Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2026 Aug 21;36(5):e70123. doi: 10.1002/hipo.70123

Dissociation of Memory for Order and Memory for Items by Concurrent Cognitive Load in Rhesus Monkeys (Macaca mulatta)

Pankhuri Singhal 1,2,✉, Victoria L Templer 1,2,3, Robert R Hampton 1,2
PMCID: PMC13498721  PMID: 42629632

ABSTRACT

Memory of the order in which events occurred addresses the “when” component of episodic‐like memory in animals. Seminal work demonstrated that normal rats remember the order in which they encountered odors, and hippocampal lesions selectively impaired this memory for order while leaving recognition performance intact. This work inspired us to develop an image‐order paradigm to model episodic memory in monkeys. Monkeys saw five images in sequence and at test reported which of two images appeared first. Unlike the case with rats, hippocampal lesions did not impair memory for order, raising questions about which memory systems monkeys used in this task. Because the memory intervals were short in the image‐order task, monkeys may have used nonhippocampal‐dependent working memory instead of long‐term memory. To test this possibility, we introduced a concurrent cognitive load between the study and test phases of the image‐order task. We hypothesized that if monkeys were using working memory to remember the order of the images, concurrent cognitive load would impair performance. We also tested recognition performance as a control, because familiarity‐based recognition is relatively resistant to concurrent cognitive load. Memory for order was much more attenuated by concurrent cognitive load than recognition. These results indicate that monkeys used working memory to solve this image‐order task, making it a poor model of long‐term episodic memory. The findings also provide clear evidence for a dissociation between working memory‐dependent order memory and familiarity‐based recognition memory.

Keywords: familiarity, memory dissociation, order memory, recognition memory, working memory

1. Introduction

Humans, and at least some other animals, remember information about “when” memories were formed. This ability has alternatively been called “episodic memory” (Tulving 1972, 1983a, 1993), “episodic‐like memory” (Clayton et al. 2001; Crystal 2010), “source memory” (Basile and Hampton 2017; Crystal et al. 2013; Johnson et al. 1993), or “memory for context” (Smith 1994; Xiong et al. 2026). It is widely thought that these kinds of memory depend on “binding” of contextual information to the target memories by processes carried out in the hippocampal formation (Hasselmo and Eichenbaum 2005; Ross et al. 2009; Vargha‐Khadem et al. 1997) and adjacent structures (Eichenbaum et al. 2007).

Episodic memory has also been defined in humans as producing autonoetic awareness, a conscious sense of having been personally present when the memory was formed (Tulving 1983b, 2005). Because the self‐reflective property of autonoesis is difficult or impossible to capture in nonhuman animals, many researchers have determined that rather than focusing on autonoesis, a more productive research approach is to examine the extent to which nonhumans retain what‐where‐when information, termed episodic‐like memory. Studies of food‐caching birds most famously provide evidence for this kind of memory (Clayton and Dickinson 1998; Feeney et al. 2009; Zinkivskay et al. 2009). Related reports have been published concerning species from invertebrates (Jozet‐Alves et al. 2013) to great apes (Martin‐Ordas et al. 2010; Schwartz et al. 2005). Some other work in monkey species, animals in which we might expect to find episodic memory, have failed to find evidence for what‐where‐when memory (e.g., Hampton et al. 2005).

One way to probe whether animals record information about when a memory was formed is to have them discriminate the order in which remembered events occurred. For example, rats remember the order in which they encountered odors (Fortin et al. 2002; Kesner et al. 2002). Rats sniffed a series of five randomly selected odors mixed in sand, and after a delay they were presented with two odors. On half of trials, rats were presented with two odors from the study list and rewarded for digging in sand laced with the odor that occurred earlier in the list, assessing memory for order. On the other half of trials, rats were presented with a single odor from the study list, paired with a new odor not from the list. Rats were rewarded for digging in the sand with the new odor. These trials tested the ability of rats to recognize previously presented odors, independent of when they occurred in the study list. Rats reliably selected the correct odor in both types of tests. Critical for our understanding of the neurobiology of memory, damage to the hippocampus impaired memory for order but not recognition performance (Fortin et al. 2002). Electrophysiological recordings also implicated the hippocampus in memory for order in this task. Hippocampal ensemble representations formed as rats sampled odors were more similar when they occurred close in time (or space) than when they were separated by longer intervals. This greater similarity of ensembles for odors experienced closer in time provides a potential mechanism for representing and decoding the order of events (Manns et al. 2007). The combination of the behavioral evidence for memory of when events occurred, the electrophysiological evidence, and susceptibility to hippocampal damage, make this odor‐order task a promising animal model of episodic memory.

Because the odor‐order paradigm of Fortin et al. (2002) was so exciting and promising, we developed a conceptually parallel image‐order task for monkeys (Templer and Hampton 2013). Monkeys were shown a sequence of five images and after a delay were rewarded for selecting the image that had appeared earlier in the study list. Monkeys performed well and follow‐up experiments ruled out the use of memory strength, list position, or mere passage of time as strategies to solve the task. These results indicate that monkeys do remember the order in which images were presented and potentially provide a model of episodic memory in monkeys (Templer and Hampton 2013).

