All procedures were approved by the Institutional Animal Care and Use Committees in accordance with the National Institutes of Health directive. Male Long-Evans rats were obtained at 8–9 months of age from Charles River Laboratories (Raleigh, NC). They were housed in a vivarium at Johns Hopkins University until they were 24–26 months old for assessment of aged rats. Young adult rats were obtained from Charles River and were housed in the same vivarium. Rats were individually housed in cages containing corncob bedding and constant ventilation. The vivarium was 25°C and on a 12-h light/dark cycle (lights on at 7 A.M.). Water and food were provided ad lib. Rats were continuously monitored for health. Pathogen-free status and necropsies were performed at the time of sacrifice. Rats that showed impaired health or disabilities that could impact behavioral performance (e.g., poor eyesight, clinical evidence of renal impairment, pituitary or other tumors) were excluded from the study. All procedures were approved by the Johns Hopkins University Institutional Animal Care and Use Committee in accordance with the National Institutes of Health directive.

Young adult (8–9 months) and aged rats (24 months) were tested in an assessment of hippocampal function prior to odor recognition memory behavioral tests. The background behavioral assessment used a well-established Morris water maze protocol as described in detail elsewhere (Gallagher et al., 1993; Tomás Pereira and Burwell, 2015). This protocol was designed to tax memory in the task with sparse training (3 trials per day) at 24-h intervals. Rats were trained for 8 days (3 trials per day) to locate a camouflaged escape platform that remained at the same location throughout training in a water maze surrounded by curtains with fixed cues. Every sixth trial consisted of a probe trial (no escape platform for the first 30s of the trial) that served to assess the development of a spatially localized search. Learning Index (LI) scores were derived from each rat’s proximity to the platform during the four probe trials. The proximity measure was obtained by sampling the position of the animal in the maze (10 times per second) to provide a record of its distance from the escape platform in 1-s averages. The learning index is the sum of weighted proximity scores obtained during probe trials, with low scores reflecting a more accurate search and indicating better retention of the platform location. A learning index cutoff was used to identify aged rats as Aged Unimpaired (“AU”) or Aged Impaired (“AI”). The cut off value was an index score of 240, with higher scores representing worse performance and reflecting scores that fall inside or outside the normative range collected from young adult Long-Evans rats over many years. Cue training was used to assess the sensori-motor and motivational status of the rats. Only rats with successful cue training performance were included in the present study. After behavioral characterization and cue training, a total of 23 rats, including Y (n = 7), AU (n = 8), and AI (n = 8) rats were selected (Figure 1B). One AI died before completing short-term delay test, and one AU did not meet minimum exploration criteria for the first long-term delay task.

Six weeks after MWM characterization, odor recognition memory was assessed over a 15-day protocol using odor sets in three within-subject assessments. A square arena (70 × 70 × 50 cm) was used for long- and short-term memory testing and a circular arena (diameter = 76 cm) was used for a second long-term memory test to provide a distinctly different context. Both arenas were grey and were surrounded by a black curtain during sample and test phases. The bottom of each arena was covered with ~2 cm of corncob bedding which was refreshed between counterbalanced groups, each consisting of AU, AI, and Y rats (Antunes and Biala, 2012). Odors were made fresh each day with 200 μL of odor liquid extract (McCormick, Hunt Valley, MD) diluted in 1 mL of distilled water and placed on a gauze pad inside a scintillation vial which was fastened to the arena floor with Velcro. In the square and the circular arenas, odor vials were placed 10 cm from the walls. Odor concentration and identity was based on previous work showing that young adult and aged rats show no preference for or aversion to these odorants under these conditions (Weiler et al, 2021). The sequence of behavioral tasks is outlined in Figure 1A.

In the preliminary habituation and acclimation phase, rats were handled for at least 5 min per day for 3 days and then allowed to freely explore the test arena for 10 min a day for 5 days. Following this, animals were given a sampling phase in which two vials containing identical odors were introduced into the arena and rats were allowed to explore for 10 min. After a 24-h delay, the animals were placed in the square arena with two vials, one containing the odorant present on the sampling phase and the other containing a novel odor (Long-term Memory 1, LTM1) (Weiler et al., 2021). After a 1-week delay, this was followed by a Short-term Memory (STM) test in the square arena in which animals were given a sampling phase with two identical odors followed by the test phase with one novel and one familiar odor, with a 5-min delay between test and sample phases. Finally, after a 1-week delay animals were given a second Long-term Memory test (LTM2) in a circular arena with a new set of odors. All animals were given the same pair of odorants for each test, with the identity of novel and familiar odors pseudo-counterbalanced across animals (LTM1: lime and orange, STM: banana and almond, LTM2: vanilla and coconut). For all sample and test phases, animals were allowed to freely explore for 10 min. All behavior occurred under red light with overhead lights off and with a white-noise generator on. All phases were digitally recorded for offline scoring between three experimenters.

