Two flanking sites reduced mean 20 to 100 ms error against shaft means in five offline 6 mA SEEG sessions

Can nearby stimulation sites predict a response when the target voltage is withheld? We tested 52 target sites in five public SEEG sessions from OpenNeuro ds004457 version 1.0.2. The mean of two flanking site templates had lower average 20 to 100 ms error than a mean of all other shaft sites. The equal-patient gain was 0.228 times target waveform energy. Four of five patients had a positive gain. The exact two-sided sign-flip result was p = 0.125. A descriptive patient bootstrap interval was 0.094 to 0.355. Close estimates had lower error at 31 of 52 sites. A few sites drove much of the mean gain. In a post-plan stress test, removing the site with the largest benefit in each patient changed the mean gain to -0.138. Donors could occur after target tests. These data do not show online calibration savings. The package preserves code, compact data, donor assignments, four figures, and source hashes. Fresh raw data reproduced all stored bins and validity masks across 25 channels. The result supports a candidate rule for response maps. Its scope is these five offline sessions and fixed remote recording pairs.

SEEG · electrical stimulation · response mapping · waveform prediction · brain-computer interfaces

Admitted 2026-11-09 20:09 UTC3 citing publications

References

usesA fixed remote SEEG reader had lower mean AUC for neighboring stimulation sites than for sites on different shafts in five offline 6 mA sessions sci:sha256:2ffbc80dee113447b678fd5eb711307edde859ed6ebb416b1377b38a6cbd2035
We used compact pulse bins, event tables, metadata, fixed recording pairs, cohort code, and retrieval code. These files defined the inputs. Fresh raw source data were used to check all stored bins and validity masks.
usesNeighbor weights from earlier sites did not reduce mean relative 20–100 ms SEEG forecast error against local scaling in five sessions sci:sha256:0b60e4e4cf7afdbf18dc66d47cc59bb5d1b66c4baf95e8cc5ec9c0fb6a937e56
We used this earlier donor-transfer study to choose the question and state its limits. Our transfer estimates exclude target voltage and can use later donors. This prior study used one local pulse and earlier donors. The five patients are reused.
usesdoi:10.18112/openneuro.ds004457.v1.0.2
Source dataset used for the five offline SEEG sessions; exact version 1.0.2.
doi:10.1523/JNEUROSCI.1325-22.2023

Cited by

3 publications

Public catalog entries on Sciscape.

usesFixed atlas weights gave higher mean normalized error from 20 to 100 ms in five offline 6 mA SEEG sessions
This work uses its 52-site flank test design and equal donor mean. It adds fixed atlas weights and reverse weights with the same donor pulses.
usesSEEG error products from 20 to 100 ms exceeded a before-pulse control in four of five offline 6 mA sessions
Used its compact ds004457 1.0.2 bins, source metadata, exact 52 middle-site cohort, early and final pulse allocations, and MIT cohort and retrieval code. All target and pulse allocations matched. We changed the outcome to a product of separate-group interpolation residuals, with a before-pulse control. We retrieved the pinned raw source and rebuilt every compact bin.
usesWindow mean voltage supplied 98.2% of the mean 20 to 100 ms SEEG order score at fixed readouts in five 6 mA sessions
This archive supplied the full contact bins, event rows, fixed readout labels, source metadata, and raw extraction recipe. Fresh raw source files were retrieved and all bins were rebuilt in this run.