Figure 1. Training performance in the appetitive spatial task. (A) Event arena and training paradigm. The appetitive spatial task was composed of two trials per session. During the encoding trial, the rat found hidden food rewards (filled circle) inside the arena. After a delay, rats would encounter 5 different sandwells (circles) with the same location being rewarded. (B) A group of aged rats (19–23 months old) were trained at the older age followed by interleaving training and probe sessions. Various tests refer to the memory tests presented in Figures 2–5. (C) Latencies to retrieving rewards gradually decline across 4 blocks of training (one-way, repeated measure ANOVA, F3, 12 = 18.8, p < 0.0001). (D) The number of errors made at retrieval was below chance (dashed line; one-sample t-tests, all p < 0.001) and stable across 4 blocks of training (one-way, repeated measure ANOVA, F3, 12 = 0.06, p = 0.96). (E) The number of errors made at retrieval during interleaving training was below chance throughout the study (all p < 0.05–0.001). (F) A second group of rats was trained and tested in young (3–5 months old), in middle age (11–13 months old), and at later age (19–23 months old). Various tests at the older age refer to the memory tests presented in Figure 2–5. (G) Latencies to retrieving rewards were stable (one-way, repeated measures ANOVA, F3, 12 = 2.31, p = 0.11). (H) The number of errors made at retrieval was below chance (dashed line; all p < 0.0001) and stable across 4 blocks of training (one-way, repeated measures ANOVA, F3, 12 = 0.89, p = 0.44). (I) The number of errors made at interleaving retrieval trials was below chance (all p < 0.002). All data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.005.