Search the site...

organization for human brain mapping
  • BLOG
  • Tutorials
  • Media
  • Contributors
  • OHBM WEBSITE
  • BLOG
  • Tutorials
  • Media
  • Contributors
  • OHBM WEBSITE

 BRAIN MAPPING

 

2017 In Review

12/22/2017

0 Comments

 
BY THE OHBM BLOG TEAM
As brain mappers start to log off their computers, comfortable in the knowledge that their OHBM 2018 abstract is finally submitted, we, at the OHBM Blog, offer a round-up of our most interesting and informative posts from 2017. This platter of self-citations should provide sustenance for those experiencing neuroscience hunger pangs throughout the holiday season. Each of our main contributors provides insight into their favourite posts - and if you feel you’d like to contribute as a guest blogger next year,
let us know!
PictureAmanPreet Badhwar
Aman:
As 2017 comes to an end, I think of the numerous ways that OHBM has promoted diversity since that first meeting in Paris, 22 years ago. I realized this while interviewing Marsel Mesulam, a longstanding academic inspiration of mine and a founding member of OHBM. I found myself entranced by Marsel’s recap of the organization’s history. He touched on the themes of discovery, flexibility, and evolution in the field of human brain mapping, and the importance of taking a step back and drawing inspiration from the brilliant diversity that is OHBM – be it the various imaging modalities highlighted, the composition of OHBM’s membership, comprised of both trainee and established members, or multidisciplinary interactions such as the annual art and neuroscience exhibits. I truly believe that OHBM draws its strength from its inherent diversity, an ingredient necessary to advance the understanding of the organization of the human brain. I look forward to ushering in 2018 with the OHBM community.


PictureJean Chen
Jean:
This year I met interesting people and learned interesting things through writing the blog posts. My favourites are: (1) the story of the first human fMRI experiment at the MGH (Mark Cohen interview); (2) finding out about the future of data sharing from David Van Essen; (3) the chaotic but pseudo-stable nature of brain connectivity; and (4) how stimulating the lateral prefrontal cortex makes people comply more with social norms. As a PI, I find that all too often, PIs and trainees may get tunnel vision, being committed to certain research findings and to propagating certain theories. It is true that we all have to publish and propagate knowledge as researchers, but the reasons that we became researchers in the first place may be something different. From working on the OHBM blogs, I got a sense of history, of different perspectives, of how successful scientists can reinvent themselves and stay true to their passions. That experience was both humbling and energizing. As we look to the start of a new year, I ask you one question: “Sure, doing science can be tough, especially these days, but what would you rather be doing with your time and your brains?”
​

PictureNikola Stikov
Nikola:
I had a lot of fun interviewing Alan Evans ahead of the annual OHBM meeting in Vancouver. However, my favorite post this year was not written by me, but by Agâh Karakuzu, a student of mine who wrote about his impressions as a first-timer at the OHBM Hackathon. The pleasure came from guiding Agâh through the labyrinthine process of introductions, interviews, standard operating procedures and gruelling team edits, only to see his efforts validated by the overwhelmingly positive response from the community. I feel like the hackathon post provided exactly what the OHBM open science SIG needs -- easy entry points for the uninitiated. I hope this post will motivate other OHBM trainees to volunteer their time and energy, be it in making science more open, or in spreading the word about the exciting initiatives coming from OHBM in 2018.
​

PictureNils Muhlert
Nils:
Whilst I was mainly involved in editing this year (including setting up the gruelling team edits), I did get the chance to interview a number of the OHBM execs - finding out about the challenges of working as treasurer and chair. But my personal favourite was interviewing the 2017 program chair Mike Greicius. His clinically-focused work covered such a wide breadth, from direct stimulation of the anterior cingulate in those with epilepsy to amyloid PET imaging in people with Alzheimer’s disease. Fleshing out the stories behind these papers and finding out his career path seemed to me to be exactly what we’re aiming for with our blog. Looking forward to our blogposts in 2018, you can expect more clinical neuroscience, open science, advice for early career researchers and coverage of brain mapping from around the globe.

