Wikimedia Deutschland/Open Science Fellows Program/Experiments BIDS

{| class="plainlinks" width="100%" height="100%" border=0 cellspacing=12 cellpadding=2 style="font-size: 95%; background-color:#dce4e8;"
 * colspan="2" style="text-align:center"|
 * valign="top" width="75%"|
 * valign="top" width="75%"|



= Open Science for Neuroimaging: Standardising Experiments for BIDS =

Project Description
A central goal of Open Science is to make experiments reproducable. Popular studies suggest that only about 10-40% of all Experiments can be reproduced. According to a survey with scientists the reasons for this reproducability crisis lie in poor laborganisation, missing documentation and the low number of replications of ones own experiments. These factors result in even fewer replications and re-analysis of experiments by other scientists. This is particularly worrisome, because a promising method to evade the reproducability crisis - meta-analyses - rely on standardized statistical methods.

For these reasons a number of neuroscientists have founded an initiative that wants to establish open standards for the data organisation in neuroimaging: The Brain Imaging Data Structure(BIDS). Because it is hard to change previous structures, a core challenge is to make BIDS as accessible as possible. In my project I want to improve on the tools that researchers are using in the neurosciences to create their datasets from the start. These 'open science' tools will target the barrier of entry into open data formats, because only very few laboratories have implemented the necessary structures yet.

In particular I am focusing on creating experimental code that complies with BIDS from the start. The idea is that scientists using a toolbox to create their experiment, e.g., psychtoolbox, psychopy etc., are given an example on how to create their BIDS-compliant output. Therefore I am creating example experiments myself and wish to collect examples from others for as many toolboxes as possible. In addition, we want to enable online experiment generators, e.g. lab.js.

Author/in

 * Name: Alexander von Lautz
 * Institution: Neurocomputation Neuroimaging Unit, Free University Berlin
 * Kontakt: alexander.von-lautz@bccn-berlin.de
 * Github: https://github.com/alexvonlautz/FFW_BIDS
 * }