Home - Brain - Large-Scale Brain Network Modeling for More Optimized Tumor Removal
Research Brief

Large-Scale Brain Network Modeling for More Optimized Tumor Removal

A new study explores the idea of incorporating a large-scale brain network model, called The Virtual Brain, into pre-surgical planning

Meeri Kim, Contributor
Wednesday, May 30, 2018


Shutterstock/Tyler Olson


Before removing a brain tumor, surgeons must determine appropriate resection boundaries to spare as much normal tissue as possible in order to preserve essential functions. Since anatomical landmarks alone cannot define areas of brain function, they currently use noninvasive neuroimaging techniques such as functional MRI to make this decision. However, even these techniques fail to capture the complex nonlinear dynamics of the brain and don't always accurately predict functional outcome after surgery.

A team of Belgian researchers explored the possibility of large-scale brain network modeling -- which integrates neuroimaging data with biophysically based models -- to determine the boundaries of surgical resection in patients with brain tumors. They simulated large-scale brain dynamics with The Virtual Brain, an open-source neuroinformatics platform, to create personalized models for each patient. The results were published by the journal eNeuro May 29.

The study included 25 patients with brain tumors from Ghent University Hospital and 11 healthy controls. All participants received three types of MRI scans: anatomical, functional and diffusion-weighted MRI. They input these scan data to The Virtual Brain, which then simulated biologically realistic dynamics for both generic and subject-specific brain anatomy. The Virtual Brain is based on the fact that a perturbation in certain areas of the brain can cause changes in far distant areas, so such a model is more accurate than functional MRI.

The authors demonstrated that adding data from the three MRI scans to The Virtual Brain resulted in a significantly improved prediction accuracy of individual functional connectivity. In patients, the model could differentiate between regions directly affected by the brain tumor, regions distant from a tumor and regions in a healthy brain. The results confirm that brain tumors have both a local and distant effect on function beyond what their mere size would reveal.

Although more research and a greater sample size is needed, the study does point to the potential of large-scale network modeling for pre-surgical planning.