PicturePanthea Heydari
Panthea:
My favorite experience of 2017 was surely the completion of my graduate studies and moving from the world of graduate student into a fully fledged PhD --- feels like I’m finally part of the cool kids club! A close second, though, was my OHBM Interview with Susan Bookheimer. Susan’s neuroimaging work at UCLA is fascinating and diverse, and her attitude and moral convictions are bold and impressive. It was refreshing to have a scientifically stimulating conversation with someone who shares such strong opinions on personal accomplishments, women in science, and the importance of life outside the PhD. ​
​

PictureShruti Gopal Vij
Shruti:
The end of this year marks my first year as a blog team member! Through the changing weathers, my work has changed colors, flourishing into an array of insightful posts on a variety of topics. It was fun liaising with the OHBM Student Postdoc SIG and writing for early career issues such as mentorship. But through all this, my favorite was to interview my postdoc advisor Lucina Uddin as an OHBM Young Investigator Awardee! It was enriching to see what makes a young investigator. I loved writing about things she is passionate about and sharing the pearls of wisdom that I, as her postdoc, have gleaned from her presence and mentorship. It was also inspirational to interview stalwarts such as Damien Fair, and see what a significant role mentorship has played in their careers. Personally, it was reassuring to know that I am surrounded by scientific experts who also value building mentee careers!
​

PictureThomas Yeo
Thomas:
In addition to contributing to this blog, I am a member of the 2018 OHBM local organizing committee. These two roles nicely intersected as my postdocs (Csaba Orban and Valeria Kebets) and I put together a blog post introducing Singapore as the location of next year’s annual meeting of OHBM. I hope everyone is as excited as I am that OHBM will be held in Singapore. Look forward to seeing everyone here!




​...and we’d like to thank all our contributors and interviewees, Sarabeth Fox for filming, and Randy Gollub, Niko Kriegeskorte, and especially Stephanie McGuire for their help in keeping the blog running!


Interested in suggesting a topic or writing a guest post for 2018? Contact us at ohbmblogteam@gmail.com

0 Comments

Open science topic: NeuroDebian

12/19/2017

0 Comments

 
By Agâh Karakuzu & Nikola Stikov
Picture
Neuroimagers face an ocean of software tools that have the potential to make research transparent, sharable, and collaborative. Open science is on the horizon, and we could use some instructions on how to ride the open software wave. We couldn’t think of better software surfers than Michael Hanke and Yaroslav Halchenko, the creators of NeuroDebian. NeuroDebian is a curated ecosystem of neuroscience research software that runs on virtually everything. It originally started as a personal convenience tool in 2005, then went above and beyond its initial purpose, finally transforming into a widely-used and globally accessible platform.

AK: What is NeuroDebian? Can you briefly explain it in layman's terms?


Michael: NeuroDebian was started almost 12 years ago, and was originally intended to provide correct software ecosystems. This is not only about shipping software to another computer, but a full system integration effort, so that independently developed software can work together in one system. NeuroDebian is basically a front for the Debian system, a decentralized and democratic effort of individuals working together to develop the universal operating system. Think of it as a neuroscience capable operating system.

Yaroslav:  Let me add a really layman definition. It’s your cellphone with the app store, but it is not just the cellphone. It can be your laptop, PC or even a computing cluster. NeuroDebian and Debian provide this whole turn-key platform where you have an app store offering accessible software. All that software is already available for you, and a dockerfile or singularity file can describe which ones to install. I think we cannot get more layman than that.

Michael: Did you say lame?
Picture
Aftermath of the layman joke with the sneak preview of a new data control platform by Yaroslav and Michael.
AK: NeuroDebian was not initially intended to be a global project. Tell us about how it grew into something bigger.

Yaroslav: It started from scratching our own itch. We were doing our PhDs (or so we thought), and needed software to analyze data. We were Debian users already, and thought it would be cool if we could share our work: I packaged and maintained PyEPL, Michael did the same for FSL. We soon realized that there are more interests besides those two software tools. The catalyst was us joining the Jim Haxby Lab at Dartmouth. In a frenzied week in 2009, we packaged lots of potentially useful software. Then we established NeuroDebian, as it is known now.

