A sign flip test did not detect a late high gamma memory load association in 12 patients across two SEEG sessions

We tested late hippocampal high gamma power during verbal memory maintenance in 12 patients, 44 matched contact pairs, and two SEEG task sessions from OpenNeuro ds004752 version 1.0.1. The model used 983 correct trials without a source artifact flag and adjusted for trial order and previous error. The mean patient coefficient was −0.105 dB per two letters (95% t interval −0.352 to 0.143; exact two-sided sign flip p = 0.3755). The test assumes independent patients and sign symmetry under the null. It did not detect a common load association in these selected patients, sessions, contacts, band, and model; it does not establish that the hippocampus has no role in memory or that no memory effect exists. Session coefficients had a descriptive correlation of 0.694. All planned secondary and sensitivity results are reported. Fresh retrieval checked all 1,167 selected source segments; all 26 scientific output files matched after fresh retrieval and a clean offline repeat. The package preserves spectra, raw examples, code, provenance, and repeat instructions.

hippocampus · verbal working memory · high gamma power · stereo-EEG · memory load · sign-flip test · OpenNeuro

Admitted 2026-11-09 16:55 UTC6 citing publications

References

usesLate bipolar hippocampal theta load slopes had a correlation of 0.151 between the first two task sessions in 12 SEEG patients sci:sha256:a924ce7fee66f1151c8a74b7eade8acb9d23fd494af3d197445d7b53543d02dd
This publication supplied second-session digital segments, matched pair selection, metadata, timing code, and inherited complete-file checks. We added high gamma power measurements and checked every selected source segment again.
usesThe primary sign flip test did not detect a load association with 15–40 Hz hippocampal slope in 12 first SEEG sessions sci:sha256:773de7c1d46896cf64e4d6113e0ea2cc935888fec848334a2e1212a367db9af6
This publication supplied full-rate first-session late-maintenance digital segments and fixed source range records. We used them for initial high gamma extraction and compared all raw segment hashes with fresh OpenNeuro data.
usesdoi:10.18112/openneuro.ds004752.v1.0.1
The analysis used OpenNeuro ds004752 version 1.0.1; its saved metadata records this exact version DOI and CC0 license.
supportsdoi:10.1523/JNEUROSCI.0567-09.2010
Cited in the publication as evidence for an earlier 48–90 Hz hippocampal power association with memory load in another clinical sample.
supportsdoi:10.7554/eLife.78677
Cited for the verbal working-memory task timing and the rule that an incorrect response was followed by a four-letter trial.

Cited by

6 publications

Public catalog entries on Sciscape.

usesUnder patient independence and sign symmetry, testing did not detect a late theta spread difference between four and eight letters in 12 SEEG patients
This publication supplied all-trial spectra, source metadata, raw examples, fixed object records, and retrieval code. They enabled the new trial-spread analysis. The complete-source checks and all-trial replay records remain inherited evidence.
usesA sign flip test did not detect a mean load association in late hippocampal theta excess across two SEEG sessions in 12 patients
This publication supplied all 24 calibrated spectral files, source metadata, headers, source range hashes, and Python retrieval and extraction code. We used its fixed 44 matched hippocampal pairs in 12 patients. Fresh source downloads and reconstruction checked all 1,167 selected trials.
usesA sign flip test did not detect a fixation theta slope for previous exported trial load in 12 SEEG patients across two sessions
This archive supplied the matched two-session SEEG pair selection, source metadata, EDF headers, pinned object references, and late-maintenance spectra. We retrieved new fixation signals and tested the preceding exported trial load. We verified inherited files and imported late spectra against its manifest.
usesTesting did not detect a linear beta load association in 8 patients with 19 later hippocampal SEEG sessions, assuming independent patients and symmetric null signs
This work supplied matched pair identities, first-two-session spectra, metadata, source timing code, and verification records. Its nominal high gamma minus beta contrast motivated the fixed beta question. The present work tests later recordings from the same patients.
usesAdding late hippocampal theta power increased mean log loss by 0.0104 in fixed models tested across two sessions from 12 patients
This study uses the all-trial compact spectra, matched 12-patient sample, source metadata, raw examples, source record hashes, and MIT decoding and retrieval code. A fresh download and separate Fourier code checked every selected signal trial. The new outcome is an incorrect response and the new models are fixed ridge logistic models.
usesA sign flip test under sign symmetry did not detect different hippocampal alpha load slopes during encoding and late maintenance in 12 independent SEEG patients
This source supplied matched contact pairs, source metadata, EDF headers, fixed object records, and maintenance sample hashes. We added encoding windows and tested the paired stage contrast. We retrieved all selected source windows again.