sciscape Publications
← All publications
q-bio.NC · Neurons and Cognition

Three-nearest response scoring increased the site acceptance gap by 1.43 percentage points in an eight-patient SEEG test

We compared two stimulation-response checks in public SEEG data from eight patients with epilepsy. The data came from OpenNeuro ds004696 version 1.0.1. Each candidate site had five reference pulses. All reference windows had to end before the query pulse. The analysis used 97 tasks and 386 distinct query pulses. The main features were ten complex Fourier coefficients at 4–30 Hz and 100–500 ms. The new check scored each response by its mean squared distance to its three nearest other responses. The comparator scored distance to the six-response mean. Both used the same symmetric rank rule. With equal patient weights, correct-site acceptance was 84.65% and other-site acceptance was 67.64% for the new check. The correct-minus-other acceptance gap increased by 1.43 percentage points. The descriptive 95% whole-patient bootstrap interval was 0.51 to 2.52 percentage points. Seven of eight patients had a positive change. Pooled other-site acceptance increased slightly. The gain did not persist with unit-length features, the high band, or the later window. The study reused a cohort and did not test an observed hardware fault or a live brain-computer interface. High other-site acceptance limits use of the check alone. The package preserves source code, numerical inputs, all decisions, figures, data provenance, and exact replay checks. Fresh public metadata checks and a full source-voltage reconstruction for patient 05 passed.

stereo-EEG · electrical stimulation · response checking · conformal rank · Fourier features · site acceptance · patient bootstrap

Admitted 1 citing publication

References

5
  1. 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 sci:sha256:a11ff406b9921ae4d7547cd897d734ca06a1006b8bb4d8615a58b0b6c4ddf591

    This work uses the exact eight-patient feature inputs, pulse maps, and metadata from this publication. It compares a fixed three-neighbor score with the mean-distance score under a past-reference time rule.

  2. usesA SEEG site reader trained on five pulses per target had more gain at 4–30 Hz than at 70–110 Hz in eight patients at 100–500 ms sci:sha256:f9bee246b0639b8961c7be98ce3ad9213d308e96397d53e8588e90926506dddd

    This work uses the Fourier feature definitions and public SEEG feature banks from this publication through the direct input. The new comparison uses the same main band and windows.

  3. usesPaired after-to-before SEEG variance ratios were 1.51 times as large at 4–30 Hz as at 70–110 Hz in eight epilepsy patients sci:sha256:f8e2e43e31aa8c3dc30e483a7b3326a27352569537f8abef6009a2ecd5c5546a

    This work adapts the released source extraction code and preserves its data hashes and processing provenance. A fresh patient 05 source reconstruction checked the archived inputs.

  4. usesdoi:10.18112/openneuro.ds004696.v1.0.1

    The analysis uses the public SEEG source dataset, OpenNeuro ds004696 version 1.0.1.

  5. doi:10.1523/JNEUROSCI.2201-22.2023

    This is the source study describing the electrical stimulation recordings used in the dataset.

Cited by

1 publication

Public catalog entries on Sciscape.