While the monkey memory for order task showed many similarities to the odor‐order task used with rats, performance in this task was found to be insensitive to hippocampal damage, raising questions about whether it is an appropriate model of episodic memory (Basile et al. 2020). In addition to the insensitivity to hippocampal damage, other properties of the monkey task raise questions about the suitability of this paradigm as a model of episodic memory in monkeys. Monkeys were trained in this task over many thousand trials, and the delays over which memory was tested were relatively brief. Because episodic memory is generally considered a form of long‐term memory, short delays between encoding and testing may not be appropriate for models of episodic memory (Crystal 2009; Tulving 1972). The training and testing parameters used in the monkey image‐order task may have favored the use of shorter‐term working memory rather than longer‐term episodic memory. If monkeys used working memory, which would presumably be dependent on frontal cortex rather than the hippocampus (Funahashi and Kubota 1994; Fuster and Alexander 1971; Miller et al. 1996; Petrides 1995), this would explain why performance was not dependent on the hippocampus.

To assess the extent to which monkeys used working memory to solve the image order task, we introduced a concurrent cognitive load between the study and test phases of trials. Because working memory is an active, resource limited process, a concurrent cognitive load interpolated during the delay interval should cause a significant reduction in accuracy when working memory is critical for accurate responding (Basile and Hampton 2013; Brady and Hampton 2018; Logie 1986; Phillips and Christie 1977). We hypothesized that if performance were impaired by concurrent cognitive load, then monkeys used working memory to remember the order of images. One limitation of this interpretation is that a concurrent cognitive load might have nonspecific effects other than those on working memory, for instance by affecting attention or motivation or other processes generally. Fortunately, like the odor‐order task of Fortin et al. (2002), this image‐order task can be conducted with a recognition performance control task. Recognition performance in monkeys can often be successful based on familiarity alone, and familiarity is much less impacted by a concurrent cognitive load than is working memory (Basile and Hampton 2013; Brady and Hampton 2018). Thus, if concurrent cognitive load affects memory for order, which cannot be accomplished by familiarity alone, more than it does recognition performance, this would further support the conclusion that memory for order depends on working memory in this paradigm.

2. Materials and Methods

The methods used in these experiments were approved by the Emory University Institutional Animal Care and Use Committee (IACUC).

2.1. Subjects

We studied six 8‐year‐old male rhesus monkeys ( Macaca mulatta ), each with approximately 4 years of computerized cognitive testing experience. Each monkey had access to his cage‐mate except during testing and during feeding at the end of the day. Before testing, monkeys were separated by insertion of dividers between cage‐mates that allowed limited visual and physical contact, but prevented access to cage‐mates' testing rigs.

2.2. Apparatus and Materials

Monkeys were tested in their home cages using computerized touch‐screen rigs that were attached to the front of each cage. Cage doors were raised, giving subjects full access to the screen during testing. Each rig consisted of a 15‐in. LCD color monitor (3 M, St. Paul, MN or Elo, Milpitas, CA) with a resolution of 1024 × 768 pixels, stereo speakers, two automated food dispensers (Med Associated Inc., St. Albans, VT), and two food cups below the screen. Food reinforcement consisted of 94 or 97 mg nutritionally complete primate pellets (Bio‐Serv, Frenchtown, NJ and Purina TestDiet, Richmond, IN). We presented stimuli and collected responses using programs written in Presentation (Neurobehavioral Systems, Albany, CA). A large set of 6000 images resized to 300 × 300 pixels was used for recognition memory and order tests. Images of birds, fish, flowers, and people were used for the perceptual classification tests drawn from a set of 1400 images.

2.3. Procedure

Our study combined procedures from three different tasks, described below:

Memory for order (Figure 1a): After the monkeys began trials by touching a green start square twice (FR 2), five sample images appeared, one after the other, centered on the screen. Subjects had to touch each image twice to advance to the next image, after which there was a memory delay. At test, two images from the study list appeared simultaneously on the screen, pseudo‐randomly assigned to the left and right sides of the screen. Monkeys were reinforced for selecting the image that had appeared earlier in the study sequence (Templer and Hampton 2013).

FIGURE 1.

FIGURE 1

Open in a new tab

Schematic of trial progressions of (a) memory for order task, (b) Match/Nonmatch recognition task (either the match or nonmatch screen appeared at test, not both), and (c) perceptual classification task. FR2 indicates that the monkey had to click twice on the image to proceed.

Match/Nonmatch recognition (Figure 1b): Trials proceeded exactly as in the memory for order tests through the memory delay, but the test phase differed. At test, an image appeared along with a “nonmatch” symbol, pseudo randomly on the left and right sides of the screen. Half of tests were matching trials on which an image from the studied list appeared. Images from each list position, from first to fifth, were equally likely to appear at test. Selecting the image from the study list was rewarded with a food pellet. The other half of tests were nonmatching trials on which a distractor image that was not from the studied list and the nonmatch symbol were presented. These distractor images were selected from a separate pool of 100 images. Monkeys were rewarded for selecting the nonmatch symbol on these trials (Basile and Hampton 2010).

Perceptual classification (Figure 1c): Monkeys classified images into four categories: “fish,” “flower,” “person,” and “bird.” After initiating the trial, a sample image appeared in the center of the screen, which monkeys touched twice. Four symbols then appeared around the sample image, one associated with each of the categories. Monkeys were rewarded for selecting the symbol associated with the appropriate category. Each category was represented equally within each session (Gazes et al. 2013).

