We recently distributed a survey to evaluate how the neuroimaging community uses BIDS in their labs. We were interested in what data types BIDS was used for, an estimate of how many subjects’ data are organized in BIDS datasets, and the usage of BIDS Apps. The survey was primarily distributed on Twitter and the BIDS-standard specification GitHub. We received feedback from 116 researchers! The raw response data can be found on figshare. In this blog post, we will report some of the demographic information and usage results of this survey.

Table of Contents

1. Responder demographics
2. BIDS Usage
3. Discussion

1. Responder demographics

To begin to understand the demographics of our responders, we were interested in their current career stage. This would provide a lens for interpreting the results of the usage section. We found a near equal distribution between: PI/Group leader (32.8%), Postdocs (27.6%), and Graduate students (30.2%). This made up over 90% of our responses with the leftover covering research staff and undergraduates.

Responders were found to be from 72 unique universities/research centers with the top 3 concentrated at: University of Iowa (7 labs), UMC Utrecht (5 labs), and Temple University (4 labs). We show below the distribution of countries from around the world.

2. BIDS Usage

One main focus was in evaluating what data types researchers use BIDS for. Overwhelmingly, MRI (solely or in combination with other modalities) is the main modality that researchers are using BIDS to organize. About 15% of respondents reported using BIDS to organize MEG, EEG, or iEEG since their inclusion (April 2018 [MEG] and Feb 2019 [EEG and iEEG]). We expect the adoption of BIDS for EEG/MEG/iEEG data to increase as awareness spreads and more tools begin to support these datasets. We look forward to supporting and covering other modalities. Below we present the current distribution of data types BIDS is used for.   

We found approximately 65,605 subjects have been BIDSified! This is based on the estimation provided by the responder.
Our last main focus was inquiring the usage of our BIDS Apps (e.g., MRIQC, fMRIPrep) and BIDS-aware tooling (e.g., PyBIDS), and 80.9% of respondents indicated they use some programs or libraries that can automatically query BIDS datasets.

3. Discussion

This survey has provided us valuable insights into who in the community uses BIDS and how they use it. These are important key performance indicators in developing our plan for continuing to advance the adoption of BIDS in research pipelines. We found the responders came from a diverse background in regards to their career stage. Our respondents mostly were made up from the United States and Europe. A future direction to improve our geographic diversity outreach would be to extend beyond these regions and target groups, centers, conferences, and journals in underrepresented regions of the world.

Looking at the types of data organized in BIDS, we aim to further reach out to non-MRI researchers and identify impediments to adoption in those domains. We are happy to observe more 80% of our respondents were using BIDS Apps. A potential limitation of this survey that may bias our results is how it was distributed. To gather feedback from more researchers, more diverse modes of distribution can be utilized.

We would like to thank all the neuroimaging community for their time and valuable feedback!