Fig. 6: Activity in the occipital cortex signifies REM sleep.

a Dynamic change of the REM-like state during NREM-to-REM transitions. Gray line, data from individual recordings. Red line, an average of 15 recordings from 5 mice. Shading is SEM. Blue line, a linear fit of the average data during preREM (the 20-s immediately before REM onset). b Quantification of PC1^REM dynamics during NREM-to-REM transitions. P = 0.036 for NREM vs. preREM; P < 0.0001 for NREM vs. REM; Two-tailed paired t-test. n = 15 recordings from 5 mice. Data are mean ± SEM. c Diagram showing different NREM periods. d An example showing dynamic changes of PC1^REM before and after a REM bout. Scale, 0.5 and 100 s. e Quantification of PC1^REM before and after REM bouts. P = 0.011; Two-tailed paired t-test; n = 10 recording from 5 mice. f Diagram describing the definition of different periods of NREM before the NREM → Wake or the NREM → REM transitions. g Box-plot showing the quantification of REM-like brain state in different NREM periods. Box, 25–75% range and median line; Whisker, 1.5 interquartile range; Blue line, data fitting with a normal distribution model. F = 14.5; P = 0.41 for NW₁ vs. NW₂; P = 0.043 for NW₁ vs. NR₁; P < 0.001 for NW₂ vs. NR₂. one-way ANOVA with posthoc Tukey’s tests; n = 85 and 74 (15 recordings from 5 mice) for NREM → Wake and NREM → REM transitions, respectively. h Diagram describing the definition of different periods of NREM that are associated with the different propensities of transitioning into REM sleep. i Box-plot showing the quantification of REM-like brain state in different NREM periods. The box-plot elements were same as g. F = 48; P = 0.48 and 0.98 for NREM₁ vs. NREM₂ and NREM₃ vs. NREM₄, respectively; P < 0.0001 for all other paired-wised comparisons; one-way ANOVA with posthoc Tukey’s tests; n = 81 REM bouts and 166 NREM bouts, 15 recordings from 5 mice. Raw data for b, e, g, i and exact P values are provided in a Source Data file.