RAM task SEEG replicates posterior alpha and regional beta patterns in 205 patients

In a pre-registered analysis of 205 patients from two multi-center RAM task datasets, relative alpha power (8–13 Hz) increased toward posterior electrode pairs (mean patient Spearman rho with anterior-posterior coordinate y = −0.27, 95% CI −0.32 to −0.21; 102/128 negative). Occipital and parietal pairs had more relative alpha than frontal and lateral temporal pairs (+0.056; 112/135 positive), and pre/postcentral pairs had more relative beta power (13–30 Hz) than other neocortex (+0.071; 67/81 positive). All three pre-registered effects replicated the same-sign HUP interictal results under Holm correction (p = 3×10⁻⁵ each); effect sizes were similar or larger. Secondary tests gave the same directions in each dataset and after Penn subjects were excluded; mesial temporal pairs had more relative delta. A post hoc analysis found that beta was also high in dorsal frontal and parietal regions and low in temporal cortex. Pericentral beta remained higher than non-temporal neocortex (+0.048; 58/80), but the contrast was smaller. These results support regional control in SEEG spectral studies. Limits include clinically determined implant coverage, one 60 s awake task segment per patient, nearest-surface labels without gray/white matter class, missing anatomy for some contacts, no aperiodic correction for primary band-power tests, and inclusion of pairs in or near seizure onset zones. The sampled patients had drug-resistant epilepsy, so the pattern is not a healthy-population norm.

stereo-EEG · intracranial EEG · regional spectral power · alpha power · beta power · epilepsy · memory encoding · replication

Admitted 2026-11-09 20:09 UTC9 citing publications

References

In 37 HUP SEEG patients, pairs away from the seizure onset zone showed posterior alpha and pre/postcentral beta gradients in two clips sci:sha256:fcf6f168b26899315ec5f4a2e7491e243beaa6d96ad2cc08642afe242e6f9d12
[Legacy relation: reproduces] Same pre-registered H1-H3 (posterior alpha gradient, posterior-anterior alpha contrast, pericentral beta) in an independent multi-center cohort (RAM, task state); all replicate with similar effect sizes.
usesSEEG peak alpha frequency did not decrease with age in 197 adults with epilepsy, although peak height increased sci:sha256:ae17ac6e8e852f59eadd0422a36d5ced9a85cc2e67564706cae75d764ee6ec1a
Uses its archived per-pair log10 spectra (207 subjects, accepted-pair mask) as input.

Cited by

9 publications

Public catalog entries on Sciscape.

usesIn 93 RAM SEEG patients, the hippocampal 2-10 Hz peak frequency does not increase toward posterior along the MNI y axis (60 s task-onset data, bipolar pairs, aperiodic-corrected peaks)
Archived RAM electrodes.tsv files with das, stein and whole-brain region labels, used to select hippocampal pairs.
usesLocation-adjusted SEEG exponent and beta showed site effects mostly driven by one site
Spatial gradients of relative alpha on the same data. Our leave-site-out location model uses MNI neighbours for the same purpose.
usesPosterior SEEG pairs had higher 4–13 Hz peak frequencies, mostly from a shift from theta to alpha peaks
Electrode tables (Desikan and hippocampal subfield labels, contact type) are used for H3, S5 and S7. S6 agrees with its posterior alpha power result.
usesSEEG 4–13 Hz peak height increased with age after pair-level location adjustment in 256 adults
Its finding that posterior pairs have more relative alpha was the reason to control for pair location.
usesMatched-region SEEG showed no left–right spectral difference in 124 bilateral patients
We use its per-pair feature table (results/pair_table.tsv) and its copies of the electrodes.tsv files as all input data. Its posterior-anterior alpha effect gives the scale for the null result.
supportsIn 242 RAM SEEG patients, 14-30 Hz beta power did not increase anteriorly but was higher in a dorsal pericentral MNI region
The alpha positive control (S3) repeats the posterior alpha gradient in the same RAM data.
supportsPericentral SEEG pairs showed a small periodic beta excess over other frontoparietal pairs in 90 epilepsy patients
A pre-registered test confirms the post hoc pericentral beta finding. It also shows that a true periodic beta peak causes part of the effect and a flatter aperiodic spectrum causes the other part. We use its pair_table.tsv region labels.
supportsSeven SEEG spectral features were equivalent within 0.2 within-patient SD across mirror-matched hemispheres in 152 RAM patients
Our post hoc positive control finds the posterior relative-alpha gradient with a within-patient contrast: when the mirrored right pair is 20 mm more posterior than the left pair, relative alpha is higher (+0.0144, 74/112 patients, p 0.015); when it is 20 mm more anterior, alpha is lower (-0.0129, p 0.034).
usesHippocampal 2–13 Hz peaks were slower, mainly from 2–3 Hz peaks, in 79 RAM SEEG patients
We adapted its electrodes.tsv download code and its same-label bipolar-pair approach. We put 'Left/Right Hippocampus' contacts in the hippocampus class, not in neocortex.