraging such 're-hacked' analyses is the goal of this library.
Accepted library examples will typically:
- explain the scientific context of the study producing the original Dandiset
- illustrate how to access and use a Dandiset
- achieve a reproduction and/or reworking of one or more findings (e.g., specific figures) from a Dandiset, potentially generating novel insights
To achieve these goals, the live script examples combine code, narrative text, images, and sometimes interactive controls.
To access the dataset, most examples use MatNWB software which is an open-source MATLAB toolbox for inspecting, reading, writing, and working with data encoded in NWB format.
The best way to find Dandisets of interest is to browse the DANDI archive web directory. Each Dandiset has a unique 6-digit Dandiset identifier.
Using Dandiset examples requires bringing together both:
- code - this example library on GitHub
- data - DANDI archive storage on an Amazon Web Services (AWS) S3 bucket
Users can elect convenience MATLAB environments which streamline all or some of these aspects, or to configure their own individual MATLAB environment for maximum flexibility.
Dandihub is a JupyterHub environment that combines these resources within a pre-configured virtual desktop that pre-mounts the example code* & Dandiset data as folders. MATLAB** and MatNWB software are pre-installed.
(*) coming soon
(**) using the MATLAB Integration for Jupyter
MATLAB Online is a web browser-based MATLAB environment with connectivity to GitHub code and AWS S3 data.
MATLAB Online uses MATLAB Drive as its cloud filesystem. Code can be cloned there from GitHub & data can be cached there from AWS.
Individual MATLAB environments on user machines or cluster nodes offer greater storage & flexibility. Utilizing remote data access functions for working with S3 data is recommended for the best possible performance.
This library collects together examples from diverse independent contributors.
Example authors can submit candidate example live scripts as GitHub pull requests for potential inclusion in the library. The library adminstrators will review the pull request.
Submitted example live scripts can be either hosted natively in this library or originated externally to this library (cloned/mirrored into the library).
See the Submitting Your Dandiset Example Live Script(s) wiki page for further information and guidelines.
The MatNWB Working Group chartered by the INCF organization serves as the administrators for this library.
- For individual examples, see the copyright information in their LICENSE files for attribution.
- The NWB data standard and software (incl. MatNWB) are supported by the NIH BRAIN Initiative® and the Kavli Foundation. See here for details on funded NWB project and here relevant publications.
- The DANDI data archive project is supported by the NIH BRAIN Initiative®
- DANDI archive AWS storage is available under the AWS Open Data Sponsorship Program.
- MATLAB is a registered trademark of MathWorks®.