In ds006136 SEEG, aperiodic detrending leaves no 50–100 Hz recall effect; broadband power change remains undetermined
This analysis tested whether lower 50–100 Hz power during the early delay (2700–3500 ms) on full-recall trials reflects a narrowband gamma effect or a broadband power shift. In OpenNeuro ds006136 v1.0.1, bipolar SEEG from MFG and MTG was analyzed for 12 participants (138 channels; MU3 had no channel with at least five trials per class). The previously reported 50–100 Hz effect reproduced (mean AUC 0.473, 95% CI 0.461–0.485, p = 0.0024). After fitting and removing a log-log aperiodic line over 20–190 Hz, the 50–100 Hz residual showed no recall effect (AUC 0.496, p = 0.59); the exponent also showed no effect (AUC 0.493, p = 0.28). Power at 110–190 Hz had the same direction (AUC 0.473, p = 0.017), but did not pass Holm correction (adjusted p = 0.051). Pre-trial features and 300–490 Hz noise-floor power showed no effect. The pre-specified interpretation is undetermined, although the pattern agrees better with a broadband decrease than a narrowband gamma effect. Five MU-cohort participants had high noise floors. Limits include 12 analyzed participants, small trial counts in some channels, no EOG/EMG channels to exclude muscle activity, a single-line aperiodic model whose fit includes 50–100 Hz, correlated exponent and offset, and low power in cohort splits.
SEEG · working memory · recall · aperiodic activity · gamma power · OpenNeuro
References
3- Early-delay gamma power was lower on full-recall trials in 12 SEEG patients sci:sha256:3bfe028f03b12bb14b534ed5fceefa1cad1192a4ab8212271b95c1992c600767
[Legacy relation: reproduces] Reproduces its DEL1 50-100 Hz recall effect with the same data and windows (AUC 0.473, p = 0.0024). It also answers its stated limit (no aperiodic model): no narrowband part remains after aperiodic removal.
- usesEarly-delay gamma power was lower on full-recall trials in 12 SEEG patients sci:sha256:3bfe028f03b12bb14b534ed5fceefa1cad1192a4ab8212271b95c1992c600767
Uses its EDF loader, trial and recall definitions, AUC and sign-flip statistics, versioned URL list and SHA-256 hashes.
- usesdoi:10.18112/openneuro.ds006136.v1.0.1
The analysis used OpenNeuro ds006136 v1.0.1; the archived provenance records versioned file URLs and SHA-256 hashes.
Cited by
6 publicationsPublic catalog entries on Sciscape.
- usesIn 13 SEEG patients, gamma bursts did not increase during encoding or late delay, while beta bursts decreased in middle frontal gyrus
We use its observation that 50-100 Hz power is near the noise floor in the MU cohort to interpret our gamma burst null.
- usesEarly-delay 50–100 Hz power remains lower on full-recall trials after trial-position control in 12 SEEG patients
This publication reports that the effect is not present after aperiodic removal. We use it to state the limit of our test: position control does not separate periodic from aperiodic power.
- supportsFull recall tracks a steeper aperiodic exponent during encoding in 12 OWM SEEG patients
Like its broadband gamma result, the encoding recall effect is mostly a spectral tilt. Here full recall has more low-frequency and less 22.5-57.5 Hz power.
- supportsCortical SEEG shows theta–gamma coupling without delay enhancement or recall differences
That work found no 50-100 Hz recall effect after aperiodic removal. Our amplitude-normalized PAC also shows no full-vs-partial recall difference in 50-100 Hz.
- usesRecall failures clustered in trials, while SEEG power tracked object recall count
We use its result that the early-delay 50-100 Hz effect is aperiodic to limit how we interpret our 50-100 Hz correlations. We did not fit aperiodic components.
- usesThe SEEG aperiodic exponent does not change significantly from fixation to encoding or early delay in 9 ds006136 patients (MFG/MTG bipolar channels, 800 ms windows, 30-48 Hz fit)
We use its broad 20-190 Hz aperiodic fit range as our secondary range. Our test (task epoch vs fixation) is different from its test (exponent vs recall).