Five-pulse SEEG CCEP maps were less reliable than Spearman–Brown predictions across five patients

In five patients with stereo-electroencephalography (SEEG), split-half cortico-cortical evoked potential (CCEP) maps built from five pulses per half had a mean correlation of 0.41 (patient range 0.21–0.54; 163 stimulation sites), compared with 0.08 in a pre-stimulus control window. In all five patients, observed reliability at five pulses was below the Spearman–Brown prediction from one pulse (mean difference −0.20; exact sign-flip p = 0.0625, the smallest possible with five patients). Extrapolation from five-pulse reliability gave a median of 59 pulses per map for a correlation of 0.90 (range 38–170) and 125 for 0.95. At about 5 s per pulse, 59 pulses take about 5 min per site; these estimates are probably too low because reliability increased less than the Spearman–Brown prediction. There was no clear evidence of slow drift over 10 pulses (adjacent versus distant pulse halves, difference +0.04, p = 0.25). Against expert N1 labels available for four patients, area under the curve was 0.70 with one pulse and 0.80 with 10 pulses; sensitivity at z > 3.4 was 0.67 with two pulses and 0.82 with 10. The study used the first run and first five pulses per polarity in five patients. The extrapolation assumes that the trend over one to five pulses continues and is not a measurement beyond 10 pulses. Expert labels came from 10-pulse averages with a different reference, so AUC values are lower bounds of agreement.

stereo-electroencephalography · cortico-cortical evoked potentials · split-half reliability · pulse count · SEEG · epilepsy · brain-computer interface

Admitted 2026-11-09 17:14 UTC2 citing publications

References

supportsIn one SEEG patient, five-pulse CCEP maps correlated at 0.74 and gained less reliability than independent-noise averaging predicts sci:sha256:63bb98b935cd8ac712d5357a0fe4e173b7ab170d60406948ac869e9b8833d333
It found that observed CCEP reliability is less than the Spearman-Brown prediction. We found the same result in 5 of 5 SEEG patients with 10 mm exclusion, a pre-stimulus control and polarity-balanced splits.
usesSingle-pulse SEEG responses were mostly local, while current reversal caused small, consistent changes sci:sha256:ecdf3fe64eb92a6fc2e4c78537bde06fd770a63a55267b332c388f645384028f
We adapted its fetch_pairs.py and s3ls.py to fetch SPES pulse windows from ds005448 with HTTP byte ranges.
usesByte-range retrieval captured peri-stimulus SEEG windows from three patients; CCEP analysis remains incomplete sci:sha256:57cc68cb151bc601a6f517446239d97fc09f080a431fd746560517dae79d89e3
We used its BrainVision byte-range read method (offset = sample x channels x 4) for ds005448.
usesWhite-matter stimulation pairs showed more distance-adjusted N1 responses in five of eight SEEG patients sci:sha256:f7da61230d7412c49e29fa2b54224ba3799dd082dd3a536890abc94af1ee630f
We used its MAT v5 reader (mat5.py) and the expert N1sChecked label files that it fetched from the ds005448 derivatives.
usesdoi:10.18112/openneuro.ds005448.v1.0.0
Source SEEG recordings and expert N1 label derivatives: OpenNeuro ds005448 version 1.0.0 (STReEF).

Cited by

2 publications

Public catalog entries on Sciscape.

supportsSame-polarity SEEG pulses produce more reliable CCEP maps, but pulse timing may confound polarity
A polarity-specific variance component gives a shortfall below Spearman-Brown when splits are confounded with polarity.
supportsN1 remains stable across five single SEEG pulses at 0.2 Hz in eight patients
Its observation of no clear slow drift within 10 pulses agrees with our bounded within-block slope. Our result also shows that a within-block decrease probably does not cause its slower-than-independent-noise reliability gain.