AK: The spirit of free software is at the heart of NeuroDebian. It is a community driven development. How do you manage all those remote contributions?

Michael: Yes, it is a global enterprise with many people contributing to it. But it is also, given its size, complexity and the amount of effort that goes into it, by far not international enough. For the amount of impact it has, it is actually a fragile enterprise. It has taken up considerable time and effort from researchers, even though the outcomes are not going straight into their PhD or research projects. Some software is more difficult to integrate than others. I don’t want to single any out, it just symbolises the whole situation and reflects the world we are in. Tools are kept behind closed doors until the people who developed them are properly credited.  

Yaroslav: In comparison to where we were 12 years ago, we are now in a very different ecosystem. In getting there we used many of the concepts of open science and open software, such as continuous integration. Without that we’d be fixing bugs for every release. To ensure scalability, we encourage people to test their software, and even though this is not directly related to NeuroDebian per se, it ensures that software remains working on all systems. Another angle is that we share the responsibility. Michael packages something and I package something else, and then we share it so that anyone can benefit. The same goes for the teams within Debian (Debian-Med, Debian-Science) which we are also part of, and we maintain many packages together with those teams. We are not duplicating anyone’s effort. Other team members do their part and we often just borrow relevant neuroscience research packages and backport them for all Debian and Ubuntu users. Everyone is happy at the end of the day. Once again, it would be impossible if these were done in an independent fashion.

AK: We hear about backporting. Is there a simple explanation for that?

​Yaroslav: It takes a lot of effort to stabilize a release. That’s why the most stable Debian releases are usually the most outdated ones. This takes us back to Michael's comment about integration. When you put together software that is supposed to work together and then you attempt to stabilize it, eventually it remains stable over years. But research software has different needs. Researchers want to adapt new methods as quickly as possible. So, we take a stable Debian release and put new software versions in it. If we were to upload fresh tools to the unstable versions of Debian and wait until it becomes stable, it would take too long. This is the point where backporting comes in handy. It allows us to provide the most up-to-date software for all Debian and Debian derivatives, such as Ubuntu.
PictureYaroslav, clearing the way for reproducible neuroscience.
​AK: The name gives the impression that NeuroDebian is exclusive to the Debian OS. What about researchers used to OSX or Windows?

Yaroslav: We saw cases where people installed NeuroDebian on a virtual machine and after a while realized that they keep their OSX or Windows in the background all the time. So they installed Debian as dual-boot at the beginning and then wiped out anything else but Debian. There are other people who prefer to keep their original OS, because they need to use Microsoft Word. You can even run NeuroDebian within NeuroDebian.

AK: How is NeuroDebian maintained, where do the resources come from?

Yaroslav: First it was our advisors. They allowed us to pursue this instead of our research. Some projects provide a small proportion of the funds needed to support NeuroDebian. But overall, NeuroDebian never received dedicated funds as a project. Whenever we apply, there is always one reviewer who says  “this effort matches Friston’s brilliancy of SPM” , and another who says “it is just some packaging, who cares”.

Michael: I think we should name drop here. Yaroslav’s advisor was Stephen Hanson, who paid for this for years. My advisor was Stefan Pollmann, who did the same. After many unsuccessful applications, I stopped thinking about grants. Once we wrote a grant where we had 40 letters of support, but reviewers questioned the letters’ legitimacy. Next time we gave them twice as much, and that was also not enough. The best comment we received was ‘this doesn’t fix any disease, try elsewhere’. Most of the stuff we do cannot be tied to a single purpose. We are not the ones achieving it, we are enablers. If any blog readers have recommendations, or are sitting on a pile of money, please contact us! Very recently, the CBBS research center in Magdeburg has started funding work that relies heavily on NeuroDebian and thereby helps to maintain it for everyone else, too.

