Introduction: Brain mapping has undergone a paradigm shift from functional localization to focusing on complex network connectivity. Central to this has been the search for the connectome or the brains wiring diagram. Modelling the effects of focal lesions using graph theory allows consideration of how important a region is to network function and the effects of its removal. Our aim is to determine the feasibility of applying connectomics to neurosurgery and determine the key topological characteristics of patients with real lesion.
Methods: Resting state functional MRI at 3 Tesla was performed with multi echo independent component analysis pre-operatively on 5 patients with glioblastoma in the right temporo-parieto-occipital region. Complex networks analysis was performed by parcellating the brain into an anatomically based 116 region template followed by wavelet based decomposition of timeseries into correlation matrices that were subsequently thresholded and binarised into individual adjacency matrices.
Results: Our dataset exhibited the key features of complex networks found in healthy controls including ubiquitous small world features of simultaneous network segregation and integration. An exponentially truncated power law fit to the degree distribution predicted findings of general network robustness and a core of highly connected and integrated hubs with disproportionate vulnerability. Real lesions produced both local and distant effects in terms of reduced connectivity, network fragmentation, as well as alterations to the topological core structure of hubs and robustness.
Conclusions: Our refined analysis pipeline confirms the feasibility of performing complex network analysis with graph theory in patients with real lesions and is a novel approach to pre-operative brain mapping. Potential discrepancies between the effects of real and simulated lesions may allow identification of mechanisms behind network plasticity. Pre-operative mapping of network hubs and robustness is a novel approach for understanding mechanisms of how higher cognitive processes are affected by and recover from real lesions.
Patient Care: 1. Improving out understanding of brain mapping using a network analysis approach will allow more radical resections of focal lesions while maintaining or even improving cognitive function and quality of life
2. Network analysis offers the potential to model dynamic processes such as seizure propagation and the effects of cortical stimulation and better understand their biological basis
Learning Objectives: By the conclusion of this session, participants should be able to:
1) Describe the relevance of the connectome to understanding modern neuroanatomy.
2) Understand how to construct and analyse the connectome from resting FMRI data using graph theory.
3) Realise how a network based approach of brain mapping focusing on highly central 'hubs' and vulnerable areas can be used to guide surgical resection.
4) Recognise how modelling simulated dynamics on networks can improve our understanding of processes such as seizure propogation, cortical stimulation, or extra-lesional resections.