Mapping brain network activity from structural connectivity using AI and Deep Learning

The structure-function relationship is fundamental to natural sciences. The nervous system is organised as a hierarchy of progressively complex and interconnected neural populations. Traditionally, anatomical information provided high biological specificity and interpretability, yet it was inadequate for characterising the differences between individuals, in part because neural circuit functionality is influenced by tissue microstructure variations.
Modern neuroimaging allows probing into, and highly detailed reconstructions of, the brain's structural and functional connectivity networks. Using these, early modelling proved the existence of functional coupling between structurally linked regions. Correlations between functional regions however depend both on the presence of pathways and signals received from the overall network. To account for this, models predicting function through higher order interactions between neural populations were developed, yet they either lacked biological plausibility or were hard to generalise across individuals. Recent statistical methods successfully used microstructure derived measures to predict functional connectivity variations while neural networks based techniques have seen some success in predicting functional connectivity; however, the goal of estimating this directly from structural information has remained elusive.
This project builds on previous work highlighting correlations between structure and function using information extracted from diffusion MRI and resting-state functional MRI by developing a DL algorithm capable of translating between an individual's structural and functional connectivities, by learning the underlying relationships between them. Algorithms such as CNNs and GANs have been increasingly used in medical imaging due to their ability to learn complex features and relationships. This task involves both learning patterns and representations of direct signal and functional correlations between regions several synapses removed. The required biological insight and variability is achieved by using large clinical imaging datasets from UK Biobank. To retain the spatial contextual correlations of structural pathways and functional information, no data reductions are used before the joint between-modality modelling. This, together with the substantially higher dimensionality required for rich between-modality modelling stimulate the development of algorithms of much higher complexity than typical DL imaging applications.
Initially, the inputs to the network are the summation of 27 white matter tracts obtained from subject-specific probabilistic diffusion tractography using standard-space protocols, ensuring sufficient spatial information is provided to allow the learning of relevant pathway and functional correlations. The targets are the rsfMRI spatial maps, with the Default Mode Network being initially used out of the 21 major functional sub-divisions obtained through group-level ICA. Additional modalities and data, such as an individual's genetics, lifestyle, cognitive and physical measures, will be later added as inputs while the outputs will incorporate all the functional sub-divisions, thus aiding in determining whether correlations can be established between them and an individual's functional connectivity (and the way in which it differs from the population average). Achieving this would contribute to the development of probabilistic models assessing an individual's deviation from the population distribution and identify specific brain regions which contribute to this. Moreover, this could also aid in the development of pre-surgical functional mapping methods for physically impaired subjects without the need for challenging explicit cognitive/motor tasks.
This project falls within the EPSRC Medical Imaging research area, but also contributes to the Artificial Intelligence Technologies and Image and Vision Computing. It is a collaboration with F.Hoffmann-La Roche

The UK's world-leading position in biomedical research is critically dependent upon training scientists with the cutting-edge research skills and technological know-how needed to drive future scientific advances. Since 2009, the EPSRC and MRC CDT in Systems Approaches to Biomedical Science (SABS) has been working with its consortium of 22 industrial and institutional partners to meet this training need.

Over this period, our partners have identified a growing training need caused by the increasing reliance on computational approaches and research software. The new EPSRC CDT in Sustainable Approaches to Biomedical Science: Responsible and Reproducible Research - SABS:R^3 will address this need. By embedding a sustainable approach to software and computational model development into all aspects of the existing SABS training programme, we aim to foster a culture change in how the computational tools and research software that now underpin much of biomedical research are developed, and hence how quantitative and predictive translational biomedical research is undertaken.

As with all CDT Programmes, the future impact of SABS:R^3 will be through its alumni, and by the culture change that its training engenders. By these measures, our existing SABS CDT is already proving remarkably successful. Our alumni have gone on to a wide range of successful careers, 21 in academic research, 19 in industry (including 5 in SABS partner companies) and the other 10 working in organisations from the Office of National Statistics to the EPSRC. SABS' unique Open Innovation framework has facilitated new company connections and a high level of operational freedom, facilitating 14 multi-company, pre-competitive, collaborative doctoral research projects between 11 companies, each focused on a SABS student.

The impact of sustainable and open computational approaches on biomedical research is clear from existing SABS' student projects. Examples include SAbDab which resulted from the first-ever co-sponsored doctorate in SABS, by UCB and Roche. It was released as open source software, is embedded in the pipelines of several pharmaceutical companies (including UCB, Medimmune, GSK, and Lonza) and has resulted in 13 papers. The SABS student who developed SAbDab was initially seconded to MedImmune, sponsored by EPSRC IAA funding; he went on to work at Roche, and is now at BenevolentAI. Similarly, PanDDA, multi-dataset X-ray crystallographic software to detect ligand-bound states in protein complexes is in CCP4 and is an integral part of Diamond Light Source's XChem Pipeline. The SABS student who developed PanDDA was awarded an EMBO Fellowship.

Future SABS:R^3 students will undertake research supported by both our industrial partners and academic supervisors. These supervisors have a strong track record of high impact research through the release of open source software, computational tools, and databases, and through commercialisation and licensing of their research. All of this research has been undertaken in collaboration with industrial partners, with many examples of these tools now in routine use within partner companies.

The newly focused SABS:R^3 will permit new industrial collaborations. Six new partners have joined the consortium to support this new bid, ranging from major multinationals (e.g. Unilever) to SMEs (e.g. Lhasa). SABS:R^3 will continue to make all of its research and teaching resources publicly available and will continue to help to create other centres with similar aims. To promote a wider cultural change, the SABS:R^3 will also engage with the academic publishing industry (Elsevier, OUP, and Taylor & Francis). We will explore novel ways of disseminating the outputs of computational biomedical research, to engender trust in the released tools and software, facilitate more uptake and re-use.