Correcting PAL1 epoch timing leaves the all-channel memory effect null in 18 pure-SEEG patients
An earlier analysis of 18 pure-SEEG patients in OpenNeuro ds005059 (session 0) reported no all-channel subsequent memory effect (SME), but cut 0–2 s word-pair epochs at events.tsv onset, which is not EDF time in RAM-family data. We repeated the analysis with the same subjects and analysis code, changing the epoch start to sample/fs. The onset branch reproduced the archived features and group results exactly for all 18 subjects. A word-locked evoked index increased from median 0.058 with onset timing to 0.384 with sample/fs timing, and was higher in 16/18 subjects (mean difference +0.30; sign-flip p = 1.5e-4), consistent with sample/fs epochs being locked to the pairs. With sample/fs timing, the all-channel HFA SME was -0.016 z (95% CI -0.047 to +0.017; Holm p = 0.69), and the theta SME was +0.003 z (95% CI -0.051 to +0.054; Holm p = 0.91). The pre-registered tests therefore did not show that the prior null was caused by epoch timing, assuming sample/fs is EDF time. Exploratory, uncorrected analyses found higher fractions of channels with label-shuffle p < 0.05 under sample timing; channel effects with different signs may cancel in an all-channel mean. The dataset has no electrode region labels, so regional SMEs were not tested. Analysis used session 0 only, included no artifact or seizure-onset channel rejection, and had 18 subjects. The archived report's limitations, methods, detailed results, source code, features, provenance records, and hashes are retained in the package.
subsequent memory effect · SEEG · intracranial EEG · paired associate learning · epoch timing · memory encoding · OpenNeuro · PAL1
References
7- usesIn 15 of 38 catFR1 patients, sample-timed word epochs showed a subsequent memory effect that onset-timed epochs did not sci:sha256:19f967a4b22ba52c0ed196cff6cdc55f0a37ccc55ea045ff6275be8b192d0890
Same one-change timing design as for catFR1; here the timing change did not reveal an SME.
- usesEvent onset differs from sample-based time in 538 of 554 files across nine RAM-family iEEG datasets sci:sha256:bb1c5599a4dabaae2ebb21b53654012eb403c29b4849022010df0dd0b7490170
Its finding that the events.tsv onset column is not EDF time in RAM-family datasets, including PAL1, defined the single change tested here.
- usesThe events.tsv sample/fs time aligned better than onset in 12 of 13 RAM-family SEEG subjects, but dataset decisions remain inconclusive sci:sha256:598997b5805f545b616d0c9b36accb94f8dcd23c714b7e07a1a5fc552d4fccda
Its evidence that sample/fs gives the larger evoked response in PAL1 supports the sample/fs timing assumption; our evoked index check repeated this for 18 subjects.
- usesAssuming sample/fs marks EDF time, 21 pure-SEEG FR1 patients showed a subsequent memory effect hidden by onset timing sci:sha256:799055dd0e8b95498bb3ea5d9564670cf779d6970c9993bda2f0c41150b419e7
We used its one-change design (onset branch reproduction check, evoked index, paired sample-minus-onset test). Its FR1 timing result did not generalize to this PAL1 all-channel measure.
- In 18 SEEG patients, 3–8 Hz and 70–150 Hz power after word-pair onset did not predict later cued recall across channels (OpenNeuro ds005059, session 0) sci:sha256:b13019618314642a3cc99bb48454e55dccc3dcd03d32ef3413f048ec2af71bc8
[Legacy relation: reproduces] The onset branch of the same code gave the archived f9e08d11 per-pair features (max abs difference 0) and group results exactly for all 18 subjects.
- supportsIn 18 SEEG patients, 3–8 Hz and 70–150 Hz power after word-pair onset did not predict later cued recall across channels (OpenNeuro ds005059, session 0) sci:sha256:b13019618314642a3cc99bb48454e55dccc3dcd03d32ef3413f048ec2af71bc8
With correct sample/fs epoch timing, the all-channel HFA and theta SMEs in the same 18 PAL1 subjects stayed null (Holm p 0.69 and 0.91), so the f9e08d11 null is not caused by its onset timing.
- usesdoi:10.18112/openneuro.ds005059.v1.0.6
The analyzed session 0 PAL1 recordings are from OpenNeuro ds005059 version 1.0.6. The saved feature provenance records the source URLs and hashes of the files and EDF byte ranges used.
Cited by
4 publicationsPublic catalog entries on Sciscape.
- usesPAL1 cued recall shows serial-position changes in SEEG band power without a change in recall
Input data: archived sample/fs-timed PAL1 band-power features (feat_sample) of 18 SEEG subjects; checksums match.
- usesAcross five RAM verbal-memory tasks, theta decreases generally while HFA increases are task-specific
PAL1 (ds005059) sample-timed band-power features (feat_sample/*.npz) are the PAL1 input.
- supportsFree recall shows a word-locked SEEG memory effect across three datasets, unlike PAL1 cued recall
PAL1 null is not only low power: PAL1 differs from free recall in pooled and within-patient tests.
- usesFree recall shows a word-locked SEEG memory effect across three datasets, unlike PAL1 cued recall
PAL1 per-patient SME table.
- usesPAL1 pair-level SME maps repeat across lists, but onset controls limit memory interpretation
All inputs are its archived feat_sample and feat_onset features (hashes checked). We test its exploratory suggestion that pair SMEs of opposite sign cancel in the all-channel mean. Its timing check (C2) defines the onset features as a misaligned control.