2.4. Training

Phase 1: Match/No Match recognition training: Monkeys had ample experience performing the order task because of a previous study (Templer and Hampton 2013). Hence, they started training with the Match/Nonmatch recognition task. Each session consisted of 100 trials. A correct choice was rewarded by positive auditory feedback 100% of the time and a food reward 70% of the time (except Subject A who received a food reward 90% of the time to maintain his motivation). Monkeys were tested 3 days a week on the recognition task until they reached a criterion of an average accuracy of 80% or better in two consecutive sessions.

Phase 2: Intermixed order and recognition training: After completing Phase 1, monkeys were tested in sessions in which half the trials were the order task and the other half were the match/nonmatch recognition task, pseudo‐randomly intermixed. Each session had 120 trials. Food reinforcement was increased to 100% on all correct trials for all monkeys at this point to increase motivation in this more challenging task. Monkeys were tested until they reached a criterion of an average accuracy of 80% or better in two consecutive sessions.

Phase 3: Recognition training for titrating delays: After Phase 2, monkeys returned to the match‐non match recognition program (100 trials each) to individually titrate their delays until recognition of each item had a 2.5 d′ over one session (adapted from Macmillan and Creelman 1991). As before, monkeys were rewarded with 70% food reinforcement for correct choices (except Subject P whose food reinforcement was increased from 70% to 80% to increase motivation).

Phase 4: Intermixed order and recognition with titrated delays training: After Phase 3, monkeys were returned to intermixed sessions of recognition tests and tests of memory for order. Recognition tests used the individually titrated delays while the order tests used a delay of 500 ms. Monkeys were trained until they reached a criterion of average accuracy of 80% or better in two consecutive sessions. After reaching the criterion, all monkeys completed 20 sessions of the intermixed trials to ensure that they maintained their accuracies over the run. As before, monkeys were rewarded with 70% food reinforcement for correct choices (except Subject P whose food reinforcement was increased from 70% to 90% to increase motivation).

Phase 5: Perceptual classification training: Lastly, monkeys were introduced to the perceptual classification task. Each session was comprised of 100 trials. They were trained on this task until they reached a criterion of average accuracy of 80% on two consecutive sessions. The food reinforcement on correct responses for this phase was 70% for all monkeys.

2.5. Testing

Testing involved four types of trials presented in pseudo random order, each making up 25% of the total trials: (1) baseline order tests, (2) baseline recognition tests, (3) order tests with concurrent cognitive load, and (4) recognition tests with concurrent cognitive load. In baseline trials, we did not add a concurrent cognitive load. Sessions consisted of 120 trials.

Trials began when monkeys touched a green start square (Figure 2). A list of 5 images, selected randomly without replacement from the set of 6000 appeared sequentially in the center of the screen. Each image remained on the screen for a minimum study period of 250 ms, and the monkey had to touch each image at least twice to move to the next image. This was followed by a delay of 500 ms, during which the screen was gray. On half of the trials, monkeys were then required to complete a classification test. There was no feedback or reward for the classification task. If monkeys made an incorrect response in the classification task, the trial aborted and was restarted from the beginning with the same images in the same order. Following correct responses in the classification task, half the trials proceeded to the order test. The other half proceeded to the match/nonmatch recognition test after the delays titrated individually for each monkey during training Phase 3. Baseline order trials proceeded to test right after the 500 ms delay. For the baseline recognition trials, in addition to the 500 ms delay, a blank gray screen appeared for the duration of the individually titrated delays. Correct choices were rewarded with food on 70% of trials (except for Subject M and Subject S whose reinforcement was increased from 70% to 80% during this phase of testing. Subject P was at 90% food reinforcement to maintain his motivation to perform the task). A positive auditory stimulus signaled correct responses 100% of the time. Incorrect choices resulted in a negative auditory stimulus and the addition of a 10 s timeout to the normal interval between trials. After correct responses, the next trial started after the normal inter‐trial interval of 5 s. Monkeys completed 20 sessions of 120 trials each.

FIGURE 2.

FIGURE 2

Open in a new tab

Schematic of trial progression during testing. Monkeys started trials by touching the green square twice (top left). They then touched each of five sequentially presented images and a delay of 500 ms followed. In baseline trials (dashed arrows), monkeys proceeded immediately to an order test or received a recognition test after an individually titrated delay. On the concurrent cognitive load trials (solid arrows), monkeys performed the perceptual classification before proceeding to either test.

2.6. Data Analysis

Trial level accuracy data was analyzed using generalized linear mixed models (GLMMs) with a binomial distribution and logit link function, fitted by maximum likelihood (Laplace Approximation) using the lme4 package in R. Subject was included as a random intercept to account for repeated measures. For the first analysis, fixed effects included Trial Type (Order vs. Recognition) and Concurrent Cognitive Load (ConCogLoad: Yes vs. No), as well as their interaction. For the second analysis, fixed effects included Symbolic Distance (SD0, SD1, SD2, SD3) and Concurrent Cognitive Load, as well as their interaction. For the third analysis, fixed effects included Image Position (Positions 1–5 and Nonmatch) and Concurrent Cognitive Load. If significant main effects or interactions were found, pairwise comparisons between baseline (no concurrent cognitive load) and concurrent cognitive load conditions were conducted within each level of the relevant predictor, with Bonferroni corrections applied to control for multiple comparisons. Results are reported as log odds ratios (β) with standard errors (SE), z‐statistics, and p‐values.

