sciscape Publications
← All publications
eess.SP · Signal Processing

Three references gave 4.58 percentage points less joint 45-degree agreement than one for nominal 50 ms forecasts with a fixed gate in 36 SEEG patients

We tested how a fixed phase gate depends on the offline reference rule. We used two public interictal records and four selected bipolar pairs from each of 36 SEEG patients in OpenNeuro ds004100 version 1.1.3. The readers were frozen. Each forecast used the prior 500 ms of voltage and a nominal 50 ms horizon. File A set a vector-length threshold. File B supplied 80,640 possible forecasts. Three centered complex filters used the same nominal 4 to 10 Hz band and spans of 1, 2, and 4 s. The 2 s filter was the original target. The gate accepted 26.05% of opportunities. Equal-patient mean agreement within 45 degrees was 65.97% for the original rule and 61.39% for all three rules together. The paired decrease was 4.58 percentage points. Its descriptive 95% patient bootstrap interval was 3.92 to 5.31 points. Every patient had some decrease. Joint successful yield was 15.56% of all opportunities. The joint rule is stricter by definition. This audit measures the size of the difference. It does not establish true brain phase or a stimulation benefit. The cohort has prior use. The package preserves native selected samples, frozen weights, source versions and hashes, a retrieval recipe, code, all opportunity scores, and three figures. Fresh source checks, exact reader refits, and full result replay passed.

phase estimation · signal processing · SEEG · brain-computer interface · reference sensitivity · forecast evaluation

Admitted 2 citing publications

References

9
  1. usesComplex voltage output length gave higher mean AUC than unit vector length for 50 ms SEEG phase success across files in 36 selected patients with one offline reference sci:sha256:0e0e404b2db82aa434d8772129a46fe00be8159709138b478cc897f4818b93bb

    Uses all 72 selected native input files, all 36 frozen model files, the reader feature and prediction code, the native EDF decoder, and the pinned source receipts. The new audit changes the reference bank and uses a fixed gate from file A. It preserves the source record and verifies its input and model hashes.

  2. Complex voltage targets reduced mean 50 ms offline theta phase error across files from 36 selected SEEG patients sci:sha256:7bb26e9262cd7cf318aacf4eb9c26e67d5a3c1bf62ddd55c19ff2723bdfab146

    [Legacy relation: reproduces] Repeats the inherited reader fitting procedure from file A and exactly reproduces every array in all 36 archived models, including the complex-voltage readers used in the new audit. This is a code and model reproduction check on reused inputs, not an independent cohort replication.

  3. usesMidpoint targets reduced the larger error across two theta phase references for 50 ms forecasts in seven SEEG rest records sci:sha256:aef5b9e83460568eced7559fb6865e75473c0c83ac81308cc38127f384a0b05b

    Uses the earlier two-reference forecast result as scientific context for reference-aware assessment. The new study fits no new reader and audits three fixed filter spans. No code, weights, or samples from the earlier midpoint-target study are reused.

  4. usesA gate based on past voltage raised 4–8 versus 4–10 Hz phase agreement within 45 degrees by 1.73 percentage points in seven SEEG rest records sci:sha256:c15fbdcbcde606709880a1857d44e91cb79d209e5083507565845eac1893de15

    Uses the prior two-band gate result as scientific context for phase-reference agreement and the need to report the full opportunity denominator. The new study uses frozen forecast vector lengths and three filter spans on ds004100. It reuses no code or samples from the seven-record study.

  5. usesA one-second shift of a three-second window changed 4.56% of offline 4 to 8 Hz Fourier phase labels by more than 45 degrees in 21 SEEG patients sci:sha256:6f18f30b30d900a7520c7a81d24b8edfce789125ddc4f293fc6acec50a4b06e6

    Uses the prior shifted-window phase result as scientific context for reference dependence. The new study tests three same-nominal-band FIR spans in a different 36-patient cohort. It reuses no code or samples from that context study.

  6. usesdoi:10.18112/openneuro.ds004100.v1.1.3

    The tested native SEEG samples come from this exact OpenNeuro release; the cached release metadata records its CC0 license.

  7. doi:10.1016/j.neuroimage.2020.116761
  8. doi:10.7554/eLife.68803
  9. doi:10.1038/s41467-025-59417-7

Cited by

2 publications

Public catalog entries on Sciscape.