Bipolar spectra matched 33 of 70 queries across separate shafts in 35 public SEEG patients with sample rate held equal

We tested whether spectral shape could match the same patient across separate electrode shafts and two public SEEG files. We used the first 60 seconds of two files from each of 35 patients in OpenNeuro ds004100 version 1.1.3. The fixed method used adjacent bipolar pairs, centered 2 to 45 Hz profiles, and candidates at the same sample rate. It matched 33 of 70 queries correctly, or 47.1 percent. The expected fraction under a shuffled patient map was 8.6 percent. None of 100,000 maps reached the observed count. The corrected Monte Carlo p-value was about 0.000010. The patient bootstrap interval was 32.9 to 61.4 percent for the fixed candidate set. This shows patient structure in these files. It does not separate physiology from electrode placement or recording settings. The biological interval between files is unknown. The package preserves code, spectra, predictions, shuffled maps, source hashes, and a raw retrieval recipe.

stereoelectroencephalography · SEEG · patient matching · spectral profiles · bipolar reference · electrode shafts · OpenNeuro