Fig. 4: Covariation between single-cell spike train auto-correlograms and SWR cross-correlograms.

a SWRs cross-correlograms (top left) and auto-correlograms across brain states (bottom) for two example HD neurons. Circular plots (top right) indicate HD tuning curves. b, c Same as a for three non-HD neurons recorded simultaneously on the same shank. Top right, average waveforms on the eight recording sites of the shank, from dorsal (top) to ventral (bottom) positions. The neuron shown in gray is the same for both panels. Note that, although the cell bodies of the two neurons in c are close to each other (see waveforms), auto-correlograms and responses to SWRs are strikingly different. d Same as a–c but for three neurons recorded on different shanks and different sessions in the same animal. Their anatomical location is shown on the left-side map of the anterior thalamus. e PCA projections of first 20ms of auto-correlograms (top) and SWR cross-correlograms (bottom) for the three example neurons of b, c. f Correlation matrix between the two series of PCA weights (left) and for shuffled neuronal identities (right) for all neurons. ρ²: total correlation. g Distribution of total correlation ρ² for actual (red line) and for random shuffling of cell identities (1000 times) between all cells irrespective of shank and recording days (black) and between cells recorded on the same shank on a given day (gray). h Distribution of total correlation ρ² for actual (vertical lines) and shuffled data between SWRs and auto-correlograms from wake (red), REM sleep (light blue) and NREM sleep (dark blue). Dashed line shows correlation between SWRs and stacked auto-correlograms of all epochs (same as in g). i Same as h but for cell identities shuffled within shanks.