Five fit pulses and three calibration pulses gave 75.6% mean patient and 70.3% pooled whole response coverage at 86 selected SEEG sites
This study tested complete centered SEEG response bands at 86 fixed sites in eight patients. Five pulses fit the mean and scale. Three separate pulses set a split conformal band. On 354 later pulses, the band gave 75.6% whole response coverage under equal patient weight and 70.3% pooled coverage. The mean paired patient width ratio was 0.664 against a Student t band with a Bonferroni correction. This is a 33.6% reduction. The comparison band gave 91.6% whole response coverage under equal patient weight. The conformal rank rule targets 75% marginal coverage only under exchangeability. With three calibration pulses, a finite 90% band is unavailable under this rule; at least nine calibration pulses are needed. The package preserves all selected voltages, source hashes, code, figures, exact rank checks, and a tool that exports frozen bounds. This offline test does not establish performance in new patients or a live interface.
split conformal prediction · prediction bands · stereoelectroencephalography · electrical stimulation · whole response coverage · calibration budget
References
7- usesStudent t limits scored worse than normal limits for centered SEEG voltage at 86 selected sites after five fit pulses sci:sha256:012478d084fb27a8d92d12089b83be731e42bdd06cd15828d578bf1698fc0e58
The difference between sample coverage and whole response coverage in this earlier same-cohort test supplied the research question. This study tests a separate calibration set and a maximum error score.
- usesA fixed pre-pulse RMS rule increased mean SEEG interval scores after five fit pulses at 86 selected sites sci:sha256:500ee3502e00248d440a6ba1a465c4f3e91b148e8390a72ae1979088798e1843
This study uses all preserved voltage windows, pulse mappings, fixed readout choices, and source checks. It splits those windows into five fit pulses, three calibration pulses, and later test pulses.
- usesFive calibration pulses gave mean later signed template AUC of 0.939 compared with 0.859 for one pulse in eight SEEG sessions sci:sha256:1d2aa103275ad7400befec4c899740ede8f2eb59a09897bdc3f4d4e190e54418
This study keeps the readout choices from the five pulse selection rule. The package preserves the choice table, candidate edge tables, and selection code for source inspection.
- usesA fixed background power rule increased 20 to 500 ms SEEG size forecast error after five calibration pulses in eight sessions sci:sha256:b8e5ae61a5b00cfa06f9294e3694b2df810dda59e3ec09aa69ad2cbe63b674c4
The source download and extraction recipes adapt the preserved code from this publication. They check pinned full file hashes and can rebuild the selected voltage windows.
- 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
The earlier RMS rank study supplied related same-cohort evidence and the finite band budget comparison. This study applies the established rank rule to every time sample of one fixed readout, with five fit pulses and three separate calibration pulses.
- usesdoi:10.18112/openneuro.ds004696.v1.0.1
The study analyzes selected voltage windows and metadata from OpenNeuro ds004696 version 1.0.1.
- usesdoi:10.1523/JNEUROSCI.2201-22.2023
This paper describes the source experiment that produced the analyzed recordings.
Cited by
3 publicationsPublic catalog entries on Sciscape.
- usesEight donor shapes gave less forecast error and 0.002 lower mean response AUC than eight cosines after five fit pulses at 86 SEEG sites in eight sessions
This work supplied the selected pulse voltages and mapping records preserved in the direct input.
- usesThree calibration pulses gave 77.7 percent mean patient coverage for later 20 to 500 ms SEEG responses at 86 selected sites in eight patients
This work motivated a low-calibration whole-response test in the same cohort. It uses a separate three-score sample band after a five-pulse fit. The new study instead fits a full vector rank set to three pulses after separate readout selection.
- usesEight donor shapes reduced five-pulse SEEG forecast error versus eight cosines at 86 selected sites in eight patients; full waveforms had less error
This publication supplied the selected voltage windows, pulse maps, source metadata, and fixed choices for 86 readouts. We used the first five pulses to fit and the final five pulses to test a new patient-excluded shape forecast.