@ARTICLE{10.3389/fnhum.2020.00315, AUTHOR={Duffau, Hugues}, TITLE={What Direct Electrostimulation of the Brain Taught Us About the Human Connectome: A Three-Level Model of Neural Disruption}, JOURNAL={Frontiers in Human Neuroscience}, VOLUME={14}, YEAR={2020}, URL={https://www.frontiersin.org/articles/10.3389/fnhum.2020.00315}, DOI={10.3389/fnhum.2020.00315}, ISSN={1662-5161}, ABSTRACT={For a long time, the relevance of the information provided by direct electrostimulation (DES) for mapping brain functions was debated. Recently, major advances in intraoperative DES for guiding resection of cerebral tumors in awake patients enabled the validation of this method and its increased utilization in basic neurosciences. Indeed, in addition to the cortical stimulation used for many decades in epilepsy surgery, axonal mapping was developed thanks to DES of the white matter tracts, giving original insights into the neural connectivity. Moreover, functional results collected during intrasurgical mapping have been correlated with neuropsychological performances before and after DES-guided resection, and with perioperative neuroimaging data. Thus, it was evidenced that DES offers the unique opportunity to identify both cortical and subcortical structures critical for cerebral functions. Here, the first aim is to propose a three-level model of DES-generated functional disruption, able to explain the behavioral consequences elicited during awake surgery, i.e., (i) DES of an input/output unimodal (e.g., somatosensory or motor) network inducing “positive” responses (as involuntary movement); (ii) DES of a distributed specialized network inducing a within-system disruption leading to specific “negative” disorders (e.g., exclusive language deficit with no other disorders); (iii) DES generating an inter-system disruption leading to more complex behavioral disturbances (e.g., the inability to perform dual-task while each function can be performed separately). Second, in light of this model, original findings gained from DES concerning the human connectome, complementary to those provided by functional neuroimaging (FNI), are reviewed. Further longitudinal multimodal investigations are needed to explore neuroplasticity mechanisms.} }