Five-pulse signed templates did not show a clear AUC gain over energy on energy-selected pairs in eight epilepsy SEEG sessions
We used public OpenNeuro ds004696 version 1.0.1. The study had eight epilepsy patient sessions, 86 stimulation sites, and 886 eligible bipolar pairs. At each site, five early pulses selected one pair by energy. The same pulses defined a mean after-pulse template. We tested five later after and before windows with both scores on that pair. The window was 20–500 ms after stimulation. Equal patient mean AUC was 0.939 for signed projection and 0.931 for energy. The mean difference was +0.86 percentage points. The exact patient sign-flip test gave p = 0.4531. The exact test does not establish a difference at the 0.05 level. The descriptive 95% patient bootstrap interval was -0.90 to +2.90 percentage points. For 100–500 ms windows, the signed projection difference was +0.55 percentage points. The package preserves exact joint waveform products, selected source waveforms, scores, figures, source hashes, and an offline scoring tool. This reused clinical cohort has no sham input, neural ground truth, or new-session test. The result does not validate a live BCI control rule.
stereoelectroencephalography · electrical stimulation · response monitoring · signed projection · template matching · area under the curve · brain-computer interface · epilepsy
References
4- usesEarly power selection gave 0.931 mean AUC between later after-pulse and before-pulse SEEG windows in eight epilepsy patient sessions sci:sha256:b0d1b4a077fff84e56c767399fe1c933ce7ad2f5c319fa66176fa90683c52a20
This publication supplied candidate tables, exact within-window products, source hashes, metadata, and retrieval code. They fixed the eligible cohort and readouts selected by early energy. New source retrieval added after-to-before products for the template scores.
- usesdoi:10.18112/openneuro.ds004696.v1.0.1
The analyzed processed trial voltages came from OpenNeuro ds004696 version 1.0.1.
- doi:10.1523/JNEUROSCI.2201-22.2023
- doi:10.1371/journal.pcbi.1011105
Cited by
7 publicationsPublic catalog entries on Sciscape.
- usesFive calibration pulses gave mean later signed template AUC of 0.939 compared with 0.859 for one pulse in eight SEEG sessions
This publication supplied exact joint waveform products, source hashes, metadata, and the fixed five-pulse benchmark. We used these inputs to test one to five calibration pulses on the same later pulses.
- usesA template rank check increased the pulse–background acceptance gap by 17.66 percentage points over distance scoring in eight SEEG sessions with five early pulses at 4–30 Hz and 100–500 ms
Its signed-template versus energy comparison in the same cohort informed our fixed low-energy comparator. We changed the question to symmetric rank acceptance at all readouts with delayed low-frequency features. We did not reuse its selected-pair AUC as evidence for our new endpoint.
- usesAn exact sign-flip test passed at 70 of 86 SEEG sites with six retained pulses and a 20–500 ms window in eight people from ds004696
Uses the exact waveform-product format and pinned source derivative lineage preserved by this earlier package, through immediate parent 851374db-d616-40da-9101-a7e4a51128a6. The sign-flip response screen and six-pulse endpoint are new in this study.
- usesSigned templates had mean AUC 0.7925 versus 0.83625 for energy at 20–500 ms after contact order reversal in eight SEEG sessions
The earlier comparison of five-pulse signed templates with energy in ds004696 motivated this transfer test. The current study applies the same type of early energy selection and signed projection to STReEF contact order changes. It adds a separate source result and does not claim a new projection method.
- usesPair choice with five reused controls raised background calls under random roles in eight SEEG sessions
This study uses its verified joint voltage products, candidate tables, source metadata, and retrieval code.
- usesPair selection at 200 to 500 ms did not show a clear SEEG AUC gain with five calibration pulses in eight sessions
The method lineage supplied the signed-template comparison on energy-selected pairs. This study preserved that score and changed the permitted samples for pair selection.
- usesPacked SEEG signs reduced float32 waveform bytes by 96.9 percent and mean response AUC by 0.028 in eight sessions at 20 to 500 ms
The full-voltage signed template supplies the main reader comparison.