The PERL explores how the brain processes emotional information and how this is influenced by brain chemicals and medicines.
This helps researchers to understand disorders such as depression and anxiety and to understand and contribute to the development of drug and psychological treatments.
Project lead: Professor Catherine Harmer, Professor of Cognitive Neuroscience, Department of Psychiatry (contact: Dr Cassandra Gould van Praag)
The Wellcome Centre for Integrative Neuroimaging WIN, of which the PERL is a part, does not have a local policy on open access publication other than the requirements of funders. The lab’s research is funded by MRC and the BRC (Oxford Health Biomedical Research Centre) is funded by NIHR, both of which have open access publication policies.
Much of the lead in open neuroscience comes from North America, for example the Poldrack Lab at Stanford Centre for Reproducible Neuroscience (as confirmed through interview with WIN group leaders). There is room for Oxford to have a significant impact as a UK driving force.
The director of PERL is supportive of increasing openness. This includes permission to spend time curating metadata in line with best practice.
The WIN has a commitment from the outset for ‘open neuroimaging’ (Theme 5): 'One of the key themes of WIN is to increase the openness of our data, to be a leader in the way that data is shared, and importantly, in how tools (such as the FMRIB Software Library), pipelines and databases can be used to ensure that the data shared is useful to other researchers ... All members of the WIN management board are committed to sharing their data openly by the end of the 5 year project, and all Centre members are encouraged to do so'.
The centre employs a Public Engagement Ambassador in each lab and is developing a role of Open Scholarship Ambassador for each lab to support colleagues to work in an open way.
A key organisation for researchers, the Organization for Human Brain Mapping (OHBM), an international society dedicated to using neuroimaging to discover the organisation of the human brain, is taking an open policy stance, and has an Open Science Special Interest Group whose mission is to advance neuroimaging research by fostering the open sharing of ideas, data, and tools between members of the OHBM community. The society has published its Committee on Best Practice in Data Analysis and Sharing (COBIDAS) report in an attempt to advance open science in neuroimaging. A new publishing platform, Aperture, is due to be released by the society. It is described as ‘a new initiative to develop an open publishing platform by and for the OHB community'. This platform will include the facility for open peer review, support reproducibility and dynamic publication. PERL and BRC Experimental Medicine researchers are following COBIDAS recommendations which provide evidence of good practice for reporting to NIHR.
Case study details
The degree of openness of research data is limited by data protection and ethics regulations because much of the data gathered by researchers comprises sensitive personal information. For example, a brain scan cannot be released without permission even if the data have been anonymised. The WIN is due to develop infrastructure and publish advice on permitted openness. At the time of writing release of summary data within figures in papers is allowed, however raw data and other data may not be released. Researchers often initially reserve data under privileged access until they have published at least one paper based on the data. Reasonable requests for access to data may be treated on a case by case basis, for example for collaboration. For the time being, researchers are being cautious with their own data until clearer guidelines have emerged.
The BRC aims to support its researchers by giving people the skills to empower themselves to make their research transparent and reproducible. Much of this initiative involves understanding programming and sharing code packages. The group is currently using tools with licences that results in the content being locked up, so a decision has been taken to move towards Python and open source software where possible in future. The group is currently reprogramming experimental tasks and stimuli into an open form (PsychoPy, an open source application allowing you run a wide range of neuroscience, psychology and psychophysics experiments). Tasks will be released with DOI (using Zenodo or OSF) alongside an empirical research publication, so the software will become a documented and attributable research output.
Training events have taken place involving the use of GitLab, programming in Python and Bash, using electronic lab notebooks (LabArchives) and working within the guidance of the Brain Imaging Data Structure standards for image file naming and metadata. Further training is scheduled for document version control using the Open Science Framework.
Experiments are routinely preregistered on clinicaltrials.gov as a required by many journals in this field (for example the Journal of Psychopharmacology 'conforms to the ICMJE requirement that clinical trials are registered in a WHO-approved public trials registry at or before the time of first patient enrolment as a condition of consideration for publication'). The group is in the process of preparing materials for more complete pre-registrations (such as full statistical analysis plans) via OSF.
WIN is building a local data repository for sharing images.
PERL is collecting metadata about the tools, analysis methods and settings in use so details can be reported well.
Articles are published open access in compliance with funders’ requirements, however there is no local policy requiring open access to publications.
MIT licences may be applied to software when their use has been approved by Oxford University Innovation and by the head of academic department.
Dissemination is aimed at patients and the general public as well as other researchers.
There is a strong focus on returning information to patients. Experiments are designed with patients in mind and to ensure their needs are accommodated (for example transport or how questions are phrased). In March 2019 Dr Jessica Scaife of PERL won a departmental award for public patient involvement and science communication.
There is an active public engagement programme as part of WIN.
Assessment, metrics and impact
PERL works with pharmaceutical companies to ensure that trials are pre-specified. PERL published the first study (to their knowledge) that locked down and pre-registered fMRI analysis pre-blinding. This information was used alongside other results to modify the ‘label’ (Summary of Product Characteristics) for the antidepressant vortioxetine by the European Medicines Agency to include cognitive impairment (Smith et al., 2018 Mol Psychiatry. 2018 May;23(5):1127-1133).
Funding is provided by MRC and NIHR. It can be used to cover costs of some publications, plus contributions towards training of open scholarship focused staff such as ambassadors and open scholarship facilitators. BRC contribute towards this, with main funding from (Wellcome Trust funded) WIN.
Benefits of open approach
- enables contribution to the field of research in the best way for the benefit of patients
- allows the best science to be undertaken in the most efficient and effective way
- the tools and processes that have been put in place are now being used to train the next group of researchers in making open scholarship normal practice
- adds an additional competitive edge to the group who can spread their research wider
- the skills needed to openly put individuals at a competitive edge because such skills are increasingly being recognised as valuable on a personal CV
Drawbacks of open approach
Although pre-registration should be the norm, researchers are worried about spending precious time on peer review of methods when involved with a short project. There is a risk the project will not be completed before the grant runs out and the researcher leaves. Better ways to work pre-registration into the grant lifecycle would be valuable.
Learning new skills when using new tools requires time investment. There is a temptation to muddle along ad hoc not using best practice rather than spending the time required to learn new skills. The result would therefore be that some of the reproducibility is lost.
Possible changes to make the research more or less open
A restricted instance of OSF (Open Science Framework) has been set up for depositing all documents around reproducibility, for example, parameters of how data were recorded.
WIN Open Scholarship working group may benefit from support in finding managing communications.
Additional freely available information and advice for all members of the group and other external researchers about working in an open research environment and how to successfully promote openness by individuals would be beneficial.