3. Results and Discussion

The average titrated delay for the memory tests was 35.83 s: Subject S, 40s; Subject G, 30s; Subject H, 25s; Subject M, 10s; Subject A, 100s; Subject P, 10s. The average time added by the competing cognitive load to the trials across monkeys was 2.1s (averages for individual monkeys: S—2.2s; G—1.6s; H—2.1s; A—2.7s; P—2.1s; M—1.6s). The computer program controlling the experiment was intended to continue sessions from exactly where they were interrupted following events like crashes or early termination at the end of the day. This feature did not always work correctly. As a result, some trials were not completed as expected, resulting in some gaps in the data, and small deviations from the intended counterbalancing of trial types. These gaps did not follow any pattern we could detect that would compromise the results, and the gaps were similar in proportion among trial types (6.78% for baseline order trials; 7.25% for order with concurrent cognitive load; 7% for baseline recognition, 7.14% for recognition with concurrent cognitive load). The average number of missing trials per session of 120 trials was 8.05 (SD = 8.96) for Subject A, 6.45 (SD = 9.14) for Subject G, 13.0 (SD = 21.0) for Subject M, 9.25 (SD = 9.85) for Subject P, and 14.0 (SD = 13.3) for Subject S. Subject H received 8 extra trials.

Concurrent cognitive load impaired memory for order but not recognition performance (Figure 3; Mixed effects logistic model: Reference level: Order baseline; Test Type × ConCogLoad, β = 0.915, SE = 0.080, z = 11.404, p < 0.001; Test Type, β = 0.269, SE = 0.058, z = 4.608, p < 0.001; ConCogLoad, β = −0.918, SE = 0.053, z = −17.339, p < 0.001; pairwise comparison with Bonferroni corrections: baseline order vs. order with concurrent cognitive load, z = 17.339, p < 0.001; baseline recognition vs. recognition with concurrent cognitive load, z = 0.046, p = 0.963). This effect of concurrent cognitive load on memory for order but not recognition suggests that only performance in the order task depended on working memory. Because working memory depends on limited cognitive resources, it is especially sensitive to concurrent cognitive loads (Basile and Hampton 2013; Brady and Hampton 2018; Logie 1986; Phillips and Christie 1977).

FIGURE 3.

FIGURE 3

Open in a new tab

Proportion correct as a function of test type and concurrent cognitive load. Accuracy in only order memory tests was affected by concurrent cognitive load. Chance is at 0.5. Error bars represent SEMs.

Because the images used in this test were drawn from a large set of 6000, monkeys were likely able to use relative familiarity to identify images that had been seen during study, as the distractor images would not have been familiar (Basile and Hampton 2013; Brady and Hampton 2018). In contrast, it is unlikely that monkeys can solve the order task using relative familiarity because all the images are familiar from study (Fortin et al. 2002; Templer and Hampton 2013). These results appear to dissociate a working memory process from a familiarity process, with the former subserving memory for order and the latter recognition.

One characteristic pattern of performance in tests of memory for order is the symbolic distance effect, according to which subjects are better able to identify which stimulus occurred first when the tested stimuli were relatively far apart in the studied list (Fortin et al. 2002; Manns et al. 2007; Templer and Hampton 2013). We replicated the symbolic distance (SD) effect in this study, and also found that the concurrent cognitive load reduced accuracy at each symbolic distance (Figure 4; Mixed effects logistic model: Reference level: SD0 baseline; Symbolic Distance—SD1: β = 0.505, SE = 0.093, z = 5.416, p < 0.001; SD2: β = 0.934, SE = 0.117, z = 7.965, p < 0.001; SD3: β = 1.849, SE = 0.209, z = 8.853, p < 0.001; ConCogLoad, β = −0.587, SE = 0.080, z = −7.37, p < 0.001; Symbolic Distance × ConCogLoad, SD1: β = −0.378, SE = 0.13, z = −3.01, p = 0.003; SD2: β = −0.791, SE = 0.151, z = −5.23, p < 0.001; SD3: β = −1.303, SE = 0.24, z = −5.326, p < 0.001). Although the interaction indicates that the magnitude of the CCL effect differed across symbolic distances, the pairwise comparisons confirm that CCL significantly reduced accuracy of memory for order at every level of SD (SD0 baseline vs. with concurrent cognitive load: z = 7.370, p < 0.001; SD1: z = 9.938, p < 0.001; SD2: z = 10.707, p < 0.001; SD3: z = 8.171, p < 0.001).

FIGURE 4.

FIGURE 4

Open in a new tab

Proportion correct in tests of memory for order according to symbolic distance and concurrent cognitive load. Symbolic distance refers to the number of images that appeared between the two tested images in the study list, such that symbolic distance 0 denotes images that were adjacent in the study list while symbolic distance 3 refers to tests involving the first and last images in a five‐item study list. Accuracy increased with symbolic distance and concurrent cognitive load reduced accuracy at all symbolic distances. Chance is 0.5. Error bars represent SEMs.

