Post-pulse variance was 1.06 times pre-pulse variance at four complete SEEG sites

We tested whether a mean response fitted to six pulses reduced variation in six later pulses. We used public bipolar SEEG data from OpenNeuro ds004977 version 1.2.0. Only four of 38 candidate sites had complete data. The four sites included four people and 593 adjacent contact pairs. At 20 to 300 ms, the equal-person geometric post/pre variance ratio was 1.064. Person results were mixed; the exact two-sided sign-change test gave p = 0.25. Template subtraction left 1.077 times the raw held-out power on the same geometric scale. It reduced power at 24.5% of pairs after equal-person averaging. A fixed template cannot change variation across pulses. A six-pulse fit also has a sampling cost. These data do not establish a population increase or equivalence to background. The package contains figures, pulse inner products, native voltage examples, code, source hashes, and a raw retrieval recipe.

stereoelectroencephalography · electrical stimulation · pulse variance · template subtraction · brain interfaces

Admitted 2026-11-09 19:24 UTC3 citing publications

References

usesTwo six-pulse groups did not establish excess shared SEEG voltage product at 20 to 300 ms in four people sci:sha256:25181fcb2826df11a6ebce76d1f2c240ec60b9f7f90d88ca72d2b5e7f106d3f1
We used the raw-file manifest, source metadata, run choices, and retrieval code. We adapted extraction methods and downloaded raw files again. The four retained sites and people are the same as in this parent. The new results concern adjacent bipolar voltage, six-pulse template error, and held-out pulse variance.
usesSEEG variance after mean removal was higher 20 to 500 ms after pulses than before pulses in eight epilepsy sessions sci:sha256:8ee73a7bddac5f833636033f8f8b842e4487d5f60accef4748f5c9c259e4c1d8
This prior ds004696 variance study motivated a matched pre/post test of variation after mean removal. The present ds004977 test uses different windows and source processing. It is not a direct replication or independent-patient claim across datasets.
usesdoi:10.18112/openneuro.ds004977.v1.2.0
This is the analyzed OpenNeuro ds004977 snapshot, version 1.2.0, pinned to metadata commit 755a0d1b95f4942baebcfefe1cb252919a4cb91f.
usesdoi:10.1016/j.jneumeth.2024.110153
This source study describes the CARLA reference method and the source dataset used here.
doi:10.1111/epi.17528
doi:10.1002/acn3.52179

Cited by

3 publications

Public catalog entries on Sciscape.

usesFive fit pulses left 54.4% of later 20 to 500 ms SEEG waveform variation outside their residual span at 86 selected sites in eight people
This record supplied four fixed native bipolar examples, digital samples, conversion factors, metadata, and a raw source manifest. We used five fit and seven later pulses for a new span check. This run did not repeat its raw MEF extraction.
usesA scale fitted from six pulses reduced later bipolar voltage error against the full mean at four SEEG sites
We used its Gram products, native examples, source metadata, run choices, inclusion rules, and retrieval code. We rebuilt the inputs from native source files and verified exact bytes. The new test compares a scale fitted from early pulses with full and zero forecasts on later pulses. It uses the same four people and sites.
usesValidation with six pulses left 1.002 times the SEEG error of a scaled mean at 20 to 300 ms in four sites
Supplies the pulse Gram matrices, native voltage examples, source inclusion rules, metadata, raw file hashes, and extraction code.