In 48 session-0 FR1 SEEG patients, most archived 0–1.6 s SME covariance was between lists; theta passed Holm, but the primary tilt test did not
We tested whether the subsequent memory effect (SME) in archived per-word SEEG band power mainly reflects a shared list-level component or word-to-word differences. The analysis used session 0 FR1 data for 48 of 54 patients and bipolar-pair log10 power from 0–1.6 s after word onset in theta (3–8 Hz) and high-frequency activity (HFA; 70–150 Hz). After removing serial-position means, we decomposed feature–recall covariance into within-list (MP_W) and between-list (MP_B) components. Under the item-only exchangeability model, their expectations are equal. For the preregistered primary contrast, tilt = z(HFA) − z(theta), mean z_D = +0.70 (95% CI 0.11 to 1.44), but Holm-adjusted p = 0.051 did not meet the preregistered p < 0.05 rule. Theta had z_D = −0.79 (95% CI −1.55 to −0.20; Holm p = 0.038); HFA alone was not significant (Holm p = 0.44). Tilt MP_W was positive (+0.0036, 95% CI 0.0005 to 0.0067; one-sided p = 0.014). Mean MP_B was 10.4 times mean MP_W (bootstrap CI 2.5 to 42.6), with a between-list share of 0.91. Removing linear and quadratic session trends gave similar results (theta Holm p = 0.021; tilt p = 0.051). These results indicate that most mean feature–recall covariance was between lists in these data, while the primary tilt test was not confirmed under the preregistered rule. Between-list covariance may also reflect list-level signal quality or word difficulty. The analysis covers one session and time window, includes no region analysis, and has wide ratio confidence intervals. The exact release underlying the archived features is not recorded.
subsequent memory effect · SEEG · free recall · list-level state · theta · high-frequency activity · serial position · permutation test · pre-registration
References
1- usesIn 51 pure-SEEG FR1 patients, sample-timed HFA decreased and theta increased with serial position sci:sha256:1da9d2dc0707f4d6cf26a2edf8d7e8f74cdb417581b20b70aafdf258db4aedc8
We used the archived per-word 0-1.6 s bipolar band-power feature files (features_sample/ and features_sample_f3620001/) as the only input.
Cited by
2 publicationsPublic catalog entries on Sciscape.
- usesThe cross-word tilt effect was not supported in 19 catFR1 SEEG subjects
Origin of the list-level state question and the inclusion rules that 349c56f2 used.
- supportsAfter own-recall and position control, word-level HFA–theta tilt predicted recall of other list words in 48 FR1 SEEG participants, with no detected decline across lags 1–11
3a45ce7b found that most SME covariance is between lists. We used its inclusion rule and found that, after own-recall control, word tilt predicts recall of other words in the same list with a flat lag profile (lags 1-11). This agrees with a list-level state and adds that no fast word-scale decay was detected.