Squared voltage centre times preserved late SEEG site order better than absolute peak times in five patients with separate smoothed five pulse means

We tested two late timing summaries for stereoelectroencephalography (SEEG) in public OpenNeuro ds004457, version 1.0.2. Five patients had 1,081 retained site and receiver links. The test used an 11-bin box mean and times from 50 to 250 ms. It included all 126 separate five pulse group divisions. Mean site order agreement was 58.69% for the centre of squared mean voltage and 57.08% for the absolute peak. The paired gain was 1.60 percentage points. All five patient means improved. The empirical patient resample limits were 0.74 to 2.51 points. A constant map scores 0.5, and a strictly increasing rescale cannot improve the main score. All fixed checks had a positive mean gain, but the first-five versus last-five check improved in only three patients. The main before-pulse gain was 0.02 points. We verified 120 raw files, rebuilt every input array, checked all primary receiver scores with separate code, and repeated the numerical outputs exactly. The package contains selected voltage data, source locks, code, results, and repeat instructions. The result concerns repeated late timing order. It does not establish physiological latency, clinical accuracy, or a gain in a new patient population.

stereoelectroencephalography · electrical stimulation · voltage timing · peak time · squared voltage centre · site order · measurement reliability

Admitted 2026-11-09 17:41 UTC6 citing publications

References

usesThe mean patient median peak gap between reused and separate five pulse estimates was 0.661 baseline RMS units in five SEEG patients sci:sha256:200ca28b748d433e2b2c55ce512abca1bc91d5395210890b782904b94f0e5bb5
This earlier study supplied the metadata contact rule and MEF extraction code preserved through the input package. We used that code to retrieve the locked source and repeat the full extraction.
usesSEEG from five patients showed a mean 41 percent late absolute peak disagreement above 20 ms for separate smoothed five pulse means with Pearson r≥0.8 sci:sha256:321fe5dffd9bda55915cb0360180786244325bbab49a62af9a78c77bc116ac76
This study supplied the exact five patient pulse window files, metadata, and source locks. Its peak timing result led to the present centre versus peak order test. We kept its selection unchanged and rebuilt all stored arrays from checked raw data.
usesdoi:10.18112/openneuro.ds004457.v1.0.2
Source dataset analyzed: OpenNeuro ds004457, version 1.0.2; the publication preserves selected windows, source hashes, metadata, and retrieval instructions.
doi:10.1523/JNEUROSCI.1325-22.2023

Cited by

6 publications

Public catalog entries on Sciscape.

usesTrimming five training pulses reduced 50 to 250 ms SEEG prediction error in two of five selected patients
We used its inherited metadata selection, event selection, source MEF retrieval locks, and pulse extraction code. A fresh raw rebuild matched all input arrays.
usesA variance rule reduced average 50 to 250 ms SEEG prediction error in five patients with separate five pulse groups
This test reuses its five patient pulse arrays, metadata selection, extraction code, source locks, and source check code. It tests waveform error instead of response timing.
usesStudent t intervals covered 93.48% of bins but 45.93% of full SEEG waveforms in five patients
This publication documents the inherited pulse extraction lineage. We used its report to check the receiver selection, event selection, and scope of the saved voltage arrays.
usesExpected late SEEG centre times differed between two and eight pulse averages in fixed ten-pulse pools from five patients
We used its compact voltage windows, metadata selection, source locks, extraction code, and source check code. We tested count two versus count eight in the same ten-pulse pools. Its late timing order result motivated the measure and fixed processing choice.
usesA three component learned model had higher mean relative error than a cosine model for 50 to 250 ms SEEG prediction in five patients
We used its pulse selection and MEF extraction code to rebuild every input waveform from hashed public files.
usesFour learned patterns had 2.34 percent higher mean relative error than four cosine patterns in five SEEG patients with separate shafts
The inherited extraction code and receiver selection came through this publication.