Functionally Informed Hand Knob Reveals Structural Connectome Differences in Motor-Eloquent Tumours

Background: Brain tumours impose complex, spatially heterogeneous disturbances on neural circuits that extend far beyond the immediate lesion site. While gross anatomical displacement of the cortico-spinal tract (CST) has been extensively studied, the topological reorganization of the functionally informed structural connectome-and its dependence on tumour molecular phenotype-remains incompletely understood. Objectives: This study aimed to characterize upper-limb functionally informed network topology in brain tumour patients, identify histological and molecular patterns of structural reorganization at the cortical and subcortical level, and determine the impact on neurophysiological parameters. Methods: Forty-eight patients with supratentorial motor-eloquent tumours (MET's) underwent diffusion-weighted imaging (DWI) as part of their preoperative motor mapping. Connectivity matrices based on streamline passing counts were extracted from 426 nodes of the HCPex atlas using DSI Studio^® upon seeding the structural connectome in the motor hotspot (best motor response) for the functional area of the upper limb identified using preoperative navigated transcranial magnetic stimulation (nTMS). Paired-sample t-tests compared tumour versus healthy hemispheres across network topology metrics. The impact of nTMS-derived excitability metrics-interhemispheric resting motor threshold ratio (iRMTr) and cortical silent period (CSP)-and tumour histological and molecular characteristics on the connectome was assessed. Results: The global network topology of the tumour hemisphere was preserved when compared to the healthy baseline hemisphere across all tumour types (p > 0.05). Subcortical analysis revealed significant hyper-connectivity in the tumour hemisphere, with elevated degree, strength, clustering coefficient, local efficiency, and eigenvector centrality (p < 0.05). Basal ganglia motor loop degree was increased in the tumour hemisphere (mean 7.37 versus 5.19; p = 0.0009). IDH-mutant tumours generated significantly more topologically organized compensatory networks than IDH-wildtype tumours (Clustering Coefficient 0.345 versus 0.273; p = 0.018). Of the cortical nodes on the side of the tumour, significant hyper-connectivity was seen in the supplementary motor area (p = 0.0011), premotor cortex (Area 6), with increased connection seen in 6 mp (medial premotor at p =0.0033) and 6 d (dorsal premotor at p = 0.025) and primary somatosensory cortex (p = 0.027). The presence of the tumour induced significant changes across all three domains: loss of CST volume (p < 0.001), prolongation of the cortical silent period indicative of intracortical inhibition (p < 0.0001), with significant prolongation in glioblastoma versus an oligodendroglioma. Conclusions: Brain tumours significantly impact the upper limb-centred structural connectome. While global network topology is preserved, tumours induce substantial subcortical and basal ganglia network reorganization by inducing compensatory hyper-connectivity. These findings suggest that structural connectomics offers a novel framework for non-invasive tumour characterization and surgical planning.