Spinal cord (SC) anatomy is often assimilated to a morphologically encapsulated neural entity, but its functional anatomy remains only partially understood. We hypothesized that it could be possible to re-explore SC neural networks by performing live electrostimulation mapping, based on “super-selective” spinal cord stimulation (SCS), originally designed as a therapeutical tool to address chronic refractory pain. As a starting point, we initiated a systematic SCS lead programming approach using live electrostimulation mapping on a chronic refractory perineal pain patient, previously implanted with multicolumn SCS at the level of the conus medullaris (T12-L1). It appeared possible to (re-)explore the classical anatomy of the conus medullaris using statistical correlations of paresthesia coverage mappings, resulting from 165 different electrical configurations tested. We highlighted that sacral dermatomes were not only located more medially but also deeper than lumbar dermatomes at the level of the conus medullaris, in contrast with classical anatomical descriptions of SC somatotopical organization. As we were finally able to find a morphofunctional description of “Philippe–Gombault’s triangle” in 19th-century historical textbooks of neuroanatomy, remarkably matching these conclusions, the concept of “neuro-fiber mapping” was introduced.