Best programming practices for open science

Last Friday (May 25th, 2018) two of the PhD students from our lab, Dominik and Maciej, organised the workshop Best programming practices for open science generously funded by GW4.

The aim of the training was to raise awareness and introduce good programming practices adopted widely in the industry, but severely underused in academia. Target group for the course were mainly post-graduate students who are using programming in their daily work for numerical simulations, data processing or statistical modelling.

The invited experts were also advocates of open source scientific software developement. Demonstrations included practical solutions on how improve the accesibility of the academic code, and in results improve reproducibility of the scientific process.

Dr Vincent Knight and Nikoleta E. Glynatsi from Cardiff University School of Mathematics presented good practices on how to increase code readability and reliability. They opened our eyes on code testing, comments and variables/function naming conventions.

Dr Cyril Charron from CUBRIC presented arguments in favour of using distributed version control systems. Participants could practice it during next GIT hands-on session. Next, we went beyond coding with Dr Daniel Finnegan from University of Bath, who presented useful tools for scientific writing (LaTeX) and reference managing (Mendeley, Zotero). Rounding up our workshop, Dr Candice Morey of Cardiff University, shed light on how to apply open science practices to directly improve the scientific work and increase overall transparency and reliability of the whole process.

According to the feedback forms received, participants had positive impressions. The workshop was full of networking, discussions and stimulating questions. Seeing young researchers being enthusisatic about improving the quality and transparency of their work was very rewarding. That begs the question, why this type of training is not included in most PhD program curricula or at least offered in doctoral academies.

New Logo

After a couple of months of fruitful discussions we finally decided to change the old logo for the bright new one. The svg version with white and transparent backgrounds might be found in images folder of this website, or just here.

2017 is quickly coming to an end. We finally got a chance to get (almost) everyone together for a photo (and some nice food). A lot happened this year. Thanks to various funding sources, a few excellent new lab members joined in, and they are now on several interlinked projects. We are also sad to say goodbye to Jacopo, who joined us in 2016 as an intern and later as a RA. Jacopo did a wonderful job in completing behavioural and EEG experiment(s) in free choices under uncertainty. We wish him all the best in his new research post in London.

Looking forward to 2018, we aim to focus on three priorities.

1. Clear the desk

This is not about keeping our office desks tidy (well, maybe we should, or should not), but to bring completed projects to closure, with new follow-ups if necessary. Everyone is excited to initiate a new study (who isn’t!), but (e.g., myself) often less so on something we did three years or even three months ago. In 2018, we will try to slim down our file drawers as much as possible, meaning to publish or preprint current drafts, including null findings, analysis scripts and stimulus.

2. See the wood for the trees

Clearing the desk frees up mental capacity. It will naturally be devoted to new adventures: new paradigms, new equipment, new collaborations, (sometimes) new grants, … (did I mention that this is exciting?). We however need to maintain a clear mind on the big scientific questions, which fundamentally interests us. Incremental or independent studies are critical to drive science forward, but may get lost if they are not signposted to a big question.

Of course, your research interests may differ from other lab members (hopefully not too far apart). The key is to frequently review them, and whether they are consistent with your current work. For PhD students, this will help deliver a coherent thesis. For research fellows, this will help shape your future independence. For the PI, well, this will help me not become multi-armed Avalokitesvara, hopping between projects. To make this emphasis, in 2018, we will discuss long-term plans for every new experiment or grant proposal, and not be satisfied solely by “a possible 4* paper” (oh and yes, we still love them).

3. Fusion and integration

We are based in an imaging centre that excels in integration and collaboration. Many of our ongoing studies benefit from our collaborators’ strong support (advice/resource/expertise). In 2018, we will continue the dedication to these successful collaborations. Meanwhile, we now have a diverse set of skills in the lab, enabling joint forces on multi-modal imaging. New projects will probe how this approach facilitates imaging-behavioural integration in understanding decision and action the human brain.

This month we welcome Beth Routley to join us as a postdoctroal fellow. After her successful PhD, Beth has done an amazing job as the Public Outreach Officer at CUBRIC, promoting Brain Games and other engagement events. Now the opportunity comes to have her inputs on an exciting project.

Beth will study data from a large group of patients with epilepsy. During epilepsy surgery evaluation, several electrodes have been implanted deep in patients’ brain, and brain activity has been recorded before and during seizure. She will first examine how seizures arise from small brain regions, using mathematical models of neural networks that describe the changes of brain cells and their connections over time. Next, She will use the parameters of the models to simulate brain activity, and validate our method by comparing simulations with the real patients’ data. Our novel analysis may lead on to new therapeutic and surgical strategies for controlling seizures and eventually enhance the quality of life of epileptic patients.

Our website migrated from a sluggish Wordpress host to a static site solution, build with jekyll + github pages + cloudflare. The site is based on the themes of Neo-hpstr and Grayscale, and the People page is based on the one from Kording’s Lab.