Further analysis of the recognition performance confirms that the effect of concurrent cognitive load was absent in recognition tests. Results from studies of recognition memory for items presented in lists are often analyzed as a function of list position (Basile and Hampton 2010; Ebbinghaus 1902; Wright et al. 1985). We replicated a common result from these studies by finding a “U” shaped accuracy function, such that the first and last items in the study list were remembered most reliably (Figure 5; Mixed effects logistic model: Reference level: List Position 1 baseline; Position 2: β = −0.475, SE = 0.179, z = −2.649, p = 0.008; Position 3: β = −0.517, SE = 0.178, z = −2.908, p = 0.004; Position 4: β = −0.320, SE = 0.182, z = −1.757, p = 0.079; Position 5: β = −0.103, SE = 0.187, z = −0.553, p = 0.580; NM: β = 0.475, SE = 0.150, z = 3.173, p = 0.002). Unlike the case with order, concurrent cognitive load did not reduce accuracy overall for recognition performance (Figure 5, Mixed effects logistic model: ConCogLoad, β = 0.153, SE = 0.193, z = 0.794, p = 0.427; List Position × ConCogLoad, Position 2: β = −0.190, SE = 0.257, z = −0.738, p = 0.461; Position 3: β = −0.261, SE = 0.254, z = −1.027, p = 0.304; Position 4: β = −0.310, SE = 0.259, z = −1.196, p = 0.232; Position 5: β = −0.145, SE = 0.267, z = −0.541, p = 0.588; NM: β = −0.101, SE = 0.216, z = −0.468, p = 0.640). Additionally, none of the post hoc pairwise comparisons with Bonferroni were significant at any list position (List position 1 baseline vs. with concurrent cognitive load, z = −0.794, p = 1.000; Position 2: z = 0.217, p = 1.000; Position 3: z = 0.653, p = 1.000; Position 4: z = 0.907, p = 1.000; Position 5: z = −0.044, p = 1.000; NM: z = −0.523, p = 1.000).

FIGURE 5.

FIGURE 5

Open in a new tab

Accuracy in recognition memory tests as a function of list position and concurrent cognitive load. Concurrent cognitive load did not significantly reduce accuracy overall, or at any of the individual list positions. Chance is 0.5. Error bars represent SEMs.

Trials were aborted due to failure to correctly complete the classification task equally often for order and recognition tests (27% of trials for each type of test were aborted and restarted at least once).

We also removed the trials that were restarted after being aborted due to incorrect classification. Repeating the same images in the same order could have impacted memory strength and thus performance. We repeated the three analyses after removing these restarted trials and our main findings were replicated. Concurrent cognitive load impaired memory for order and recognition differently (Mixed effects logistic model: Reference level: Order baseline; Test Type × ConCogLoad, β = 0.770, SE = 0.086, z = 8.945, p < 0.001; Test Type, β = 0.264, SE = 0.058, z = 4.520, p < 0.001; ConCogLoad, β = −0.927, SE = 0.057, z = −16.205, p < 0.001). In a minor contrast to our main analyses, we found that while the effect of concurrent cognitive load was greater for memory for order, there was also a significant effect of concurrent cognitive load on recognition, a result that was not significant in the main analysis (Pairwise comparison with Bonferroni corrections: baseline order vs. order with concurrent cognitive load, z = 16.205, p < 0.001; baseline recognition vs. recognition with concurrent cognitive load, z = 2.443, p = 0.015). Further analyses found that the effect of competing cognitive load on memory for order was robust and replicated at each symbolic distance, as found in the main analysis (Mixed effects logistic model pairwise comparisons with Bonferroni: SD0 baseline vs. with concurrent cognitive load: z = 6.714, p < 0.001; SD1: z = 10.278, p < 0.001; SD2: z = 9.240, p < 0.001; SD3: z = 7.921, p < 0.001). In contrast, even though we observed an effect of concurrent cognitive load on overall recognition accuracy, we did not find a significant impact at any individual list position (Mixed effects logistic model pairwise comparisons with Bonferroni: List position 1 baseline vs. with concurrent cognitive load, z = 0.441, p = 1.000; Position 2: z = 1.004, p = 1.000; Position 3: z = 1.948, p = 0.309; Position 4: z = 2.288, p = 0.133; Position 5: z = 1.630, p = 0.619; NM: z = −0.175, p = 1.000). Thus, this additional analysis with restarted trials removed reinforces the conclusion that the effect of concurrent cognitive load is robust on working memory but weak for familiarity.

The robust effects of concurrent cognitive load on memory for order in contrast to no effects on recognition memory performance suggest that two independent memory processes were in play in this study. Memory for the order in which stimuli occurred appears to have depended critically on working memory, whereas memory for whether or not a particular image appeared in the study list depended on familiarity. These results indicate a dissociation of working memory and familiarity. Our previous work with monkeys has found both single (Basile and Hampton 2013) and double (Brady and Hampton 2018) dissociations between working memory and familiarity. The current results are consistent with this same dissociation.

The different delay periods we used to approximately equate accuracy in the order and recognition tests support the interpretation that two separate memory systems were recruited in this study. While both working memory and familiarity could have been recruited during the recognition task, the much longer titrated delay periods on recognition trials likely reduced any contribution of working memory. On the other hand, working memory would be still effective at the much shorter delay during the order task (Brady and Hampton 2018).

