| Title: | Leveraging the power of Big Data to build more accurate models of neuro-anatomical and genetic variation: applications in neurodevelopmental research |
| Lead Institution: | King's College London |
| Principal investigator: | Emma Robinson |
Application 53775
WARNING: the interactive features of this website use CSS3, which your browser does not support. To use the full features of this website, please update your browser.
About
Preterm births are the commonest cause of loss of disability adjusted life-years in children under five years; about one third have neurodevelopmental impairments (such as learning disabilities and Autism), and the population faces a 7-fold increased risk of serious mental illness. We know that to give these children the best possible outcomes we must intervene at the earliest possible time. The causes of neurodevelopmental conditions are complex and varied but we have learnt a great deal from studying brain imaging data from these populations. More importantly, by looking at brain images in combination with genetic information, our group proposed new targets for clinical therapies for vulnerable preterm babies. We seek to extend and improve upon these successes by performing even more sensitive analyses. The problem is, that to bring together imaging and genetics to make very accurate predictions of neurodevelopmental outcome requires enormous data sets, far more than we can ever hope to collect due to the thankfully small numbers of very preterm babies born in our hospital. In addition, brain imaging data collected from developmental data sets is quite challenging to use on account of the significant challenges of imaging vulnerable newborns. Furthermore, it is important to stress that we do not yet even have a completely accurate understanding of how healthy brains work from which to compare our vulnerable population. For this we need the UK Biobank. Using Biobank data will allow us to compare results we have form work in newborns with a large adult dataset and establish if associations between genetic variation and brain structure in newborns continue into adulthood. The Biobank dataset will allow us to trial new methods for investigating associations between imaging and genetic variation, not feasible in our smaller newborn sample and then explore the results in newborns. In short, the goal of this study is to build much more accurate models of how the healthy adult brain works, studying how it varies across different people, and how these differences relate to genetic variation. Once this is done we will adapt these models to work for state-of-the-art developmental data that we have acquired through own Developing Human Connectome Project (dHCP, http://www.developingconnectome.org/). In this way we will be able to identify at birth which babies are most at risk of long-term brain health complications. Our work will lead to the development of treatments and therapies that will improve outcomes for these children.1 Publication
| Pub ID | Title | Author(s) | Year | Journal |
|---|---|---|---|---|
| 19304 | Unsupervised multimodal surface registration with geometric deep learning | Mohamed A. Suliman (+3) | 2025 | Medical Image Analysis |
Enabling scientific discoveries that improve human health