Fig. 3: Modulation of neuronal (co-)FRs by SOs, spindles and ripples.

a, Single-participant, single-neuron example. Raster plots show action potentials across time (x axis) for individual events (y axis), with the mean SO, spindle and ripple event-related potentials (ERPs) of the session superimposed (magenta). ERPs were band-pass filtered in the SO and spindle detection range, and from 0.1 to 120 Hz for ripples, to preserve the sharp-wave component for visualization. Right, magnetic resonance imaging–computed tomography scan (in MNI space) depicting the macro contact and the microwire bundle (arrow) from which this single unit was isolated. b, MUA FRs relative to pre-event baselines (black; left y axis), averaged across sessions for SOs (left), spindles (middle) and ripples (right) (n = 20, recorded from ten individuals; see Supplementary Fig. 5 for mean ± s.e.m.). Grand-average ERPs across sessions are superimposed (magenta; right y axis). c, Maximal FRs per session (n = 20, recorded from ten individuals), illustrating the stepwise increase across SOs, spindles and ripples. Bars show means ± s.e.m. of conditions. Individual lines represent individual sessions, with sessions from the same participants grouped by color. SOs versus spindles: t(19) = 2.21, P = 0.040; spindles versus ripples: t(19) = 3.96, P = 8.45 × 10^–4 (both two-sided paired-samples t tests). d, Event-locked cross-correlograms (CCGs) of neuronal firing among pairs of microwires, relative to shift predictors and non-event surrogates. Right, CCGs for SOs (black), spindles (blue) and ripples (red) scaled between 0 and 1 (mean ± s.e.m. across sessions; n = 20, recorded from ten individuals), illustrating the stepwise narrowing of co-firing windows. Horizontal green lines indicate FRs versus 0, P < 0.05 (corrected via cluster-based permutation test). *P < 0.05, ***P < 0.001 (two-sided paired-samples t test).