A four-pair SEEG reader kept 73.1% four-site accuracy under 16 ideal recording sign patterns in eight patients
We tested a fixed stimulation site reader in eight public STReEF SEEG patients. Each patient had four sites and four selected recording pairs. Five pulses per site fit the reader. Five separate pulses per site tested it in the 20 to 100 ms window. The reader selected the nearest site mean after it allowed either sign for each recording pair. It identified 117 of 160 test pulses (73.125%). All its scores and decisions stayed unchanged under all 16 ideal pair sign patterns. The signed reader identified 115 pulses (71.875%). The mean paired difference was +1.25 percentage points, with a descriptive 95% patient interval from -3.750 to +6.875. An amplitude-only reader also identified 117 pulses and ignored sign changes. The data do not show an accuracy gain from waveform shape. We rebuilt all 320 public voltage ranges and matched all features and numerical results. The package includes code, compact data, figures, source records, a proof, and repeat instructions. The test was offline and retrospective. It did not test a live BCI or noise and contact failure.
stereoelectroencephalography · brain-computer interface · bipolar recording · sign invariance · site classification · signal processing · STReEF
References
7- supersedesA four-pair SEEG reader kept 73.1% four-site accuracy under 16 ideal recording sign patterns in eight patients sci:sha256:b91eaa99399d155244986c772adb8a226f468dcd8af46bf0bac0923daf668f8a
This version fixes the package file check after sci retrieval. The check ignores the metadata directory added by sci fetch. The source data, reader code, fit states, predictions, scores, figures, and scientific results are unchanged.
- usesThe nearest half of SEEG pairs had a higher mean rate of repeatable waveforms across recorded contact order reversal in eight STReEF patients sci:sha256:399c1198c4184a27bd9be26923336885a19d4fdd7d9a5d49a4ca8f9cffac5cae
We used its metadata selection and pinned retrieval helpers through the direct input publication. These helpers replayed sample selection and retrieved source metadata and raw ranges. We kept the code license notices.
- usesGlobal amplitude removal lowered mean site accuracy after recorded contact order reversal in eight STReEF patients with four SEEG pairs sci:sha256:84285597246fb5fb9c594f376f9def4dfde8ac5b145d2ad68c66a8f5b3927cf3
We used its compact voltage features, fixed sample, source records, and signed and amplitude-only reader rules. We checked its 320 primary baseline decisions. We retrieved all 320 pinned voltage ranges again and matched all input arrays. The new test adds independent recording pair sign distances and all 16 sign patterns.
- usesSelection for distinct response patterns had an uncertain mean accuracy gain at four SEEG pairs in eight STReEF patients sci:sha256:143d7f91f05a486692072b6c1e61b67efa46729f12681baae0430cec9016d9ad
We used its four-site sample in each of eight STReEF SEEG patients through the direct input publication. The condition order, fit and test pulse groups, and source feature construction remained fixed.
- usesdoi:10.18112/openneuro.ds005448.v1.0.0
Source data: STReEF, OpenNeuro ds005448 version 1.0.0; the saved source metadata records this exact version and declares CC0.
- doi:10.1016/j.ynirp.2025.100274
- doi:10.1007/s12021-022-09567-6
Cited by
4 publicationsPublic catalog entries on Sciscape.
- usesOne pulse at a known site restored signed SEEG reader decisions under fixed output sign reversals in five sessions
Its separate SEEG sign-invariant reader study supplied context for the recording orientation problem. We do not use its dataset or fitted reader. The present study tests recovery of signed scores with one pulse from a known site.
- usesFull covariance gave no mean gain in offline four-site accuracy after stimulation contact order reversal with four SEEG pairs and five fit pulses per site in eight patients
Used its fixed STReEF sample, compact cosine voltage features, raw data recipe, source metadata, and MIT extraction and selection code. Rebuilt all 320 epochs and reproduced all 160 original signed-reader decisions before testing covariance.
- usesMinimum separation selection gave 68.1% later accuracy versus 71.9% for total separation with four SEEG pairs in eight patients
This work supplies the fixed four-site sample, compact features, source metadata, and raw range hashes. We reproduced its 160 primary reference predictions.
- usesPulse time gave 80.8 percent site accuracy after recorded contact order reversal in offline tests of 12 sites in each of eight SEEG patients
We used its four-site feature arrays, pinned source records, voltage reader, and retrieval code for a matched comparison. We retrieved all 320 source epoch ranges again, rebuilt all features exactly, and matched all 160 signed after-pulse predictions and scores. We added schedule controls and a before-pulse reader with independent pair selection.