Fig. 2: Item-level cortical high-gamma pattern around hippocampal ripples during the learning stage.

A Depiction of hippocampal contacts in all patients (N = 11). Each color denotes one patient, and for each participant, we select one hippocampal contact to identify ripples (marked by large spheres). B Example of ripple events in hippocampal recordings during the learning stage. From top to bottom: raw hippocampal LFP; ripple-band filtered LFP (70–180 Hz); normalized ripple-band envelope above the threshold were detected as ripple events. C, D Grand average peri-ripple field potential and spectrogram during learning phase (n = 7763 ripple events from 11 patients). E Distribution of inter-ripple intervals. F Normalized ripple rate (mean values ± SEM) triggered by cue onset during the learning stage. The yellow bar indicates time period of significant increases in ripple rate (p = 0.032, cluster-based permutation test based on one-sided paired-sample t-test, N = 11). G Distribution of all intracranial cortical contacts in all patients (N = 11). H RRS (retrieval-retrieval similarity) schematic illustration between different learning rounds. Cortical activity pattern around ripple onset (±500 ms) was correlated either between trials with the same items (WI) or between different items (BI). Item-level representation was calculated by subtracting BI similarity from WI similarity. I Averaged WI similarity. J Averaged BI similarity. K Averaged Item-level representation. L WI similarity (averaged across ±0.5 s) was significantly higher than BI similarity (t(10) = 4.54, p = 0.001, two-sided paired-sample t-test, N = 11). Data are presented as mean values ± SEM. **p < 0.01, RRS: difference of retrieval-retrieval similarity. WI within item. BI between item. The schematic illustrations (clock and computer) are sourced from The Noun Project (https://thenounproject.com, icons created by Thea Graph and Maxicons, respectively) under a Royalty-Free License.