Data Sharing under FAIR principles

Introduction

FAIR principles are the 4 pillars of the open science movement.

They are an easy-to-remember way to enumerate several features a shared scientific product (a manuscript, a dataset, a piece of code, etc) should have.

They turn out to be really relevant when one has to deal with sharing datasets, because for this kind of scientific product, they fit in a particularly significant way.

The acronym stands mainly for:

  • F - Findable: A shared scientific product should include a persistent, maintained address where other people could find it (a DOI, etc);
  • A - Accessible: A shared scientific product could be retrieved using standard communication protocols (on a web page using http, etc);
  • I - Interoperable: Information included in a shared scientific product should be arranged using shared ways of data representation (Brain Image Data Structure for neuroimaging data, etc)
  • R - Reusable: A shared scientific product should include instructions on how to re-use the information contained in it (licenses, etc).

For further information, check:https://www.go-fair.org/fair-principles/

General indications

To properly share data online, check the following three steps:

  1. Pick an open, well-maintained, EU-based repository
    Examples: https://zenodo.org/, Open Science Framework - OSF (pick an EU-based repository)

  2. Choose which data to include

    1. Remove from your data all the information relative to Name, Surname, Date and Place of Birth (for neuroimages retrieved at CIMeC, this is already accomplished by the Retrieving System) and every information that could be related to participants personal data

    2. De-identify your data: if there is any specific physical information which might be used to identify the participants, remove it (e.g.: through defacing)

    3. Do not include structural data at first. You can share them on request through a Data Use Agreement

  3. Prepare your dataset arranging the information in a standard way
    Brain Imaging Data Structure is a standard way to arrange neuroimaging data which covers most of the common CIMeC settings (behavioral, MEG, EEG, MRI, physiological registrations, non - invasive brain stimulation, etc).

and proceed to publish.

CIMeC public repository

CIMeC maintains a public page on gin.g-node repository (a repository thought to be used by neuroscientists) whose goal is to gather the public dataset production in an easy to find and standardized way, featuring:

  • DOI request and maintaining (F-Findable)

  • Single file HTTP access and downloading (A-Accessible)

  • Datasets should be uploaded arranging the information using BIDS Standard (I-Interoperable)

  • Datasets should be uploaded including licensing information and (if any) Data Use Agreement instructions (R-Reusable)

Should you be interested in uploading your dataset on gin.g-node, please contact thinkopen.cimec@unitn.it specifying Data Sharing in the subject.