Although these results are consistent with a dissociation of memory systems, a single memory system account might also be viable. In the case of memory for order, monkeys would seem to have to remember both what images occurred and when they occurred, placing a heavy demand on working memory. The memory load for recognition performance is lower because monkeys only have to remember what images were presented. So, it is possible that concurrent cognitive load had a greater impact on memory for order because that task is more demanding of memory overall. Two considerations suggest this is not the case, however. The first consideration is that monkeys could not tell in advance, during encoding, whether they would receive a test of order or a recognition test. If working memory were near capacity with information about both what images appeared and when they appeared, it is not obvious that the when component should be more vulnerable to the competing cognitive load. If information about what and when were equally likely to be displaced by concurrent cognitive load, then we would not obtain the apparent dissociation observed. The second consideration is that earlier findings in monkey recognition tests showed that working memory is much more important for recognition performance when monkeys are tested with a small set of repeating stimuli. By contrast, when a large set of nonrepeating images was used, familiarity is sufficient for accuracy (Basile and Hampton 2013; Brady and Hampton 2018).

In this experiment we used a large set of images, favoring use of familiarity for recognition performance. Partly because familiarity worked so well for the interleaved recognition memory tests, monkeys may have been motivated to disengage working memory on trials with competing cognitive load leading to poor performance on tests of memory for order. Thus, it may not be that concurrent cognitive load impaired working memory, but rather that it led to a disengagement of working memory. Whether poor performance resulted from an attenuation of working memory or from a strategic disengagement of working memory, the dissociation between working memory and familiarity still stands, and contributes to the body of related work dissociating these memory systems (Basile and Hampton 2013).

This study was modeled after the influential rat odor‐order paradigm developed as a test of episodic memory (Fortin et al. 2002; Kesner et al. 2002). The current results indicate that, as adapted, this paradigm does not measure episodic memory in monkeys. While our adaptation captured many features of the procedure and the results of the work of Fortin et al. (2002), a critical difference is likely the much shorter memory intervals which appear to have enabled monkeys to use working memory instead of episodic memory. Development of a robust model of episodic memory in monkeys may require designs involving longer memory intervals.

In addition to shorter delay intervals, our study differs from the work with rodents on which it was modeled in lacking a manipulation of the location in which stimuli were presented. The location of the platform on which the odor cues were presented changed across trials in the rodent study (Fortin et al. 2002). This change in spatial context between trials likely helped animals ignore experiences with odors from previous trials. While space or location is often part of the “context” for episodic memories, temporal context is an essential component (Tulving 1983b). The hippocampus, the region most often implicated in episodic memory, has also been implicated as critical for preserving both temporal and spatial components of episodic memories (Manns et al. 2007; Ranganath and Hsieh 2016; Vargha‐Khadem et al. 1997). For example, specialized “time‐cells” in the hippocampus encode temporal context (Eichenbaum 2013, 2017; Ranganath 2019). In fact, differences in patterns of activity in ensembles encoding temporal context predicted memory performance in the rat odor‐order task, whereas the absence of differences in these ensembles indicated errors. In contrast, hippocampal spatial representations of the odor events were similar between correct and incorrect trials and thus were not a reliable predictor of memory performance (Manns et al. 2007). Considering this crucial role of time in episodic memories, our goal in the work described here was to focus on the order of events.

A major reason that the Fortin et al. (2002) paradigm has been so influential is that memory for order, but not recognition performance, was impaired in rats with lesions of the hippocampus, a structure widely implicated in episodic memory (Hasselmo and Eichenbaum 2005; Ross et al. 2009; Vargha‐Khadem et al. 1997). Further reinforcing the view that our adaptation of the paradigm did not capture the same memory functions, hippocampus removal did not impair memory for order in this task (Basile et al. 2020).

In conclusion, this paradigm was developed in monkeys with the aim of establishing a nonhuman primate model of episodic memory. The patterns of performance observed in monkeys (Templer and Hampton 2013) and rodents (Fortin et al. 2002; Kesner et al. 2002) in this paradigm are closely parallel, including symbolic distance effects and the dissociation of memory for order and recognition memory. However, it appears that differences in the implementation of the paradigm—most likely the shorter memory intervals used with monkeys—caused recruitment of different memory systems between the two species. Thus, the present results do not recommend this procedure as a model of episodic memory in monkeys, but they do document a striking dissociation of memory for order and recognition in monkeys and suggest this order memory paradigm recruits working memory processes.

Funding

This work was supported by the National Science Foundation BCS‐0745573 and the National Institutes of Health's Office of the Director, Office of Research Infrastructure Programs (P51OD011132).

Ethics Statement

All procedures used in the study were approved by the Emory University Institutional Animal Care and Use Committee (IACUC).

Conflicts of Interest

The authors declare no conflicts of interest.

Acknowledgments

The authors thank Tara Dove‐VanWormer, Emily Brown, and Steven Sherrin for technical assistance.

Data Availability Statement

The data that support the findings are available from the corresponding authors upon request.