Picture
Ready to attack large datasets. PyMVPA course, HWK, Germany, 2014
AK: We know custom installations of neuroscience research software can be painful. How easy is it to install them on NeuroDebian?

Michael: It really depends on the complexity. You can have a really easy installation after downloading gigabytes of binaries that are pre-compiled for your system. If you want to install it again, you need to repeat the same steps. However in Debian, you have all the benefits of automatic upgrades. You would say “I use 150 software packages on this computer and I need to download and update all of them”. The more complex the system you use for your research, the greater the benefit of using a package manager system.

Yaroslav: Do you remember how difficult it is to install apps on your phone and maintain them? Not really! You just click, install them and they keep updating themselves automatically. This is primarily the same thing. Once your research software is installed you don’t need to think about where it comes from or how to update it. It just gets done and it works.

Michael: Most people do not have administrative privileges on the hardware or the execution environments they use. Think about institutional clusters, think about workstations managed by someone else. In general, you need to call a system administrator and say which software you need installed. If you are on Debian, this only takes a few minutes.

AK: Most of the processing pipelines have a heterogeneous working environment with strict version dependencies. This is a serious threat for sustainability and reproducibility. How does NeuroDebian tackle this problem?  

Yaroslav: A while ago, we created the NeuroDebian virtual appliance. Downloading this virtual machine image,  different people can have the identical environment. There are additional projects that we can benefit from. One of them is http://snapshot.debian.org/, which takes snapshots of the entire Debian archive repository twice a day. So you can recreate any Debian system you used in the past. We now do the same thing for NeuroDebian, although it is not fully public yet. This way, if you know that your software environment previously existed, you can re-establish it. It can also be used to validate the compatibility of your current system. We are pursuing this through the ReproNim project that can track versioning for even more crowded environments.

​Michael: You can keep using the script-based generated environment for as long as the underlying components exist. But in my experience, many claims of strict version dependencies are convenient excuses, but not real dependencies. Often it’s because somebody spent time manually assessing whether a software works as it should. That’s why they have that strict dependency, because they are not willing to go through that manual effort again.  In this case you are following a trust and hope model. No manual validation that is done on one machine scales to the other. Remember the paper showing surprising differences in Free Surfer outputs depending on the execution environment. You need high level tests for machine based verification so you can make sure that when you ship it to your users, it will work. NeuroDebian offers solutions to this problem by enabling inclusion of build-time and run-time tests.

AK: How can developers get their software into NeuroDebian?

Yaroslav: You should email us to start with, or email the Debian mailing list. For some people who reached out to us, we packaged and maintained their software. Others packaged it themselves and gave it to us to review and upload. So, there are multiple ways to achieve this.  But if you mail us first, we can figure out the optimal way. Just remember, if it is for us to package, it might take longer because of the previous questions.
Picture
NeuroDebian is an efficient way to make your software available to a wide public. You can get the ball rolling with an e-mail to the address above.
PictureThe Michael Hanke test: if you can answer these questions, you are ready to make first contact with the NeuroDebian team
Michael: There are so many people and so many different projects in this field, not necessarily limited to neuroimaging. It doesn’t matter who you contact, as long as your initial e-mail contains all the relevant information. There are key questions you should be able to answer immediately before you contact them.

AK: How do you see the future of open science? Which role will NeuroDebian play along the way?

Michael: I am happy to say that open science is the present, but surely it is also the future. Open Science is not just doing science in a different way, it is the only way. All the excuses people had not to be fully transparent while spending public money for public research can be justified only by the technology limitations of the past. Today, we have surpassed those limitations. There is no point in keeping things secret. In addition, open source software development is the role model of open science. If you look at collaborative coding platforms like GitHub, you will see how low the threshold is for people to work together. The same will happen with open science, and things will become much more interesting and faster.

Yaroslav: Open science has already benefited from various standardization projects. For example, the open brain consent forms enable you to be open from the beginning. You don’t have to reveal all your cards, but you kind of pave your way, so that later on you can open up your research when the time is right. 

