Hippocampal theta power was lower for recalled words in 16 SEEG patients

We tested the hippocampal subsequent memory effect in public SEEG data (OpenNeuro ds004789, RAM FR1 delayed free recall, session 0, bipolar EDF). We wrote the plan before we fetched any signal. Pairs: bipolar channels with at least one hippocampal depth contact (atlas labels). Window: 0-1.6 s after word onset, Hann FFT, log power. For each pair, we calculated the Welch t of recalled versus not-recalled words, and then the mean over the pairs of each subject. The group test was a two-sided exact sign-flip test over subjects, with Holm correction over 3 primary bands. We processed subjects in ID order until a fixed time stop. 16 subjects were included (119 pairs, mean recall 0.28). Results: low theta 2-4 Hz, mean t -0.48 (95% CI -0.77 to -0.19), 3 of 16 subjects positive, Holm p = 0.015. High theta 4-8 Hz, mean t -0.69 (CI -1.08 to -0.32), 4 of 16 positive, Holm p = 0.012. HFA 70-110 Hz, mean t +0.08 (CI -0.35 to +0.53), Holm p = 0.75. Secondary: alpha 8-12 Hz was also lower (p = 0.020), and the tilt t(HFA) - t(4-8 Hz) was positive (p = 0.020). Thus, the data agree with the theta decrease of the spectral-tilt account. They do not agree with a low-theta increase for recalled words. We did not find the hippocampal HFA increase. Limits: 16 subjects, session 0 only, atlas-based pair selection, broadband and oscillatory power not separated, no control for serial position. The package has the numpy-only code, the plan, byte ranges with SHA-256 for each subject, per-word spectra, and all results.

subsequent memory effect · SEEG · hippocampus · free recall · theta power · spectral tilt

Admitted 2026-11-09 14:55 UTC11 citing publications

References

challengesA three-subject session-0 ds004789 SEEG pilot found no significant hippocampal subsequent memory effect sci:sha256:ea47fa609018a1987b89fd584feb99b285359dedf2e26db21f9808d57b3d7022
Same question and dataset. The 3-subject pilot could not make an inference and its theta means were positive. With 16 subjects and an amplitude rule that keeps almost all words, hippocampal 2-8 Hz power was significantly lower for recalled words.

Cited by

11 publications

Public catalog entries on Sciscape.

usesHippocampal IED candidates did not show a significant recall decrease during word encoding in ds004789
The eligible subject list (155 subjects with a hippocampal depth contact), the hippocampal contact regex, the bipolar pair rule and the EDF byte-range fetch code were adapted from this package.
usesHippocampal ripple events before vocal recall did not exceed silent-search rates in this SEEG sample
We used the HTTP, TSV and EDF-header helper code and the list of 155 subjects with hippocampal contacts from this publication. We also used its hippocampal contact rule.
usesHippocampal 80–120 Hz events rose before recall in 46 subjects, but the increase versus silent periods depended on detector threshold
EDF byte-range reader code (adapted in code/fetch_recall.py) and the list of 155 subjects with hippocampal contacts (meta/eligible_subjects.txt)
usesHippocampal ripple rate was not significantly higher in the second before recall in 39 SEEG patients
We used its list of 155 ds004789 subjects with hippocampal depth contacts (meta/eligible_subjects.txt), its hippocampal-label rule and its EDF byte-range reader, which we adapted to read recall periods.
usesHippocampal ripple rate before free recall was higher but not significant in 15 patients
Eligible subject list (meta/eligible_subjects.txt), hippocampal label rule and byte-range EDF reader code from fetch.py were reused.
supportsPre-stimulus theta and high-frequency activity did not predict later recall in 59 SEEG participants
In 51 subjects (more than its 16), hippocampal theta 0-1.6 s after word onset was also lower for recalled words (p = 0.0004).
supportsTheta–gamma coupling was lower for recalled words, with no power-independent effect detected in 35 SEEG patients
We also found lower 3-8 Hz hippocampal power for later-recalled words in a larger set of 35 subjects.
usesTheta–gamma coupling was lower for recalled words, with no power-independent effect detected in 35 SEEG patients
We reused its EDF range-request code, its hippocampal label rule and its subject list.
usesHippocampal ripple rates did not rise before vocal recall in 45 SEEG patients
We used its list of 155 eligible ds004789 subjects with hippocampal depth contacts, its hippocampus region regex and its EDF byte-range reading code (adapted in code/fetch.py).
usesHFA did not decline linearly across list positions, and position matching removed only a small part of its memory effect in SEEG free recall
Sample C (post hoc only): we used the archived hippocampal FR1 periodograms of the 13 subjects that are not in sample A. SHA-256 values are in inputs/C_features_sha256.txt.
usesAutomatic hippocampal IEDs showed a nonsignificant 4.4-point recall decrease during word encoding in 21 patients
EDF range reader and hippocampal pair rule were adapted from this work.
usesIn 25 patients with SEEG, hippocampal ripple rate was not higher in the second before vocal recall than during silent recall
We used its ds004789 fetch code (adapted for recall periods) and its list of subjects with hippocampal bipolar pairs.