Résumé
INTRODUCTION:The brain displays degrees of functional reorganization, i.e., neuroplasticity, that enable interventions that were previously thought to be impossible (e.g., resection of "Broca's area"). Studies have described the static, single-point in time neuroplastic potential in diffuse low-grade gliomas (DLGG) following one surgery.METHODS:We studied n=103 patients with a DLGG that underwent multiple brain resections achieved up to critical neural networks identified under awake mapping by means of direct cortical and subcortical electrostimulation. We used normalized postoperative images to generate probability overlap, atlas-based cortical and subcortical parcellation, and functional compensation maps for both surgeries. We performed statistical analysis correcting for multiple comparisons across all cortical parcels and tracts to determine differences at the cortical and subcortical levels between surgeries.RESULTS:We show for the first time the dynamics of neuroplastic changes occurring in DLGG that undergo multiple awake resections. We created residual and resection cavity overlap probability maps; cortical parcel damage and subcortical tract disconnection maps and the difference between surgeries; and functional compensation maps and subcortical volumes. We discovered the following four main patterns that constrain how and to what degree this dynamic reorganization occurs between surgeries: cortical vs subcortical, superficial vs deep, unimodal vs multimodal, and left vs right. We found greater plasticity at the cortical level, superficial terminations of white matter tracts, multimodal networks, and left ventral and right dorsal networks.CONCLUSIONS:These newly discovered patterns of reorganization provide a unifying model to understand the dynamics of neuroplasticity in DLGG patients undergoing multiple resections. These dynamics with its patterns that constrain mechanisms of neuroplasticity, can help guide neurosurgeons to predict which patients are likely to benefit from a repeat resection.