Fig. 6: Reduced neuronal ensemble track-specific depiction of future and reactivation of past distinct multiple linear track experiences during sleep after sphere-rearing.
a Proportions of preplay and ‘specific-track’ preplay after cuboid- and sphere-rearing. Yellow/black stars: p-values for statistical comparison with shuffle and chance. b Experiences on distinct tracks during Day2Run2 recruit neuronal ensembles from a larger repertoire of Pre-Day2Run1 sleep frame clusters in cuboid- vs. sphere-reared rats (individual rat examples). Proportion of significant frame pairs in Pre-Day2Run1 sleep remains significantly higher in sphere- vs. cuboid-reared rats (bottom-left, p = 6.0 × 10−16, two-sided Z-test for 2 proportions), which does not translate to a significant difference in the number of discrete clusters (bottom-center, p = 0.16, two-sided rank-sum test). Specificity (bottom-right) of cluster recruitment from Pre-Day2Run1 sleep by each future track is higher in cuboid- vs. sphere-reared rats (p = 0.030, t-test). c Proportions of replay and ‘specific-track’ replay after cuboid- and sphere-rearing. Yellow/black stars: p-values for statistical comparison with shuffle and chance. d Cell assembly plasticity (sleep reactivation exceeding pre-activation) for multiple experiences in P16-25 cuboid-reared and P25 sphere-reared rats. Cuboid-reared P24-25 rats exhibit multi-fold increase in experience-dependent cell assembly plasticity (sleep reactivation) vs. younger rats. Cell-assemblies in sphere-reared rats (P25) have impaired plasticity compared with P24-25 and age-matched (P25) cuboid-reared rats, and similar reactivation to P16-23 cuboid-reared rats. e Average neuronal ensemble similarity across frame-pairs in Pre-Day2Run1 sleep is significantly correlated with average population vector correlations across future track pairs during run (Pearson’s correlation, N = 24 rats). Shaded region is 95% confidence interval. f Contribution of Pre-Day2Run1 sleep or run firing properties to pattern separation of multiple spatial experiences estimated using multiple regression (N = 24 rats). Sleep individual neuronal firing rates (p = 0.020), across-neuron firing rate variation (p = 0.016), neuronal ensemble similarity across frame-pairs (p = 5.4 × 10−5) are all significantly related to population vector correlations across future distinct track pairs, while run spatial information has no further significant relationship (p = 0.11, likelihood ratio tests). Data in a-d are represented as mean ± SEM. N (rats/group x 2 directions/rat for each track): (a–c) Cuboid, P24-25 (8); Sphere, P25 (10). ***p < 0.005. **p < 0.01. *p < 0.05. ns = not significant. Source data are provided as a Source Data file.