References

  1. Basile, B. M. , and Hampton R. R.. 2010. “Rhesus Monkeys ( Macaca mulatta ) Show Robust Primacy and Recency in Memory for Lists From Small, but Not Large, Image Sets.” Behavioural Processes 83, no. 2: 183–190. 10.1016/j.beproc.2009.12.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Basile, B. M. , and Hampton R. R.. 2013. “Dissociation of Active Working Memory and Passive Recognition in Rhesus Monkeys.” Cognition 126, no. 3: 391–396. 10.1016/j.cognition.2012.10.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Basile, B. M. , and Hampton R. R.. 2017. “Dissociation of Item and Source Memory in Rhesus Monkeys.” Cognition 166: 398–406. 10.1016/j.cognition.2017.06.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Basile, B. M. , Templer V. L., Gazes R. P., and Hampton R. R.. 2020. “Preserved Visual Memory and Relational Cognition Performance in Monkeys With Selective Hippocampal Lesions.” Science Advances 6, no. 29: eaaz0484. 10.1126/sciadv.aaz0484. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Brady, R. J. , and Hampton R. R.. 2018. “Nonverbal Working Memory for Novel Images in Rhesus Monkeys.” Current Biology 28, no. 24: 3903–3910. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Clayton, N. S. , and Dickinson A.. 1998. “Episodic‐Like Memory During Cache Recovery by Scrub Jays.” Nature 395, no. 6699: 272–274. 10.1038/26216. [DOI] [PubMed] [Google Scholar]
  7. Clayton, N. S. , Griffiths D. P., Emery N. J., and Dickinson A.. 2001. “Elements of Episodic–Like Memory in Animals.” Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences 356, no. 1413: 1483–1491. 10.1098/rstb.2001.0947. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Crystal, J. D. 2009. “Elements of Episodic‐Like Memory in Animal Models.” Behavioural Processes 80, no. 3: 269–277. [DOI] [PubMed] [Google Scholar]
  9. Crystal, J. D. 2010. “Episodic‐Like Memory in Animals.” Behavioural Brain Research, Special Issue on Episodic Memory 215, no. 2: 235–243. 10.1016/j.bbr.2010.03.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Crystal, J. D. , Alford W. T., Zhou W., and Hohmann A. G.. 2013. “Source Memory in the Rat.” Current Biology 23, no. 5: 387–391. 10.1016/j.cub.2013.01.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Ebbinghaus, E. 1902. Grundziige der Psychologie. Veit & Comp. [Google Scholar]
  12. Eichenbaum, H. 2013. “Memory on Time.” Trends in Cognitive Sciences 17, no. 2: 81–88. 10.1016/j.tics.2012.12.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Eichenbaum, H. 2017. “Time (And Space) in the Hippocampus.” Current Opinion in Behavioral Sciences, Memory in Time and Space 17: 65–70. 10.1016/j.cobeha.2017.06.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Eichenbaum, H. , Yonelinas A. P., and Ranganath C.. 2007. “The Medial Temporal Lobe and Recognition Memory.” Annual Review of Neuroscience 30(Volume 30, 2007: 123–152. 10.1146/annurev.neuro.30.051606.094328. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Feeney, M. C. , Roberts W. A., and Sherry D. F.. 2009. “Memory for What, Where, and When in the Black‐Capped Chickadee ( Poecile atricapillus ).” Animal Cognition 12, no. 6: 767–777. 10.1007/s10071-009-0236-x. [DOI] [PubMed] [Google Scholar]
  16. Fortin, N. J. , Agster K. L., and Eichenbaum H. B.. 2002. “Critical Role of the Hippocampus in Memory for Sequences of Events.” Nature Neuroscience 5, no. 5: 458–462. 10.1038/nn834. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Funahashi, S. , and Kubota K.. 1994. “Working Memory and Prefrontal Cortex.” Neuroscience Research 21, no. 1: 1–11. 10.1016/0168-0102(94)90063-9. [DOI] [PubMed] [Google Scholar]
  18. Fuster, J. M. , and Alexander G. E.. 1971. “Neuron Activity Related to Short‐Term Memory.” Science 173, no. 3997: 652–654. 10.1126/science.173.3997.652. [DOI] [PubMed] [Google Scholar]
  19. Gazes, R. P. , Brown E. K., Basile B. M., and Hampton R. R.. 2013. “Automated Cognitive Testing of Monkeys in Social Groups Yields Results Comparable to Individual Laboratory‐Based Testing.” Animal Cognition 16, no. 3: 445–458. 10.1007/s10071-012-0585-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hampton, R. R. , Hampstead B. M., and Murray E. A.. 2005. “Rhesus Monkeys ( Macaca mulatta ) Demonstrate Robust Memory for What and Where, but Not When, in an Open‐Field Test of Memory.” Learning and Motivation 36, no. 2: 245–259. [Google Scholar]
  21. Hasselmo, M. E. , and Eichenbaum H.. 2005. “Hippocampal Mechanisms for the Context‐Dependent Retrieval of Episodes.” Neural Networks, Computational Theories of the Functions of the Hippocampus 18, no. 9: 1172–1190. 10.1016/j.neunet.2005.08.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Johnson, M. K. , Hashtroudi S., and Lindsay D. S.. 1993. “Source Monitoring.” Psychological Bulletin 114, no. 1: 3–28. 10.1037/0033-2909.114.1.3. [DOI] [PubMed] [Google Scholar]
  23. Jozet‐Alves, C. , Bertin M., and Clayton N. S.. 2013. “Evidence of Episodic‐Like Memory in Cuttlefish.” Current Biology 23, no. 23: R1033–R1035. 10.1016/j.cub.2013.10.021. [DOI] [PubMed] [Google Scholar]
