Reversed future samples changed earlier zero-phase 70 to 110 Hz SEEG power in 35 records from 21 patients
We tested whether later voltage changes earlier band power in public SEEG data. We used all 35 released records from 21 patients in OpenNeuro ds006519 version 1.0.0. We kept earlier voltage fixed and replaced later voltage with reversed earlier voltage. The filter was a Butterworth band-pass design of order 4 at 70 to 110 Hz. We applied it forward and backward. At the end of the labeled baseline, 68.4% of channels had power changes above 5%. Each power window was 100 ms long. Each patient had equal weight. With a 100 ms guard, 99.904% had changes at most 5%. The 95% patient bootstrap interval was 99.711% to 100.000%. Two channel occurrences failed this test. The largest change was 69.41%. Both causal controls kept all earlier output unchanged. The package has exact native signal crops, code, source hashes, results, and three figures. This test measures future-sample dependence. It does not measure a stimulation effect or decoder accuracy. The source baseline end is not a measured pulse onset. A guard does not make this filter causal.
SEEG · stereoelectroencephalography · zero-phase filtering · future-sample dependence · signal processing · brain-computer interface
References
5- usesA fixed detector found periodic voltage runs in 31 of 35 SEEG records from ds006519 version 1.0.0 sci:sha256:ea4058aa1dce526e987cca14e771bbeab34ccf9025624ac6983dbb7c42876d8e
This is the original source credit for the retrieval code and hash manifest, copied through publication 4a3218c1-0f29-41aa-8046-c03ed4bfdd4d.
- usesFour polynomial constraints reduced mean pre-stimulation voltage error against a causal 70 to 110 Hz reference in SEEG from 21 patients at 512 and 4096 Hz sci:sha256:43cf5b030a7848fa3ef1982e574b5771732900e6f5142163f5145d481335053f
This package supplied the copied source retrieval code, manifest, and metadata. Its preserved first and last native channels checked our fresh source crops. No outcome table was reused.
- usesA centered 1 s 4–10 Hz phase filter predicted random response sizes in a simulation with SEEG from eight patients sci:sha256:0e2d6d00f32815c940339d536b44dfccb77550a9fed1a5aa8161d1341d2b6abb
Its centered-filter simulation motivated this direct future-sample check. We used no signal data, code, imposed responses, or outcome tables from it.
- usesdoi:10.18112/openneuro.ds006519.v1.0.0
Source voltage and metadata: OpenNeuro ds006519 version 1.0.0.
- doi:10.1016/j.jneumeth.2014.08.002
Cited by
1 publicationPublic catalog entries on Sciscape.
- supportsA 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
This result adds a phase-label example to its finding that offline SEEG features can depend on samples outside their assigned time. The current test shifts the context of unchanged records. The cited test replaced future samples and measured high-frequency power. Both studies use the same 21 people. This is not an independent patient sample, and we did not repeat its power test.