Figuring out how the world works, for the sake of knowledge itself, is a vital human endeavor. Time and again we’re reminded that curiosity-driven basic science - wanting to know how something works - can have unexpected benefits later down the line. It can also be a wonderful occupation, with the joys of problem-solving, discovery, and collaboration with smart and enthusiastic people.
We expect everyone in the MetaLab to take advantage of these opportunities and show a real committment to and a passion for research. Hard work and good science comes in many forms and styles (see below - it’s definitely not about some required number of hours per week).
We want the lab to be a place where people of any gender, race, ethnicity or background can feel welcome and thrive. Steve feels this is important not only to ensure people feel comfortable and enthusiastic about joining us, but also because new ideas in science are most likely to be generated by groups with diverse backgrounds and perspectives. Our team is conscious of the multiple ways in which inequality is manifested in neuroscience and psychology, especially in light of the Black Lives Matter movement and acknowledged gender disparities. Built into our lab timetable is a programme of internal meetings to talk about these inequalities and create an open forum for discussion.
The lab, and the university, is committed to ensuring a safe, friendly, and accepting environment for everyone, and is an environment that must be free of harassment and discrimination. If you notice someone being harassed, or are harassed yourself, tell Steve immediately. If Steve is the cause of your concern, then have a word with your Head of Department or another trusted faculty member. UCL has a very useful and comprehensive Research Staff Code of Conduct that we adhere to, downloadable here. This lays out expectations of both researchers within the lab, and Steve as PI. Please make yourself familiar with its recommendations. We strive for a welcoming and inclusive atmosphere and encourage open and honest intellectual debate, which allows everyone in our local and international communities to do their best work and be respected. Call out misconduct when you see it; you can always approach Steve, or other faculty members in the department, if you want to discuss something confidentially.
With this privilege of doing science comes the responsibility of doing it well (or at least, in the best way we can). We’re responsible to taxpayers for getting the most out of their hard-earned pounds and euros; to our academic communities for not misleading each other purposefully or wasting time on dead-ends; and to ourselves for spending our time well. To maximize the usefulness of our science, we must do our work ethically and openly. This means being honest (with ourselves and others), transparent, and always open to learning and improving how we work. There are some obvious practical steps we can take to make all this happen:
We endeavour to publish papers as preprints (on bioRxiv, arXiv, psyArXiv or OSF). All papers submitted after January 2021 need to be in open-access compliant journals (in alignment with Plan S), and you should add the following statement to your Acknowledgments upon first submission:
“For the purpose of Open Access, the author has applied a CC-BY public copyright licence to any author accepted manuscript version arising from this submission”
You can use this to explore which journals are possible, but it can also be a bit complicated to navigate so make sure to speak to Steve before setting your heart on a particular journal. We also endeavour to use Twitter to write accessible summaries about our findings: it’ll even help your career.
As a general rule, we aim to choose free / open-source software and code over expensive and proprietary alternatives. In the lab we are a broad church when it comes to programming and software, with different projects requiring a blend of code written in Matlab, R, Python together with the range of probabilistic programming languages that interface with these wrappers (STAN, JAGS, Webppl, etc). We are practical when it comes to these choices - we don’t want to reinvent the wheel for the sake of it, as that is a waste of time and resources. This might mean using Matlab more often than we’d like, although Matlab is not universally free (UCL has a blanket license). But if there are alternatives, Steve is always happy to hear about them.
In general - keep up with the latest developments, and think about ways to improve the work we do.
A good scientific paper is engaging and interesting to read, but this should never come at the expense of transparency in report. For example, unexpected findings and post-hoc changes to the analysis pipeline should be presented to the reader as what they are. Whenever feasible we pre-register our study plans and hypotheses before data collection. Later, when we write about interesting and surprising aspects of the data, we have our pre-registration to remind us, and the readers too, of what we expected to find before seeing the data.
We are committed to being open in our sharing of resources and data with others. Within the lab, you should share code and data with whoever you like, and you should aim to help others where possible with problems you’ve encountered and solved yourself in the past. Outside the lab, please check with Steve before sharing code and data with others. Generally, we will aim to make the code and data freely available around the time of publishing the associated paper, but there may be some exceptions to this general rule, for instance if the data cannot be anonymised.
Reproducible research is an essential part of doing good science, and an expectation for all projects in the lab. For results to be reproducible, the analysis pipeline must be organised and well documented, and have a clear link to the anonymized data files (not preprocessed or otherwise modified). This means keeping good notes throughout the design, data collection and analysis process, so that you can document the methods as part of the paper. It also means a healthy amount of commenting of your code.
Towards the end of a project, and before submitting a paper, there is an expectation that all relevant code to reproduce key results will be reviewed and uploaded to the lab’s Github. It is much easier to do this when you have been keeping track of the analysis pipeline throughout the lifecycle of the project.
Take care of yourself: research can be exhausting. As Weiji Ma has said: we’re professional doubters, so doubting ourselves is a real occupational hazard. A certain amount of impostor syndrome is normal. Research can be tedious, frustrating and boring, scientific setbacks are normal, and peer review can be downright hostile. To stay sane and grounded, remember you’re a whole person: don’t neglect your life outside the lab. Spend time building and maintaining relationships, both inside and outside science: while enthusiasm over results, papers and grants will wax and wane, your peers and friends will not. This podcast episode has great advice on career satisfaction and well-being.
If you struggle with mental health, know that you’re not alone, and that resources are available. UCL has student psychological and counselling services that you can use freely and confidentially. The NHS too offers counselling. Finally, you can always talk to Steve in confidence about changes in your professional or personal circumstances or for advice.
In general, everyone making a substantial contribution to a project will have the opportunity to co-author the resulting paper. Early in the life of a project we will usually hold a meeting to explicitly discuss authorship expectations. If you are leading on data collection, analysis, and writing, the default position is that this will entail first authorship. Things get a little more complicated when these responsibilities are split, for instance if a Masters student project is being done in collaboration with a PhD student in the group. In cases such as these, please make sure to discuss your expectations about authorship with Steve earlier rather than later so that we can find a reasonable solution. One caveat to the above is that if a substantial time period passes without a paper being completed after someone leaves the lab (after 2-3 years of finishing data collection), then we may need to reassign the project to a new first author. The policy here is to prevent rare and useful data from remaining unpublished, while giving a healthy window for writing the paper to the person who initially led the project.