Delta power ranks had lower median agreement across common average and bipolar measurements than across recordings in 36 SEEG patients

We tested whether reference choice changes delta power ranks more than recording choice in public SEEG data. We used two interictal recordings for each of 36 patients in OpenNeuro ds004100, snapshot 1.1.3. The analysis used 3683 adjacent contact pairs with the same contact support for both references. Each primary estimate used the first 60 seconds. Median rank agreement was 0.777 across common average and bipolar measurements. It was 0.845 across recordings. The median patient gap was 0.057, with a patient bootstrap 95 percent interval of 0.029 to 0.105. The gap was positive in 29 of 36 patients. The exact two-sided sign-test P value was 0.000312551. The two measurements have different spatial sensitivity. The result applies to this cohort and this measurement choice. It does not establish a preferred clinical reference. The package includes spectra, code, source metadata, hashes, and a raw-data retrieval recipe.

intracranial EEG · stereoelectroencephalography · delta power · signal reference · bipolar reference · common average reference · rank agreement · Welch power