For scoring during the sample and testing phases, exploration time was defined as the time the rat’s snout was either in or directly above the vial. A Recognition Index (RI) score [(novel)/(novel + familiar)] was calculated, as previously described for object and odor recognition (Ennaceur and Delacour, 1988; Weiler et al., 2021). An RI of 0.5 represents no odor exploration preference, reflecting a lack of recognition memory of the sample phase odor. All scoring was performed by raters who were blind to experimental conditions and scoring was confirmed across three different raters (Supplementary Figure S1). Any rat that explored odor vials for less than 5 s during the sampling or testing phase was excluded from further analysis.

The Kolmogorov–Smirnov test was used to test for normal distribution of the data, and Bartlett’s test was used to test for equality of variances for all variables. Memory retention for sample phase odors was investigated using one-sample t-tests, comparing RI scores to 0.5 (i.e., comparing investigation preference for the novel odor to chance). To assess differences in recognition index memory in each task, one-way ANOVAs were performed with group as the independent factor with Tukey’s HSD post-hoc tests and Cohen’s d statistics where appropriate. Correlations between learning index and recognition index were determined with Spearman rank correlation coefficient and correlations between repeated recognition index scores were determined with Pearson correlation coefficient. Statistical tests were performed using GraphPad Prism version 10.0.0 for Windows (GraphPad Software, Boston, Massachusetts United States, www.graphpad.com). Full results of statistical tests not reported in the text are reported in Supplementary Table S1.

As illustrated schematically in Figure 1A, all rats used in this study were first assessed for individual differences in spatial learning followed by three tests of odor recognition memory at different delays. Spatial memory performance was assessed using the hidden platform water maze protocol developed in this study population and optimized for sensitivity to detect individual differences in aging apart from confounds due to physical disability or pathological conditions (Gallagher et al., 1993). Higher learning index (LI) scores signify worse performance by reflecting search at a greater distance from the escape location during memory probe tests. Figure 1B shows the learning index distribution derived from task performance over the four probe trials for young adult control (n = 7) and all aged rats (n = 16) used in this study. A repeated-measures, two-way ANOVA confirmed rats improved with training block (F (4, 80) = 187.8, p < 0.001) with the last day showing differences between age groups.

As found in this study population and in the subset of animals used here, AU rats’ performance is on par with Y rats while AI rats fell outside the range of normative distribution of young adult rats (LI scores ≤240 and > 240 were classified as AU and AI, respectively). An overall one-way ANOVA of learning index demonstrated significant differences across age (F (2, 20) = 29.24, p < 0.0001). Y and AU groups differed significantly from AI rats (both groups p < 0.0001) but LI scores were not different for AU and Y rats (p = 0.98).

To assess the relationship between hippocampal-dependent spatial learning performance and odor recognition memory, the characterized rats in this study were given a series of odor recognition memory tasks with varying delays and a Recognition Index (RI) score was calculated [(novel/novel + familiar]). Two long-term memory tasks were performed in which rats were tested 24 h after concluding the sample phase (Figure 2). In the first of these tests (LTM1, Figure 2A), Y and AU animals preferentially explored the novel odor vial more than expected by chance (Supplementary Table S1; one-sample t test, mean difference compared with 0.5: Y: p = 0.011, AU: p = 0.007). AI animals, however, preferentially explored the familiar odor more than chance (p = 0.0394), resulting in lower RI scores as a group, significantly differing from both Y and AU (One way ANOVA: F (2, 19) = 13.83, p = 0.0002, Tukey post hoc: Y vs. AU: p = 0.997; Y vs. AI: p = 0.0007; AU vs. AI: p = 0.0006), suggesting that they may have treated the familiar odor as if it were novel, as reported elsewhere (see Burke et al., 2010). This was not driven by a failure to sample the familiar odor during initial presentation, as total exploration time in the LTM1 sampling phase was similar across groups (F (2, 19) = 1.304, p = 0.295) and was not correlated with RI scores (Spearman, r (15) = 0.282, p = 0.31), and not due to a failure to investigate the odor vials during the test phase (Figure 2C; Supplementary Table S1; F (2, 19) = 0.08697, p = 0.917). To further examine the relationship between this impairment in long-term odor recognition memory and hippocampal-dependent memory in the aged animals, we plotted RI scores for LTM1 against LI scores for each animal (Figure 2B), identifying a significant negative correlation between the two test measures (Spearman, r (13) = −0.736, p = 0.0025).

