Extended Data Fig. 5: Sequential activations and burst-to-burst similarity are not present after shuffling.

a, Same raster plot visualization as Fig. 1a after shuffling. The population firing rate remains the same after shuffling and is shown by the red solid line. Population bursts exist in the same frames after shuffling and are denoted by local maxima (black dots) that exceed 4× r.m.s. fluctuations in the population rate. The burst duration windows remain the same after shuffling and are marked by the shaded gray regions, which denote the interval in which the population rate remains above 10% of its peak value in the burst. The average firing rate per unit remains the same after shuffling. b, Same instantaneous firing rate visualization as Fig. 2b. The same ordering is used as in Fig. 2b. c, Same average burst peak-centered firing rate visualization as Fig. 2c after shuffling. The burst peak is indicated by the dotted line. The unit order is the same as Fig. 2c. Please note that the progressive increase in the firing rate peak time relative to the burst peak, as well as a spread in the active duration for units having their peak activity later in the burst are not present anymore after shuffling. The average firing rate is normalized per unit to aid in visual clarity. d, Same burst-peak-centered spike times and pairwise burst-to-burst correlations as in Fig. 2d after shuffling. For (i) and (ii), the consistent firing patterns relative to the burst peak as exemplified in Fig. 2d are not present anymore and the average burst-to-burst correlation scores have decreased from 0.96 to 0.69 and from 0.82 to 0.53, respectively. Meanwhile, the average burst-to-burst correlation for the non-rigid unit exemplified in (iii) decreased from 0.51 to 0.49.