Paradigm, behavioral performance, and brain oscillations.

(A) Auditory tasks (here with 6-tone sequences, 2 s retention): “Same” trials: After a delay, the first melody was repeated. “Different” trials: One tone was changed in the second melody of the pair in comparison to the first melody (red rectangle). Memory load (3 or 6 tones) and duration of the retention period (2, 4, 8 s) varied in separate blocks. Source data can be found at https://osf.io/m7dta/. (B) Accuracy in terms of d prime presented as a function of the duration of the retention period (left panel; N = 6) and memory load (right panel; N = 6). Colored circles depict participants (one color per participant). Asterisks indicate significance (p < 0.05, nonparametric tests; see text for details); NS, nonsignificant. Source data can be found at https://osf.io/m7dta/. (C) Response time (s) presented as a function of the duration of the retention period (left panel; N = 6) and memory load (right panel; N = 6). Colored circles depict participants (one color per participant; same color coding as in Fig 1B). NS, nonsignificant. Source data can be found at https://osf.io/m7dta/. (D) Left panel: T-values in the time-frequency domain (t test relative to baseline −1,000 to 0 before stimulus onset, FDR corrected in time and frequency domains) of SEEG contacts located in the right and left Heschl’s gyrus (displayed on the single subject T1 in the MNI space provided by SPM12) for a trial time window (−1,000 to 6,000 ms) for the condition 6-tone memory load, 2 s retention period (n = 5). Right panel shows the PSD, power spectrum density (zscore) average over a trial time window (0 to 5,000 ms) that was used to define frequency for phase and frequency for amplitude for the PAC analysis. Shaded error bars indicate SEM. Source data can be found at https://osf.io/m7dta/. (E) Left panel: T-values in the time-frequency domain (t test relative to baseline −1,000 to 0 before stimulus onset, FDR corrected in time and frequency domains) of SEEG contacts located in the right and left hippocampus (displayed on the single subject T1 in the MNI space provided by SPM12) for a trial time window (−1,000 to 6,000 ms) for the condition 6-tone memory load, 2 s retention period (n = 14). Right panel shows the PSD, power spectrum density (zscore) average over a trial time window (0 to 5,000 ms) that was used to define frequency for phase for the PAC analysis. Shaded error bars indicate SEM. Source data can be found at https://osf.io/m7dta/. (F) SEEG contacts modelled with 4 mm radius spheres (see Methods) in the MRI volume showing a significant increase in oscillatory power (FDR corrected) relative to baseline in theta (4 Hz) and gamma (30–90 Hz) ranges (Hilbert transform averaged over time) during encoding, retention, and retrieval in all memory conditions in all participants (n = 16). All results are displayed on the single subject T1 in the MNI space provided by SPM12. Source data can be found at https://osf.io/m7dta/.