To determine whether this impairment in odor recognition memory is stable over time for a given animal and across different odor pairs, animals were given a second Long-term Memory test (LTM2, Figure 2D) using different odors and in a different arena. Y and AU groups had above chance exploration of the novel odor, while AI performed at chance levels (Supplementary Table S1; one-sample t test, mean difference compared with 0.5: Y: p < 0.003, AU: p < 0.039, AI: p = 0.131). Again, RI scores for Y and AU rats were similar, while AI rats had significantly lower RI scores (Supplementary Table S1; one way ANOVA: F (2, 19) = 7.612, p = 0.0037, Tukey post hoc: Y vs. AU: p = 0.996; Y vs. AI: p = 0.0083; and AU vs. AI: p = 0.0079). Similar to LTM1, aged animals RI scores on LTM2 were negatively correlated with water maze LI scores (Figure 2E; Spearman, r (13) = −0.525, p = 0.0471). While total test phase odor vial exploration time in LTM2 was lower in the aged rats relative to Y (Figure 2F), there was not a statistical difference between the groups for sample phase exploration (Supplementary Table S1; one way ANOVA: F (2, 19):1.976, p = 0.1661) or test phase exploration (Figure 2F; Supplementary Table S1; one way ANOVA: F (2, 19) = 0.580, p = 0.5695) and sample phase exploration did not correlate with RI scores (Spearman, r (15) = 0.054, p = 0.853). Furthermore, there was a significant correlation between performance on LTM1 and LTM2 across all animals (Pearson, r (22) = 0.451, p = 0.040). Taken together, the results of LTM1 and LTM2 suggest that for aged animals, the ability to recall previously experienced odor cues after a long-term delay is a stable phenotype and that this impairment has a direct relationship with an individual’s MTL-dependent cognitive abilities.

The failure of AI rats to explore the novel odor in these tests could result from an olfactory deficit or from a lack of novelty exploration preference. To determine whether this is the case, all rats were given a short-term memory test with a 5-min delay between sampling and testing phases (Figure 3). All young rats preferentially explored the novel odor vial presented in the test phase, demonstrating intact spontaneous novelty exploration under these conditions (Figure 3A). Similar to young adult animals, all aged animals preferentially explored the novel odor significantly more than chance in the test phase (Figure 3A; Supplementary Table S1; one-sample t test, mean difference compared with 0.5: Y: p = 0.0021, AU: p = 0.0001, AI: p = 0.001), with no statistical difference between groups (Supplementary Table S1; one way ANOVA: F (2, 20) = 2.13, p = 0.1446). Importantly, short-term memory RI scores were not correlated with MWM learning index scores (Figure 3B, Spearman, r (14) = −0.41, p = 0.113), indicating that the low RI scores for AI animals in the LTM tests was not due to an inability to identity the novel odor or a lack of innate novelty exploration preference. Aged rats as a group tended to have lower test phase exploration times relative to Y in the STM test (Figure 3C), but this was not significantly different (One way ANOVA: F (2, 20) = 1.319, p = 0.2896). The AU and AI, furthermore, did not differ in exploration during the sample phase (Supplementary Table S1; one way ANOVA: F (2, 20) = 4.372, p = 0.0266, Tukey’s post hoc: Y v AU: p = 0.3219, Y vs. AI: p = 0.0205, AU v AI: p = 0.2996) and sample phase exploration did not correlate with RI scores (Spearman, r (16) = −0.128, p = 0.635). These results demonstrate that under these experimental conditions, both AU and AI animals have intact odor recognition, can form a memory of a recently presented odor cue, and have intact novelty exploration preference after a short delay when the relevant cues are odorants.