A cross-species, cross-modal approach to computational neuroanatomy

In vivo methods for mapping brain connections are increasingly used in systems neuroscience but are yet to have significant clinical impact.

We propose to develop new computational approaches that bridge information from precise but invasive methods in animals to enhance in vivo methods in humans.

We translate this framework into clinical care in surgical patients: In Aim 1, we will leverage a unique resource of macaque tracers currently being digitised.

We will use state-of-the-art machine learning to automate quantification, and image processing for mapping histology to MRI, to produce a unique resource of macaque ground truth connectivity.

In Aim 2, we will develop computational approaches to enable macaque anatomical tracers, alongside multimodal MRI in macaques and humans, to be used to improve the accuracy of methods developed in humans.

In Aim 3, we will build end-to-end approaches for connectivity-based functional localisation, and deploy this as a tool to aid pre-surgical planning for localisation of subcortical targets in deep brain stimulation and for localisation of eloquent cortex in tumour surgery.