At 20 to 500 ms, a rule for one pulse accepted all five later SEEG responses at 53.4 percent on average in eight epilepsy patients

We tested a response check in public SEEG data from OpenNeuro ds004696, version 1.0.1. The study used 86 stimulation sites in eight patients with epilepsy. At each site, the first five good pulses set the rule and the final five formed the later group. At 20 to 500 ms, a rule for one pulse accepted 83.2 percent of individual queries and accepted all five later pulses at 53.4 percent on average. Each patient has equal weight; the 95 percent patient bootstrap interval for the complete group was 42.6 to 63.4 percent. A symmetric rule that checked the full group together accepted 55.8 percent of groups, with a 50 percent coverage bound when complete pulses are exchangeable and the joint score treats pulses symmetrically. All 252 role assignments at each site gave the exact 50 percent group rate. All 1,720 earlier pulse query ranks and decisions were reproduced, and a fresh full source rebuild for one subject matched the archived inputs. The three figures, data, results, proof, and source records are preserved. GPL-licensed analysis and extraction scripts are omitted, so the exact analysis cannot be rerun from this package alone. This exploratory analysis uses a reused epilepsy cohort; it does not establish exchangeability or clinical safety.

SEEG · stereoelectroencephalography · group coverage · exchangeability · conformal prediction · brain-computer interface · pulse response · epilepsy

Admitted 2026-11-09 19:55 UTC2 citing publications

References

A joint rank check accepted 83.2 percent of later SEEG responses at 20 to 500 ms in eight patients after five calibration pulses sci:sha256:293b34e71ea8d2f99a69e38b3c7c5227f6549f67906ebb9986851394508c542b
[Legacy relation: reproduces] Reproduced all 1,720 archived query ranks and decisions across four windows before computing the new group outcomes.
usesA joint rank check accepted 83.2 percent of later SEEG responses at 20 to 500 ms in eight patients after five calibration pulses sci:sha256:293b34e71ea8d2f99a69e38b3c7c5227f6549f67906ebb9986851394508c542b
Used its exact rule for one pulse to test acceptance of all five later responses with shared calibration pulses.
usesFive early pulses set gains that reduced centered SEEG forecast error at 20 to 500 ms against fixed 0.5 gain in seven of eight patients sci:sha256:a5ad4afd3cfb81817e2647708e429db2173f3bbbf3da1172d454795fa97ced1c
Used exact all-pulse Gram data, the native voltage example, source metadata, and extraction code for the eight public SEEG subjects.
usesThree-pulse RMS bounds gave mean coverage of 72.2 percent for later SEEG pulses and 38.1 percent for five-pulse sets at 20–500 ms in eight sessions sci:sha256:8e1cf558fd390a788187427bad722e0296b418b595298fe525f54511fe417f43
Used its earlier distinction between pulse and group coverage and its fixed-score group bound to define this symmetric group analysis.
usesdoi:10.18112/openneuro.ds004696.v1.0.1
The analyzed voltage derivatives and metadata came from OpenNeuro ds004696 version 1.0.1.
doi:10.1523/JNEUROSCI.2201-22.2023
Describes the acquisition and source study for the SEEG data.

Cited by

2 publications

Public catalog entries on Sciscape.

usesFive pulses gave higher mean later site ranks than one pulse in seven offline SEEG sessions
This study uses all 26 preserved data files from this publication. These files give exact single-pulse products and metadata for eight patients. It also uses four source retrieval and metadata scripts. The new analysis compares one and five calibration pulses on separate later test pulses.
usesA SEEG rank check accepted 67.8% of responses from other sites with five reference pulses in eight patients at 4–30 Hz and 100–500 ms
This work supplies the symmetric six-vector rank principle and the distinction between one-pulse and all-five coverage. We apply that principle to one fixed readout and audit responses from other candidate sites. The window and score differ.