Exact sample markers reduced expected 70 to 110 Hz SEEG power error under independent uniform dither in 35 records

We tested all 35 released SEEG records from 21 people in OpenNeuro ds006519 version 1.0.0. Six bit packets combined shared subtractive dither with exact float32 exceptions. We derived a power correction that used the exact sample markers. Under independent uniform dither, the correction removed expected quantization bias without extra packet bytes. The primary expected excess normalized squared power error of the all-level correction was 6.204e-6. This score used record channel medians and person means. In 64 numerical repeats, the mask correction reduced the measured post-window score by 3.71%. The packets saved 63.27% of bytes in a person-weighted comparison with equal-window float32 packets. Full source hashes, independent decoding, and a complete numerical replay passed. The package includes the proof, source recipe, code, actual packets, tables, and figures. The result concerns stored 70 to 110 Hz voltage power. It does not establish live BCI performance or a biological stimulation effect.

subtractive dither · SEEG · power estimation · quantization · BCI telemetry · exact exceptions · signal processing

Admitted 2026-11-09 19:18 UTC1 citing publication

References

usesA dither correction reduced six-bit 70 to 110 Hz power error in seven SEEG rest records with fixed voltage limits sci:sha256:af44577f1aaa93d39072929cfbd90a046eea0db4fcfeec522394c48b8ec068d1
The earlier study supplied the subtractive-dither power question. We extended the correction to exact sample exceptions near source stimulation labels. We did not reuse its code or numerical scores.
usesExact exceptions reduced six-bit SEEG power error in post-stimulation windows from 21 people with early fixed ranges sci:sha256:a17682a051300f5700c90c110920c3c21cd784e738434756efb657722a55b62c
We reused the cohort and window design. We copied the retrieval code. We adapted the header parser, frame edges, calibration layout, and bit packing. We did not reuse numerical scores.
usesdoi:10.18112/openneuro.ds006519.v1.0.0
The analysis used all 35 records from this exact OpenNeuro release; the pinned Git commit and complete-file hashes identify the retrieved files.

Cited by

1 publication

Public catalog entries on Sciscape.

usesPower correction raised mean Brier error in eight fixed SEEG state readers with six bit codes and a fixed floor
This method supplied the question and exact-sample mask principle for subtractive dither. We derived the finite Hann periodogram diagonal with mean removal and tested downstream log-power reader error on another dataset. No code or data were copied from this method.