NeuroHackademy, an event which brought together dozens of neuroscientists over a two-week period to solve problems, learn new data science techniques, and network, concluded earlier this month. Organized by Ariel Rokem, senior data scientist, and Tal Yarkoni, director of the Psychoinformatics Lab (University of Texas at Austin), the event offered tutorials and lectures on machine learning, bioethics, programming, version control, and more.
We offered participants the opportunity to give feedback on the event; here’s what they had to say. Some content was lightly edited.
NeuroHackademy exceeded my expectations for a summer course and hackathon. I feel incredibly fortunate for the Neurohackademy experience and will view it as a pivotal point in my research career. I learned the tools and resources to advance my research project that I’ve been struggling to find on my own. Most importantly, NeuroHackademy offered a community valuing open science and collaboration.
We came away with a vast array of open source code and tutorials, as well as resources for accessing open datasets and cloud computing. This has not only benefited me personally, but has been easily shareable with my entire lab. The organizers did an amazing job of teaching so many different topics, but also set the tone, and continued to emphasize, that the environment of NeuroHackademy was one of continued learning and collaboration, where each person brought an important set of skills.
Such a positive environment was just as important and remarkable as everything we learned from the lectures or tutorials. This was my first hackathon and it was so much fun that I will be actively looking to participate in hackathons in the future!
As a fourth year Ph.D. candidate, I was ready to shift all my focus to advancing my projects with little to no distraction. To my surprise, attending the two-week NeuroHackademy event not only helped broaden the horizons to my scientific knowledge, it also rejuvenated my curiosity as a researcher.
The students attending the event came from diverse backgrounds and various academic status; we even had industry developers and a newly appointed professor join us in learning about open science, data sharing, and reproducibility. The first week was filled with informative lectures given by leaders in the neuroimaging field. The lessons ranged from techniques in neural data processing and machine learning to an array of tools for code testing and integration.
All lecturers were aware of each other’s work through the open science community, and their passion for transparency, reproducibility, and collaboration in science was evident in their talks and our casual social gatherings at the end of the day. By the end of the Hackademy, I had the principles of open and reproducible science ingrained in my brain, and I was burning with excitement to bring all these practices back to my lab.
I’m extremely thankful to the organizers and UW’s eScience Institute for hosting such an incredible event, and I’m even more thankful to the open science community for welcoming me with open arms.
NeuroHackademy was an extremely positive and rewarding experience. It was an honour to learn about data science, computer programming, and open science from experts in these fields, and to collaborate on interesting neuroimaging projects during the hackathon. I greatly appreciated the hands-on nature of many of the lectures, which helped us to put the new skills we learned to practical use.
We covered a broad range of topics, such as machine learning, statistics, data visualization, Github, Docker, Jupyter, BIDS, Nipype, neuroethics, and reproducibility. Importantly, many of the tools we learned about can be used to make our science more open and accessible to the neuroimaging community at large. We also had the opportunity to engage in meaningful discussions about how to foster open science and collaboration, and best practices for creating positive and inclusive open science communities.
What was truly incredible about NeuroHackademy was the sense of community that was evident from the first day of the program. The instructors and participants were welcoming, inclusive, and very willing to help others. Everyone had such unique skill sets, which made for a highly supportive group in which everyone could feel comfortable teaching and learning from each other. It was inspiring to see such a diverse range of skills and interests being applied to unique projects during the hackathon, and I am grateful to have had the opportunity to work with such a wonderful community of passionate scientists.
NeuroHackademy reinforced the importance and benefits of open science and collaboration, and provided us with a strong skill set that will allow us to practice open science and build supportive communities. I left NeuroHackademy feeling strongly motivated to continue to develop my skills in these areas, and to share what I have learned with my colleagues back home. I sincerely thank all of the organizers, instructors, and participants at Neurohackademy for a truly enriching experience, and I highly recommend this program to anyone who is interested in neuroimaging, data science, computer programming, and open science.
Dr. Patrick Beukema
Throughout my years in graduate school, I had always wanted to be a part of, and contributor to, the open science community. I finally had that opportunity at NeuroHackademy. I had just completed a PhD in neuroscience from the University of Pittsburgh and Carnegie Mellon and I was looking to meet and learn from some of the most vocal and active open science contributors.