Picture
Michael: One thing to plug is DataLad. Many things we learnt in collaborative open source projects have manifested themselves in our Datalad project. It is the fusion of code and data, with reliable dependencies on datasets and versions. So everything we talked about regarding software can be mapped one-to-one onto data. In the future, you’ll probably see more and more efforts like this.

AK: Tell us more about DataLad

Michael: DataLad is for data what Git is for repositories. I use it for my students, who have no idea about Git. We use it for projects, such as the StudyForrest project. DataLad can move data in a version controlled fashion from one machine to the other: you can publish unilaterally to Github and cloud storages. In the next release, we will provide extremely enhanced metadata support, so you can make data discovery between datasets that you don’t even have. Furthermore, you can ask DataLad to watch a website! For example, if you were following  an fMRI dataset, and there was a change, DataLad will let you know what has changed, why it has changed and how the code was modified.

AK: Anything we did not ask? Would you like to add something?

Yaroslav: Do not be silent. If you use something, say it. If you use something and it does not work, say it. I am considered one of the biggest complainers in the community, but if we keep quiet it doesn’t get fixed. Constructive feedback is very much appreciated here. This is not just about NeuroDebian.

Picture
Michael: The NeuroDebian page shows the popularity statistics. There are about 300-500 machines a week that install NeuroDebian (or at least fill out the registration form). On the other hand, if the number of feedback emails goes above 10, it is considered a heavy week. There are two ways to interpret this: 1) This thing works! 2) It is a silent climate, and you are dealing with an undefined object. We don’t track people. It would be nice if people even told us that there is nothing broken.
Picture
0 Comments

OHBM Student Postdoc SIG Mentoring Programme: Q&A with Mentors and Mentees

12/6/2017

1 Comment

 
BY MICHELE VELDSMAN AND SHRUTI VIJ
Academia provides a unique set of challenges throughout one’s career. It is often highly competitive and uncertain. The evolution of science is unpredictable and this can leave researchers, at all stages, unsure of what next steps to take, how to manage their careers, or build their confidence. Mentorship is key to navigating a career in the face of this uncertainty. Successful mentorship requires an unbiased perspective from an experienced individual within academia who is dedicated to your personal and professional development.

This year, the OHBM Student and Postdoc SIG launched an international, online mentoring programme. This novel programme, pairs researchers of all levels across the globe. By pairing individuals across the international community, mentors can bring a fresh, objective perspective to the relationship while mentees provide a unique window into the changing landscape of research. Over 400 OHBM members enrolled in the programme in the first 6 months!

To provide a real platform for mentoring, pairs were encouraged to launch their mentoring relationship in person at the annual meeting. More than 180 pairs met for the first time in Vancouver and have continued their relationships online via email and video chat. In an effort to get more human brain mappers into active mentoring relationships outside of their current environments, new enrollments have begun for another round of pairing mentors and mentees. To sign-up, visit this website.

Here we present the first of a series of short interviews on the experience of mentors and mentees that give a personal insight into the benefits of the programme! The interviews also include general advice that mentors have for trainees, and discussions on the challenges that early career researchers face. We start with our Blog Team Captain Nils Muhlert and his experience in the mentorship program. He was paired with Professor Robert Turner, one of the pioneering physicists responsible for the discovery of MRI and fMRI.
Picture
Michele Veldsman (MV): Nils, you signed up to the programme looking for a mentor and volunteering to mentor trainees. Why do you think mentoring is important in academia?

Nils Muhlert (NM): As a PhD student, and early postdoc, you’re still safely tucked under a senior researcher’s wing. While you can become more independent in these stages, there’s still someone offering (largely) independent advice. As you progress into faculty positions, these sources of advice are no longer a formal part of your work.

Despite mentors being less common in later stages, I’ve always wanted a good source of advice in my career. For instance, one of my concerns has been not having a clearly defined research area. It’s useful to speak to those who have reached a high level in their career – who can look back and see what might have been useful for them, and the missteps that may have frustrated others. This was an issue on which Robert Turner (my OHBM mentor) offered helpful advice.