  24. Kesner, R. P. , Gilbert P. E., and Barua L. A.. 2002. “The Role of the Hippocampus in Memory for the Temporal Order of a Sequence of Odors.” Behavioral Neuroscience 116, no. 2: 286–290. 10.1037/0735-7044.116.2.286. [DOI] [PubMed] [Google Scholar]
  25. Logie, R. H. 1986. “Visuo‐Spatial Processing in Working Memory.” Quarterly Journal of Experimental Psychology Section A 38, no. 2: 229–247. 10.1080/14640748608401596. [DOI] [PubMed] [Google Scholar]
  26. Macmillan, N. A. , and Creelman C. D.. 1991. Detection Theory: A User's Guide, 407. Cambridge University Press. [Google Scholar]
  27. Manns, J. R. , Howard M. W., and Eichenbaum H.. 2007. “Gradual Changes in Hippocampal Activity Support Remembering the Order of Events.” Neuron 56, no. 3: 530–540. 10.1016/j.neuron.2007.08.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Martin‐Ordas, G. , Haun D., Colmenares F., and Call J.. 2010. “Keeping Track of Time: Evidence for Episodic‐Like Memory in Great Apes.” Animal Cognition 13, no. 2: 331–340. 10.1007/s10071-009-0282-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Miller, E. K. , Erickson C. A., and Desimone R.. 1996. “Neural Mechanisms of Visual Working Memory in Prefrontal Cortex of the Macaque.” Journal of Neuroscience 16, no. 16: 5154–5167. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Petrides, M. 1995. “Impairments on Nonspatial Self‐Ordered and Externally Ordered Working Memory Tasks After Lesions of the Mid‐Dorsal Part of the Lateral Frontal Cortex in the Monkey.” Journal of Neuroscience 15, no. 1: 359–375. 10.1523/JNEUROSCI.15-01-00359.1995. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Phillips, W. A. , and Christie D. F. M.. 1977. “Interference With Visualization.” Quarterly Journal of Experimental Psychology 29, no. 4: 637–650. 10.1080/14640747708400638. [DOI] [PubMed] [Google Scholar]
  32. Ranganath, C. 2019. “Time, Memory, and the Legacy of Howard Eichenbaum.” Hippocampus 29, no. 3: 146–161. 10.1002/hipo.23007. [DOI] [PubMed] [Google Scholar]
  33. Ranganath, C. , and Hsieh L.‐T.. 2016. “The Hippocampus: A Special Place for Time.” Annals of the New York Academy of Sciences 1369, no. 1: 93–110. 10.1111/nyas.13043. [DOI] [PubMed] [Google Scholar]
  34. Ross, R. S. , Brown T. I., and Stern C. E.. 2009. “The Retrieval of Learned Sequences Engages the Hippocampus: Evidence From fMRI.” Hippocampus 19, no. 9: 790–799. 10.1002/hipo.20558. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Schwartz, B. L. , Hoffman M. L., and Evans S.. 2005. “Episodic‐Like Memory in a Gorilla: A Review and New Findings.” Learning and Motivation, Cognitive Time Travel in People and Animals 36, no. 2: 226–244. 10.1016/j.lmot.2005.02.012. [DOI] [Google Scholar]
  36. Smith, S. M. 1994. “Theoretical Principles of Context‐Dependent Memory.” In Theoretical Aspects of Memory, vol. 2, 168–195. Routledge. [Google Scholar]
  37. Templer, V. L. , and Hampton R. R.. 2013. “Cognitive Mechanisms of Memory for Order in Rhesus Monkeys (Macaca mulatta) .” Hippocampus 23, no. 3: 193–201. 10.1002/hipo.22082. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Tulving, E. 1972. “Episodic and Semantic Memory.” In Organization of Memory, 381–403. Academic Press. [Google Scholar]
  39. Tulving, E. 1983a. Elements of Episodic Memory. Oxford University Press. [Google Scholar]
  40. Tulving, E. 1983b. Elements of Episodic Memory. Oxford Univ. Press. [Google Scholar]
  41. Tulving, E. 1993. “What Is Episodic Memory?” Current Directions in Psychological Science 2, no. 3: 67–70. 10.1111/1467-8721.ep10770899. [DOI] [Google Scholar]
  42. Tulving, E. , Terrace H. S., and Metcalfe J.. 2005. “Episodic Memory and Autonoesis: Uniquely Human.” Missing Link in Cognition: Origins of Self‐Reflective Consciousness 3: 56. [Google Scholar]
  43. Vargha‐Khadem, F. , Gadian D. G., Watkins K. E., Connelly A., Van Paesschen W., and Mishkin M.. 1997. “Differential Effects of Early Hippocampal Pathology on Episodic and Semantic Memory.” Science 277, no. 5324: 376–380. 10.1126/science.277.5324.376. [DOI] [PubMed] [Google Scholar]
  44. Wright, A. A. , Santiago H. C., Sands S. F., Kendrick D. F., and Cook R. G.. 1985. “Memory Processing of Serial Lists by Pigeons, Monkeys, and People.” Science 229, no. 4710: 287–289. 10.1126/science.9304205. [DOI] [PubMed] [Google Scholar]
  45. Xiong, S. , Sheridan C. L., Harrison T., et al. 2026. “Rats Replay Episodic Memories in Context.” Current Biology 36, no. 4: 1060–1066.e5. 10.1016/j.cub.2025.12.049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Zinkivskay, A. , Nazir F., and Smulders T. V.. 2009. “What–Where–When Memory in Magpies ( Pica pica ).” Animal Cognition 12, no. 1: 119–125. 10.1007/s10071-008-0176-x. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The data that support the findings are available from the corresponding authors upon request.

Articles from Hippocampus are provided here courtesy of Wiley

RESOURCES