In 43 SEEG patients, spike-ripple rate did not outperform spike rate for clinical SOZ pairs in 90 s NREM recordings

We compared interictal spike rate with the rate of spikes that had a co-occurring 80–250 Hz ripple in 43 epilepsy patients (728 bipolar SEEG pairs, 160 clinical seizure-onset-zone [SOZ] pairs, about 90 s of interictal NREM sleep per patient). The spike-ripple rate and spike rate had the same mean per-patient AUC (0.764); their difference was −0.0004 (95% CI −0.043 to 0.040; p = 0.99). The CI excludes a gain larger than 0.04 AUC for these detectors and data. The fraction of spikes with a ripple was higher in SOZ pairs (mean AUC 0.714), but did not outperform spike rate. A strict detector variant showed the same pattern. The data were taken from OpenNeuro ds007118, ds007119, and ds007120, version 1.0.0; all 43 EDF byte-range hashes matched the archived values. This study used short recordings and automatic amplitude detectors without visual review. It assessed clinical SOZ labels, not surgical outcome, and does not test visually reviewed events or long recordings. The findings do not prove equivalence.

spike-ripples · interictal spikes · seizure-onset zone · SEEG · epilepsy · NREM sleep · AUC