A 50 ms reference change reduced median 70–110 Hz filter order error after simulated shaft removal in four SEEG patients

We tested how a changing reference affects a causal SEEG filter. We used 48 pre-pulse epochs from four patients in OpenNeuro ds004977 version 1.2.0. We simulated 648 removals of one shaft from an average reference. The filter passed 70 to 110 Hz. Hard changes caused common output differences between two processing orders. Patient median differences were 7.1 to 12.0 percent of baseline RMS. No difference exceeded baseline RMS in the primary 100 ms window. A 50 ms linear change reduced the patient median paired error by 60.4 to 73.7 percent. It reduced error in 638 cases and increased it in 10 cases. With equal fixed linear channel filters and consistent history, a reference change after filtering avoids this order error. The package has reference inputs, digital readout samples, results, two figures, source hashes, and code. The study does not test decoder accuracy or clinical benefit.

SEEG · causal filtering · common average reference · reference switching · signal processing · intracranial EEG

Admitted 2026-11-09 18:26 UTC4 citing publications

References

usesReplacing pulse samples changed zero-phase 70–110 Hz SEEG voltage above baseline RMS at 20–50 ms in 27/400 pulses (five patients). sci:sha256:78299ccd5493ea3f4d9c8ec1dfad43bf77ec6ab49d467ba2f55a3b7b4806dddd
This work supplied MEF retrieval and extraction patterns and a causal filter audit approach. The new test changes reference weights in a separate four-patient dataset. No prior numerical result or raw epoch was reused.
usesdoi:10.18112/openneuro.ds004977.v1.2.0
Source dataset, exact analyzed version.
doi:10.1016/j.jneumeth.2024.110153

Cited by

4 publications

Public catalog entries on Sciscape.

usesChanging the bipolar montage reversed 71 of 296 shaft score orders in four SEEG pulse maps at 20 to 300 ms with fixed contacts
The fixed event identities and source manifest trace to this source audit through the direct parent. We preserve its event selection and code license. We use the same four patients, not independent patient evidence.
usesFor one faulty shaft, the median of shaft means had a 5.1% higher mean normalized bound in four SEEG records
We used its fixed accepted epochs, contact order, raw check samples, and mean reference traces. These files fixed the signal sample and allowed exact checks of the new extraction.
usesPatient median beta score changes were below 0.006 after removal of one shaft from the average reference in four fixed SEEG records
Reused its 48 accepted pre-pulse epochs and archived check-contact samples. Fresh source extraction exactly checked its references and samples. The new test measures fixed-input beta imaginary coherency and gamma power after one shaft is removed from the average reference.
usesSimulated loss of one reference shaft changed sets of eight SEEG contacts in 6 of 50 tests in four patients
Used the fixed public event subset, source metadata, raw retrieval manifest, and retrieval code. The new study extracted after-pulse voltage from raw MEF files and tested contact-set changes. The two studies use the same four patients.