Fig. 2: Formation of stable and orthogonal spatial activity patterns across VR contexts relates to dynamic spatial memory.

a Representative SM grid cell raster plots from sessions 1 vs 6 (left vs right subpanels) in young (Y), MA, and aged (A) mice. Dots are individual spikes. The right axis indicates trial type (dark [black], context A [pink], context B [dark blue]). Subpanel titles indicate mouse, session (S), and cell number (c.). Arrowheads indicate when context-sorted alternation trials started. b Dark firing rate (FR) autocorrelations of cells in (a) (top to bottom), revealing distance-tuning that identifies putative grid cells. c Probability densities of maximum peak prominence of cells’ real vs the spike-shuffled dark FR autocorrelation differed significantly (n = 6508 model pairs, grid vs shuffle, 0.25 ± 0.0010 vs 0.01 ± 0.0008, Wilcoxon signed-rank test, p < 0.0001). Vertical line indicates peak threshold. Bin size was 0.01. d Box and whisker plot of grid cell density by age group (n = 54 young, 58 MA, 55 aged sessions). Dots are colored by mouse (see Supplementary Fig. 1c). Box edges, center, whiskers, and black circles indicate age group interquartile range (IQR), median, 1.5 × IQR, and outliers. Age did not alter session grid cell density (young vs MA vs aged, 16.17 ± 1.14% vs 13.38 ± 0.79% vs 15.05 ± 1.24%, Kruskal–Wallis test, H = 2.58, p = 0.27). e Cross-trial correlation matrices of cells in (a) (left to right), omitting dark and gain change trials and context-sorting alternation trials. The color bar indicates the correlation value. f Effects of age and session on mean grid cell spatial firing stability across SM task phases fitted by linear mixed effects models (LMMs) (n = 2441 young, 2035 MA, 2032 aged grid cells). Large dots and vertical bars indicate the age group mean and SEM. Small dots represent LMM-fitted session averages, jittered by age group. g As in (f) for the mean grid cell spatial firing similarity ratio. h Task performance (fraction [frac.] of rewards requested) vs mean grid cell stability across age groups and task phases (n = 106 young [left], 114 MA [middle], 106 aged [right] phases). Black vs colored dot outline indicates block vs alternation phase. Lines represent linear regression fits for significant correlations. i As in (h) for mean grid cell similarity ratio. See Supplementary Figs. 2 and 3.