MV: What has been your experience of the mentoring programme so far? Have you seen any benefits?

NM: I was pretty impressed to see that Prof Turner would be offering advice. I knew his work, and had seen him give some good talks in the past (and ask some tough questions). I sent along my CV and explained the stage I was at. I also mentioned my main concern: how to balance multiple research interests with dwindling time, particularly given lecturing workload. The response I received was insightful:
Picture
Advice from Professor Turner:
“What worked for me as a scientist is having the determination to focus on crucial neuroimaging questions, and to put my available time fruitfully into the development of new techniques to address them. At important stages of my earlier career, I tried to avoid investing large amounts of time on sideline projects (though I have always been very happy to be a contributor to some papers).

Great outcomes tend to follow from intense concentration. When I was working on gradient coil design, back in the 1980s, I was doing algebra and writing the ensuing novel software 16 hours per day. As a result, I came up with the shielding equation that revolutionized gradient design. When you have broad interests, it's very important to reflect on what questions really matter to you, and what you are really skilled at--because to put in the necessary commitment you absolutely need the enthusiasm. You seem to have already proved yourself to the extent that you should be able to pick and choose what you are most enthusiastic about. You can always put other interesting problems on the back boiler--with luck you can pick them up again when the time is absolutely right."

MV: What do you think are the biggest challenges in navigating a career in neuroimaging?

NM: As discussed, I find knowing what to focus on to be difficult. There seems to be different advice on this – for some, like Tianzi Jiang, looking to predict future trends has proved useful. Robert’s view was that we should also aim to be ambitious enough to create future trends: “The key to this is to work out what are the right questions--questions that are fundamental, overlooked, simple, and answerable.” Alternatively, in a careers-advice talk, Ralph Adolphs suggested considering a varied portfolio of research, so as to be eligible for a broad variety of funding. Clearly, different strategies work for different people – but having tried the broad approach, and following Robert’s advice, I feel that for me, moving back towards fewer areas of research fits my current career stage. Whether I can fit in working 16 hours a day alongside looking after a messy 3 year-old is, however, another matter!

MV & SV: As a window into an evolving and developing programme, we have received overwhelmingly positive feedback from mentees and mentors from all academic stages. We will continue to highlight the feedback in a series of similar posts from other mentors and mentees from a range of backgrounds and career paths. Meanwhile, if you are interested in learning more about the programme, please contact the OHBM Student Postdoc SIG at ohbmtrainees@gmail.com and in order to sign up, visit this website  to complete a short questionnaire that will help match you with a suitable mentor or mentee. This sign up round will close on the 17th December 2017 and new pairs will be assigned shortly after.
1 Comment

    BLOG HOME

    ​TUTORIALS

    ​MEDIA

    ​contributors

    ​OHBM WEBSITE


    ​

    OHBM OnDemand 
    ​Education Platform


    RSS Feed

      blog survey

    Submit

    Archives

    February 2021
    January 2021
    December 2020
    November 2020
    October 2020
    September 2020
    June 2020
    May 2020
    April 2020
    March 2020
    February 2020
    January 2020
    December 2019
    November 2019
    October 2019
    September 2019
    August 2019
    July 2019
    June 2019
    May 2019
    April 2019
    March 2019
    February 2019
    January 2019
    December 2018
    November 2018
    October 2018
    August 2018
    July 2018
    June 2018
    May 2018
    April 2018
    March 2018
    February 2018
    January 2018
    December 2017
    November 2017
    October 2017
    September 2017
    August 2017
    July 2017
    June 2017
    May 2017
    April 2017
    March 2017
    February 2017
    January 2017
    December 2016
    November 2016
    October 2016
    September 2016
    August 2016
    July 2016
    June 2016
    May 2016
    April 2016

stay connected with ohbm!


become a member

Telephone

952-646-2029

Email ohbm

EMAIL BLOG TEAM
Header image created by Thiebaut de Schotten & Batrancourt  
www.brainconnectivitybehaviour.eu