Extended Data Fig. 9: PGOw in LGN.

a, Two-dimensional projection of type I and type II PGOw (red and blue LFP traces, respectively) using Laplacian eigenmaps for one exemplary PBn-LGN experimental session. Each trace corresponds to one event. Events are classified on the basis of their peri-event MUA (801–2500 Hz), also used to compute the projection. Note the segregation of type I PGOw and type II PGOw in this space. b, Zoomed-in version of two patches enclosed by a dashed square in a. Type I PGOw usually display one biphasic cycle, whereas type II PGOw display more oscillatory behaviour. c–e, PGO-wave LFP (c) spectrogram (d) and MUA (801–2,500 Hz) (e) averages are shown for the selected events of this experimental session exhibited in a and b. f, g, SDR results for putative PBn-LGN and LGN-VCX causal interactions during PGOw. Unlike that of neocortex, note the significantly increased SDR in the PBn → LGN interaction direction, indicating a putative causal effect of PBn over LGN PGO wave activity. Here we asked whether the brainstem exerted a causal influence on LGN, and whether LGN exerted a similar influence on VCX activity during the occurrence of PGOw. SDR analysis confirmed that PGOw propagate from PBn to LGN (P < 0.004 sign test, N = 7 experimental sessions; Panel f). In addition, we did not find statistically significant evidence for the direction of causation for the case of LGN and VCX recording pairs (P = 1 sign test, N = 7 experimental sessions; Panel g).