Joint limits covered 92.6 percent of later nominal 70 to 230 ms SEEG maps with 19 calibration pulses in six people
We tested limits for complete stimulation response maps in public OpenNeuro ds006065 version 1.0.0. The sample had six people and eight fixed-site post-block exports. Each block used 10 fitting pulses, 19 calibration pulses, and all later pulses. At nominal 70 to 230 ms, joint limits covered 92.6 percent of later maps with equal person weights. Separate limits covered 95.1 percent of individual values but 1.5 percent of whole maps. Their mean scaled width was 23.8 percent of joint width. Random calibration partitions averaged 95.0 percent joint coverage. The stored result was within their spread. Known rank and tolerance laws show that 19 calibration pulses give 95 percent marginal coverage under exchangeability. Under independent continuous scores, a maximum-score limit needs 59 calibration pulses for at least 95 percent population coverage with 95 percent confidence. The dataset cannot test that larger budget. Acquisition order and real-time operation are not verified. The package includes four figures, source and array hashes, compact data, code, all test decisions, and reproduction instructions.
split conformal prediction · joint coverage · calibration budget · stereoelectroencephalography · brain-computer interface · prediction limits
References
7- usesIn stored order, a running mean reduced 70 to 230 ms SEEG forecast error by 4.9 percent against a fixed mean in six people sci:sha256:b6b028713157225b96536418641eda3c5d487a3f86ef36c8891f61613192cc2a
This study uses its eight fixed-site post-block compact exports, source manifest, metadata, and timing audit. New endpoints measure joint response map coverage and calibration variation. All inputs were checked against fresh pinned public bytes.
- usesWith 10 calibration trials, amplitude updates had lower error than waveform replacement at fixed SEEG pairs in six people sci:sha256:5fbf5699f54eed2770f2feed9404402fd27dc4a13216f041c985185e72ab0782
This study uses the original pinned source manifest through the parent package. Its source hashes, S3 versions, and contact rules enable exact source checks.
- usesSeparate pair bounds covered 71.5 percent of individual SEEG responses and 17.0 percent of whole pulses at 20 to 500 ms in eight sessions sci:sha256:1a0ecb0d254415256c85ec04c2b9bce25a0f6893277ff47091e5d6c673436869
This prior result motivated the distinction between separate and whole-pulse bounds. It used three calibration pulses in a different SEEG dataset. This study tests 19 calibration pulses for pair-by-time maps and supplies a 95 percent calibration-budget audit.
- usesCross-group tests did not establish SEEG response map change across six people sci:sha256:c1ee3af2654efdf1006112a9bbce8e06fd2be8451b3d580e001e611877e2e45f
This study uses its original compact extraction method and retrieval code through the parent package. The extraction defines bipolar pairs, baseline, polarity, and nominal bins.
- usesdoi:10.18112/openneuro.ds006065.v1.0.0
Source data used in this secondary analysis; exact dataset version 1.0.0 is identified in the saved source metadata.
- doi:10.1038/s41467-025-59417-7
- usesdoi:10.5281/zenodo.14735080
The original MATLAB archive was consulted for source timing, polarity, trial counts, and stimulation-contact metadata; it was not executed and is linked externally under GPL-3.0.
Cited by
1 publicationPublic catalog entries on Sciscape.
- supersedesJoint limits covered 92.6 percent of later nominal 70 to 230 ms SEEG maps with 19 calibration pulses in six people
This version states the full distribution assumption in the summary and opening description. The 59-pulse budget requires independent scores from one continuous distribution, conditional on a separate fit.