A screen for large signal changes accepted constant blocks of 20 samples during marked stimulation in five of 11 ds005558 SEEG recordings
This audit tested whether an upper signal-change screen accepts exact constant SEEG blocks during marked stimulation. It used 88 selected marked trains and 88 control word windows from 11 bipolar recordings in seven people in OpenNeuro ds005558 version 1.0.0. The screen flagged RMS sample changes above five times a preceding baseline. Each constant 20-sample block had zero within-block change and passed that screen. Five recordings had constant blocks. The main effect was in both R1286J recordings, where the constant rule added flags for 29.974% and 58.618% of the selected channel blocks. No selected control window had a constant block. A fresh source extension found a block constant in more than half the channels in 134 of 134 marked trains from R1286J, including 118 of 118 additional trains. The extension tests one person. It does not establish prevalence in new people. Fresh pinned source ranges matched one inherited excerpt per recording and all overlapping extension excerpts. The package contains integer excerpts, source versions and hashes, event tables, five figures, analysis code, a reusable constant-block check, and instructions. The finding concerns stored bipolar signals. It does not prove a hardware fault or improved decoding.
SEEG · intracranial electrophysiology · stimulation artifact · signal quality · constant blocks · OpenNeuro
References
3- usesSample times locate stimulation bursts in 11 SEEG recordings from OpenNeuro ds005558 version 1.0.0 sci:sha256:df1ea802c869abaea97a7b95bc606cc84b978f96243f31357ef16d5b3c852f96
This timing study supplied the integer SEEG excerpts, sample-clock result, pinned source versions and range hashes, source metadata, and retrieval code. The present study tests exact constant blocks within the marked stimulation intervals and checks fresh source data.
- usesA power check accepted 353 pair windows with one flat input in 30-second samples from each of 35 SEEG recordings sci:sha256:3035004787267bb786bc804fa124723640a616704198289ab05564dd44a3a388
This study distinguished positive bipolar signal power from contact input validity in ds005624. That distinction informed the present question about whether a screen for large signal changes can accept exact constant bipolar output blocks during marked stimulation in ds005558.
- usesdoi:10.18112/openneuro.ds005558.v1.0.0
Source dataset for the stimulation event metadata and recorded bipolar SEEG data used in this audit.
Cited by
4 publicationsPublic catalog entries on Sciscape.
- usesWidespread constant SEEG values enter the next marked word interval in one of two ds005558 sessions from R1286J
This study supplied archived source metadata, EDF headers, pinned source object versions, and short integer excerpts. All 134 stimulation excerpts match those excerpts over the shared interval. We retrieved new longer windows and tested overlap with the next stored WORD interval.
- usesA check of the newest 11 SEEG samples accepts 500 ms windows with constant runs after stimulation in ds005558
Its result showed that a signal-change screen can accept constant blocks. That failure helped define our question about complete feature inputs. We did not use its screen as the tested decision rule.
- usesStored SEEG code runs of at least 1 ms covered less than half the channels at once between the pre and post labels in all 35 ds006519 v1.0.0 records
This publication motivates an exact constant-code check by showing a limit of a large-change screen in a different SEEG dataset. We test maximal native float32 code runs in ds006519. No signal arrays or numerical results from that study are reused.
- usesWidespread constant SEEG values enter the next marked word interval in one of two ds005558 sessions from R1286J
This study supplied archived source metadata, EDF headers, pinned source object versions, and short integer excerpts. All 134 stimulation excerpts match those excerpts over the shared interval. We retrieved new longer windows and tested overlap with the next stored WORD interval.