IDCC 09: Anne Trefethen Presents NeuroHub

Anne Trefethen introduced NeuroHub, which is a much smaller project than the US projects presented earlier in the day. Their aim is to work with neuroscientists to enable them to share their data – work which they help will fit into the wider picture and generate some useful tools.

The project involves Oxford , Reading and Southampton universities, together with their research council STFC e-Science Group. She explained how late trains brought two main players in the project together in conversation, which helped to develop the platform for collaboration which evolved between the universities.

Trefethen then focused in on the science involved. She noted that their work will only of value if they help to deliver the science. She explained the specific projects each university group are working on – including studying the way neural networks of insects work to enable them to move their limbs. She asked us to take note of the types of images that her slides showed – demonstrating the types of data that the scientists want to share with their colleagues – in raw form. She drew our attention to the note that some of the diagrams were stored with Spike 2 software – exemplifying the need to be aware of the wide range of tools when storing and sharing data. Data does not just include images, but also video. She explained that one of the apparently small, but significant considerations is that the scientists do not want to have to use a USB key in order to share their data.

She emphasised that it is very important to identify what the data we are are using actually are. Experiments do not necessarily create metadata to make it easy to find and share the results later. She also noted the complex range of software products being used in different ways to collect, process and publish the data.

To overcome this, the NeuroHub project involves embedding the developers in the neuroscience labs in the early stages to gain insight, combined with structured and unstructured interviews to establish how all of these issues mesh together.

Trefethen then moved on to look at the challenges they are facing – a list that she admitted could have been taken from Douglas Kell's opening keynote. The variety of interdisciplinary teams, different expectations, cultures, requirements, and understanding of shared terms, which can all obstruct data sharing. They have been using an agile development process to try and resolve some of these challenges and ensure that they develop tools that actually work for the scientists in practice.

She explained that their aim is “jam today, jam tomorrow” i.e. doing simple things that can make a big difference. This can include things like format conversions and proper annotation to help facilitate data sharing.

Trefethen then introduced some of the related projects that they are interacting with – including myExperiment (“Facebook for scientists” - socialising the data and providing annotation) and CARMEN, which is larger than NeuroHub, but more focussed on one area and works in the same community – promoting standards. There is a lot out there that they can integrate into NeuroHub.

In explaining the environment architecture of the project, Trefethen emphasised that they did not want to develop a large, monolithic system, but rather something that is in their workspace and creates an environment that empowers the researchers.