NeuroHackademy is an incredibly unique event. You get to meet a community of open science advocates from all over the world who are passionate about building software. The school cultivates that passion with hands on tutorials from leaders in open science and data science on the best practices in open source development. For example, we walked through the latest in JupyterLab with Fernando Perez, and scikit-learn with Gael Varoquaux and Jake VanderPlas. Lectures were supplemented with Jupyter notebooks which meant that everyone in the room could follow along and run the code themselves.
NeuroHackademy culminates in a hackathon where, as part of a team, you build a software project. Our team initially built a deep learning pipeline that generated T2 weighted images from T1 images using the Human Connectome Project (HCP) database. However, after an entire day spent data wrangling instead of writing code, we quickly realized that a more useful contribution would be a tool that directly interfaced with and preprocessed data from HCP’s cloud storage on Amazon Web Services. After some encouraging feedback from Ariel and Tal, we pivoted and got to work designing and building that library instead. We hoped this would eventually save valuable research time and make it easier to conduct reproducible analyses on such a massively rich dataset. The whole experience was supremely fun and productive, and there was even time to explore some of the many outstanding local breweries.
Practicing open research, especially as a young scientist, can be challenging, especially if you are taught that impact factor is more important than rigor. NeuroHackademy is a testament and a model for doing science a different way, and the community is getting stronger every day.
NeuroHackademy is a super-fun and useful summer school for anyone in neuroscience interested in obtaining more computational skills for open source collaboration and reproducible research. You leave with a lot of concrete tools such as how to collaborate on open source projects through github, design more reproducible research, write better code to avoid big mistakes, speed up your computer processing time, use machine learning for neuroscience, and much more.
Everyone was very supportive and inclusive, and it was clear we had total freedom to explore ideas throughout the hackathon (which was super fun, by the way). So there was no pressure and it was good for people with different skill sets.
Seeing how much can be achieved through open science with many important people in the field is an eye-opener for me and I think for everyone who participated. So you leave with new values and views about science and many new friends. Plus, paddleboarding and hiking through Seattle was pretty amazing…
I come from a lab that is very open towards learning all the latest and greatest methods and practices, but does not have the in-house experts to help us get there. In order to learn, I have to craft my own path of discovery by trying to glean as much information as I can from blogs, twitter posts, and online forums. While I have learned a great deal from this process, looking back I’ve realized how many roadblocks could have been overcome with a couple of helpful words from someone who knew what they were doing.
NeuroHackademy offered many helpful words from their knowledgeable instructors (and participants), benefiting experienced and novice coders and neuroimagers alike. I was blown away by the content at NeuroHackademy and my only wish is that this would have been offered during my first year of grad school. The skills I picked up through trial and error over the course of weeks on my own were covered succinctly in a couple hours at NeuroHackademy. Instead of doubting myself about what the “best practices” are, I can listen and pose questions to instructors that have much more experience: a resource that simply wasn’t available at my home institution.
In addition to the series of lectures/tutorials, we were given the opportunity to apply what we’ve learned during a hackathon. The hackathon portion not only taught me about the technical aspects of the projects I was working on, but also about building communities around the projects. The idea that a group of people can come together, sharpen their ideas towards a singular goal, and create something together in such a short amount of time is astounding. I met and worked with numerous wonderful people and I was amazed with how many helped with my project.
Neurohackademy provided the unique environment necessary to make such collaborations and communities possible. I learned invaluable skills I will carry with me for the rest of my career, and I am excited to continue collaborating with the all the friends I made over the two weeks at NeuroHackademy.
“I’ll switch to python in the next project; I already have a pipeline in MATLAB that does almost exactly what I need for this one”. “There’s no point in uploading my current analysis scripts to Github, they’re just standard processing – I’ll upload scripts when I have new analysis methods”. There are always good reasons to wait – just a bit – before taking the plunge.
NeuroHackademy provides a great blend of the different ingredients needed to transition to more open science – tools, connections and confidence. One of the most important things you realize is that it’s not all or none, and that taking a step in the right direction, any step is enough. The rest will follow.
Hearing talks by people who devote so much of their time to creating tools for the benefit of other scientists is inspirational, and the general enthusiasm for open science is catching.