Read on Twitter
1/14 Once upon a time (>2 years ago), we started the journey of making the PREVENT-AD data open. We had to reconsent the participants, develop the platform to host and share the data (thanks to Alan Evans and @NeuroLibre teams), clean the data and create the data dictionary.
2/14 This ended up being a tremendous amount of work and resources because the dataset was not initially designed to be shared openly.
3/14 Almost all of us are in favor of open science. However, only a fraction of the data ends up being shared openly. While the “fear of being scooped” might be one of the reasons, I see many other obstacles that limit data sharing.
4/14 First, someone needs to be willing to collect the data. The contribution of data collection (and sharing) is often overlooked by funding agencies. Reviewers positively rate good science but rarely assess the beneficial impact of data released for the research community.
5/14 It is obviously way faster to analyze existing data than write a R01 or a CHIR grant to get the funding to collect new data and share them. We however constantly need new data to test new ideas and prevent the fatigue effect of re-analysing the same datasets.
6/14 For us, sharing our data has been exhausting. Sharing data takes some skills and money. Individual lab collecting data do not necessarily have the platform, the funding, or the knowledge to easily share their data.
7/14 Sharing data takes time. A LOT of time. I often see on tweeter people complaining about asking a researcher for data and never getting an answer. I have been on the other side and I realized that sharing well organized data (AND answering the questions) takes days of work.
8/14 My first PhD student spent nights and weekend sharing data to collaborators and answering their questions for other students writing papers on the data she has collected! Some might say that this should have been done by an RA, yes of course, but you need money for that.
9/14 What can we do as a community to accelerate data sharing? I think (but please share your thoughts) that the funding agencies should join forces and facilitate data sharing. At least recognize the added value to the research proposal.
10/14 Funding opportunities may be prioritizing for researchers that want to share the collected data. Standardized consent form templates would be provided to investigators. A secured platform, easy to use, would be proposed for free for data sharing by these same agencies.
11/14 Key demographic information/MRI sequences… would be collected in all funded projects and shared with the granting agency as soon as the data are collected. The funding agencies would also have a liaison person that can answer basic questions about the different datasets.
12/14 Now, keeping the best for the end. It is with a lot of proud that I am sharing the link to the updated PREVENT-AD dataset. It contains MRI, cognitive test and medical info on 348 participants with a family history of AD that were all cognitively normal at entry. 
13/14 These have been followed longitudinally over 5 years (and still going); a total of 1559 MRI, 1300 cognitive evaluation, 476 CSF samples and now share for the Stage 1 of this initiative at https://portal.conp.ca https://www.biorxiv.org/content/10.1101/2020.03.04.976670v2
14/14 End: Special thanks to @prevent_ad, J. Breiter, J. Poirier, @acehigh1952, @NeuroLibre, @jbpoline, J. Tremblay-Mercier , C. Madjar, Samir Das, Jordana Remz, @alexa_pichetb, @evp82, @CIHR_IRSC @DouglasResearch, @FRQS1, @SciChefQC @TheNeuro_MNI
, C. Madjar, Samir Das, Jordana Remz, @alexa_pichetb, @evp82, @CIHR_IRSC @DouglasResearch, @FRQS1, @SciChefQC @TheNeuro_MNI
14b/14 
here is the easiest way to access the full dataset https://registeredpreventad.loris.ca/
here is the easiest way to access the full dataset https://registeredpreventad.loris.ca/
