Using DataShare in linguistics research
Dr Bert Remijsen talks about his experience archiving and sharing his Linguistics research data.
Dr Maria Valdes Hernandez is a Lecturer in Medical Image Analysis at the Centre for Clinical Brain Sciences. Her “Step-by-Step Pipeline for Segmenting enlarged perivascular spaces from 3D T2-weighted MRI” has recently been approved and archived in the Human Imaging collection within DataShare.
DataShare is the University of Edinburgh’s digital repository. Research data which is potentially useful to the research community and/or part of a publication can be uploaded onto the platform, after which it becomes discoverable and a persistent identifier is created (DOI).
More about DataShare View Brain Imaging Pipeline
The pipeline was developed for the Mid Stroke Study 3, which examines possible contributing factors and clinical implications of small vessel disease. Small vessel disease is a leading cause of dementia and stroke, and can be seen on magnetic resonance imaging (MRI). In the Mid Stroke Study 3 study, adults who suffered mild-to-moderate stroke where assessed on various clinical, cognitive and lifestyle factors, as well as their MRI brain images.
More info on PubMed
The DataShare deposit contains the Matlab code which was used to analyse the brain MRIs obtained during the study. The code helped identify which regions of the brain are of particular interest to further understand the natural history of small vessel disease.
What is particularly interesting about this DataShare submission is the way in which the Matlab code is presented: “Rather than an end-to-end encapsulated software, the code is presented as a step-by-step pipeline starting by the brain parcellation (step 03), with comments justifying and explaining every block of operations, as well as alternatives, to facilitate its use.”
Schematic of brain imaging pipeline, as seen here.
DataShare is a key tool in making research findable, and in increasing outreach and impact. Maria explains: “DataShare is a great resource that for me has been invaluable in making freely available the algorithms and data from our publications to the international community, helping us to comply with the requirements of transparency and reproducibility in reporting the outcomes of our analyses. It also is very useful for training and teaching in addition to research purposes.” Examples of DataShare deposits that Maria uses for training and teaching can be found here and here.
DataShare is managed by the Research Data Support Team, who have created a step-by-step guide to help with the submissions process. If you want to submit to DataShare, you need to choose a suitable location for your dataset amongst the existing collections. Then you will also have to describe your data using the metadata fields and choose the correct licence. The Research Data Support Team are available for questions. They will help you ensure that the data you upload is in the best possible structure and format to maximise findability, reproducibility and usefulness.
View Research Data Support
The pipeline that Maria and her team created is widely applicable, as it contains a huge number of adult MRI brain scans and associated vascular abnormalities. However, to become relevant to specific cohorts, it requires further fine-tuning.
Overall, the brain images, in combination with the other assessments of the patients, will help advance the understanding of small vessel disease and identify the people who may be at increased risk. In the end, 229 participants were recruited, and the study is still ongoing as patients are being followed-up. Preliminary outcomes are currently in peer-review.
This case study was written by Dr Sarah Janac, Research Facilitator for the College of Medicine and Veterinary Medicine.